Running Head: GENDER AND STRESS AS PREDICTORS OF DEPRESSION

Gender and Stress as Predictors of Depression

Zae’Cari Nelson

California Baptist University

Gender and Stress as Predictors of Depression 1

Gender and Stress as Predictors of Depression

More than 17 million adults in the United States experience the ill effects of depression,

making it perhaps the most well-known mental illness in the U.S.A. Depression influences an

expected one out of 15 adults. What’s more, one out of six individuals will encounter depression

in their life (What is Depression?). There are a mind-boggling number of elements that can

prompt depressive symptoms in male and female individuals, one of which is held to be a rise in

stress hormone disturbances. Male and female genders rely on different coping strategies to

handle stressors adequately and effectively. The purpose of this study is to explore gender as a

predictor of depression controlling for stress.

Depression is a medical condition that is both common and serious. It influences how an

individual think, act, and feel. Depression also results in a loss of interest in things that were

once enjoyable. An assortment of emotional and physical problems as well as a decrease in one’s

ability to function at home and/or in a public setting can be prompted by the onset of depression

(What is Depression?). Physicians and researchers have studied depression for several years and

have led research on numerous individual symptoms along with clustered symptoms, and an

excess of factors that might be the reason for developing depressive symptoms. Some research

studies identify participants by gender and age. Depressive symptoms based on life stressors of

medical students ages twenty-six to thirty were examined in a recent study. The study showed

that depressive symptoms based on life stressors of medical students ages twenty-six to thirty

expanded by 71% over a 3-month, 6-month, 9-month, and 12-month time span (Fried, 2015).

This specific study comprised of an aggregate of 3,021 people: 52.6% female and 48.4% male.

Additionally, depressive symptoms when explained by both males and females, are almost

1

Gender and Stress as Predictors of Depression 2

identical when life stress is a contributory factor of the major depressive disorder. This suggests

that gender may not be a significant predictor of stress and depression.

Literature Review

Stress. Stress is characterized as how much you feel overpowered or incapable to adapt

because of pressing factors that are unmanageable (Stress MHF, 2021). At any rate, stress is our

body’s reaction to life occurrences as well as circumstances (Sha, 2006). Social conditions,

financial conditions, environment, and hereditary qualities are overall contributors of an

individual’s degree of stress. Numerous contributors of stress can shift vastly from one individual

to another, male and female, and so does how reactions are processed. Given the way that good

and regrettable life changes can both be a stressor for one individual, it would not be implausible

if stress factors like separation, work layoff, death, or monetary troubles prompt sentiments of

depression. It has been determined that men and women manage stress factors differently.

Though the two sexes regularly report comparable feelings of anxiety, yet report fluctuating

physical and emotional symptoms (Shih, 2004). Many depression symptoms carry somewhat of a

comparative result as having life altering stressors. In this way, to appropriately identify the

aspect of gender in predicting depression, it is important to govern for the possible effect of

stressors.

Gender. According to The World Health Organization, gender is the qualities of men,

women, boys, and girls that are communally developed also including standards, practices and

functions related with being a man, woman, boy or girl, just as correlations with one another. In a

socially productive way, gender will differ from one environment to another and can change after

some time. Gender is estimated to affect depression since an individual’s stress hormones and an

individual’s genetics proposes predominant controlling factors that add to a mental illness like

2

Gender and Stress as Predictors of Depression 3

depressive disorder. Ongoing studies have discovered that women are more probable than men to

encounter depression (Zwicker and DeLongis, 2010). A few examinations show that 33% of

women will encounter a substantial depressive episode during their life. (Hyde, 2020). There is

around a 40% chance of heritability when first-degree family members such as parents, children,

and siblings, have depression (Sowa and Lustman, 1984). Existing studies that concentrate on the

correlation between gender and depression has been led primarily centered around age groups,

the current study will expand the literature by researching if depression is predicted by gender.

To state the hypothesis, depression controlling for the impact of stressors gender will largely be

predicted by depression.

Method

Participants

101 adult Online and Professional Studies undergrads at California Baptist University

(CBU) were the participants. California Baptist University is a private university of the Christian

faith located in the Southern region of California; Riverside, CA and has an enlistment of around

11,317 students. The participants ages extended from 18 to 54 years; 12 males and 88 females. In

addition to providing spiritual and social expansion opportunities, the Online and Professional

Studies division at CBU offers over 40 fully accredited online degree programs to their students.

Measures

Sex/Gender. Respondents were queried to specify their gender. Feedback choices

included male (12%) and female (87%).

Stress. A four item self-report measure of stress levels in adult college students was used

to measure stress levels of the participants. This subscale is called the DASS-21 Stress Subscale,

3

Gender and Stress as Predictors of Depression 4

it’s responses consist of stress states such as “I feel that I utilize a ton of anxious energy”) on a

7-point Likert-type scale going from 1 (doesn’t concern me by any stretch of the imagination) to

7 (more often than not). Participants recorded their responses using the applicable corresponding

stress states. In past studies, proof of moderate to strong internal reliability was found, with a

range of .59 to .81 for Chronbach’s alphas (Osman et al. 2012). The validity of measure is

supported by positively correlating scores from the measure of mixed depression and

anxiety/stress and DASS-21 Stress subscale scores.

Factor analysis. To inspect the factor structure of the four-item scale, an exploratory

factor analysis (EFA) was executed.

According to Kaiser’s K-1 rule, there is a one-factor solution due to only one factor

having an eigenvalue greater than 1. The principal factor clarified 57.57% of the variety in

scores. To establish if the scale acted for a unidimensional construct, a scree test was

administered. The point of inflection is plainly indicated by the scree plot at the second

component. This suggests a one-factor solution. Afterward, to establish if all items loaded

heavily on the first factor, the component matrix had to be inspected. As the result, each of the

four items were determined to have loadings with absolute values more than .40 on the first

factor.

4

Gender and Stress as Predictors of Depression 5

Reliability analysis. To establish the stress scales’ internal reliability, a reliability analysis

was administered. As the result of the reliability analysis, there was a Chronbach’s alpha score

of .75. This was acceptable for psychometric purposes.

Distribution of composite scores. At last, the four-item scale was used to generate a

composite score, which produced a homogenized (M=0.00, SD=1.00) stress level score for all

respondents in the sample, going from – 1.52 to 2.71. The distribution was slightly leptokurtic

with a moderate positive skew as the result. Scores were on the low to mid-range with very few

showing significant degrees of stress for respondents.

5

Gender and Stress as Predictors of Depression 6

Depression. The DASS-21 Depression Subscale was used to measure the levels of

depression for respondents. This subscale is a four item self-report measure of depression levels

in adult college pupils. Respondents provided responses to items showing a depressed state (e.g.,

“I feel that I have nothing to anticipate”) on a 7-point Likert-type scale going from 1 (doesn’t

concern me by any means) to 7 (more often than not). In earlier studies, indications of moderate

to strong internal reliability was found, with Chronbach’s alphas going from .63-.70 (Osman et

al. 2012). It was found that scores on a measure of depression and anxiety/stress combined

correspond positively with scores from the DASS-21 depression subscale. Therefore, the validity

of the measure is supported.

Factor analysis. To inspect the four-item scale factor structure, an Exploratory Factor

Analysis (EFA) was ran.

6

Gender and Stress as Predictors of Depression 7

An eigenvalue greater than 1, was reported for only one factor and 71.74% of the

variation in scores was explained by the first factor. Thus, in accordance with Kaiser’s K-1 rule,

the data offer that there is a one-factor solution. Similarly, to establish the unidimensional

construct within the scale, a scree test was ran. Since the point of inflection is shown plainly at

the second component on the scree plot, a one-factor solution is implied. Moreover, to establish

if all the items loaded heavily onto the first factor, the component matrix was inspected. As result

of the inspection, it was determined that loadings with absolute values over .40 on the first factor

was apparent for all four of the items.

7

Gender and Stress as Predictors of Depression 8

Reliability analysis. To establish the internal reliability of the depression level scale, a

reliability analysis was ran. As result of the reliability analysis, the Chronbach’s alpha score was .

86, which was satisfactory for psychometric purposes.

Distribution of composite scores. Using the four-item scale, a composite score was

generated. Resulting in a homogenous (M=0.00, SD=1.00) depression level score for all

respondents in the sample going from – .81 to 3.48. Subsequently, the distribution result was

leptokurtic with a positive skew. Participants scored mostly on the low end of the depression

level scale with only a few signifying high levels of depression.

8

Gender and Stress as Predictors of Depression 9

Procedure

Questionnaires were anonymously administered to the adult student participants that were

enrolled in the Online and Professional Studies division of CBU for the psychology major.

Because the online questionnaire was completed outside of class hours, students were given

additional credit.

Results

Descriptive Statistics

Before visually inspecting the correlation between independent and dependent variables,

the figures had to be established first. Next, an examination was completed to establish the

correlation between sex and depression. Since sex is a categorical predictor, a boxplot was used

to execute the examination.

Median depression level for both genders was established through a visual examination

of the boxplot. The visual examination holds that the median depression level for females is like

that of males. For the male scores there is more variability as some males scored comparatively

9

Gender and Stress as Predictors of Depression
10

high whereas the females scored slightly low with some potential outliers. A slight positive

correlation amongst stress levels and depression was apparent in a visual assessment of the

scatterplot. Generally, as stress levels increased so did depression levels. Though, some

respondents scored low in both. To inspect the correlation amongst each variable, both

independent and dependent, a correlation test was run.

The results of the correlation test concluded that stress and depression had a significant

moderate positive correlation, r (96) = .49, p<.01. However, gender and depression presented a

very small positive relationship r (96) = .06, p=.59. An issue with collinearity was not presented

in this model as the two independent variables do not significantly correlate with one another.

Checking Assumptions

Homogeneity of variance. The analysis was comprised of both independent variables

held in the model to test the assumptions. Predicted values and unstandardized residuals were

kept as new factors in the dataset. The predicted values were placed on the x-axis while the

residuals were placed on the y-axis of the scatterplot. This was done to test for homogeneity of

variance.

10

Gender and Stress as Predictors of Depression
11

The distribution of residuals was revealed through a visual examination of the scatterplot.

As the result of the visual examination, it was determined that the residuals are not evenly

distributed. Clusters and slight fanning appear on the scatterplot also. Since equal predictive

power for persons at the high end of the line and the low end of the depression level scale is not

present in the model, the assumption of homogeneity of variance is not supported.

Normally distributed residuals. A histogram of the distribution was constructed to find

out how residuals were distributed. The goal was to determine if residuals were normally

distributed. To do so, a normal curve was superimposed on top. As the result, the assumption of

normally distributed residuals was not supported because they are not normally distributed.

11

Gender and Stress as Predictors of Depression
12

Multiple Regression Analysis

Lastly, to establish if the hypothesis was supported, an interpretation for each of the

output tables in the multiple regression analysis was conducted.

12

Gender and Stress as Predictors of Depression
13

The model resulted in significantly improved predictions than predictions based solely on

the mean because the F-statistic was statistically significant, F(2, 93) = 16.28, p < .001.

According to the table, gender is not a statistically significant predictor of the

respondent’s depression levels. However, stress level is. Controlling for stress, my hypotheses is

not supported.

The regression equation is:

The expected depression level score for a student that has an average level of stress is .03.

Controlling for stress, there is a .10 standard deviation decrease in gender which does not

correlate to a .03 standard deviation in depression levels.

13

Gender and Stress as Predictors of Depression
14

Discussion

In view of the outcome, my hypothesis was rejected. With the reason being because when

controlling for stressors, depression is not significantly predicted by sex. However, when using

sex as an independent variable, it makes for an unbalanced participant pool since the number of

participants total to 101, but with only 12 males and 89 females. A more equally balanced

number of participants would have essentially prompted an improved and more precise outcome.

In spite of the fact that review discoveries are positively precise, there are a few limitations. To

begin, this study included a wide array of participants from ages 18 years of age to 54 years of

age. I view such large age fluctuations as a substantial limitation. As an individual’s level of

maturity has much contingency on the way they process a stressor. Regardless of sex, adults that

have more life experiences then to process stressors more maturely than adults that have less

experiences in life.

The results of this investigation liken results from a study that was conducted in February

2015 by Eiko I Fried, an Associate Professor in Clinical Psychology at Leiden University. In the

study led by Fried, male and female participants described their depressive symptoms similarly

to how they were described in this study. Which suggests that when correlating stress and

depression, sex may not be a significant factor.

This study, specifically, can undoubtedly be summed up to the college populace by just

interpreting the results from an sample populace and applying it to the populace at large: Making

the investigation more equivalent as it relates to sex and increasing the sample size to be greater

than 101 participants. Hypotheses formed by many speculators that theorize that males have less

life stressors than females and therefore suffer from symptoms of depression less than females

are only stating a hypothesis not back by evidence that any statistical significance is present.

14

Gender and Stress as Predictors of Depression
15

In light of the results of this study, future scientists ought to follow an alternate

methodology to acquire more accurate and exact discoveries. Future scientists would benefit

from choosing an equivalent measure of male and female participants just as dividing the

participants by age. This set out as a better procedure for future studies and establishes a

clearer path for impending research to trail.

My experimental results conclusion rejected my hypothesis on sex assuming a huge part

in stress and depression. Truth be told, this study completely annulled my underlying speculation

though I firmly held males would have less stressors and thusly lower levels of depression than

females. My general impression, which is not in accordance with my referenced study or the

present study, is that the exploration study was very amazing, and I would be keen on proceeding

with it with a more unambiguous scope of participants to decide whether there is any chance of

truth to my original investigation and research.

15

Gender and Stress as Predictors of Depression
16

References

Stress. Mental Health Foundation. https://www.mentalhealth.org.uk/a-to-
z/s/stress#:~:text=Stress%20can%20be%20defined%20as,of%20pressures%20that
%20are%20unmanageable.

Shih, J. H.-F. (2004). Sociotropy/autonomy and depression: Gender differences and the
mediating role of stressful life events ProQuest Information & Learning]. APA PsycInfo.
http://libproxy.calbaptist.edu/login? url=https://search.ebscohost.com/login.aspx?
direct=true&db=psyh&AN=2004-99010- 099&site=ehost-live&scope=site

Zwicker, A., & DeLongis, A. (2010). Gender, stress, and coping. In J. C. Chrisler & D. R.
McCreary (Eds.), Handbook of gender research in psychology, Vol 2: Gender research in
social and applied psychology. (pp. 495-515). Springer Science + Business Media.
https://doi.org/10.1007/978-1-4419-1467-5_21

Sowa, C. J., & Lustman, P. J. (1984). Gender differences in rating stressful events, depression,
and depressive cognition. Journal of Clinical Psychology, 40(6), 1334- 1337.
https://doi.org/10.1002/1097-4679(198411)40:63.0.CO;2-8

What is Depression? American Psychiatric Association.
https://www.psychiatry.org/patientsfamilies/depression/what-is-depression

Sha, T. (2006). Optimism, Pessimism and Depression; The Relations and Differences by Stress
Level and Gender. Acta Psychologica Sinica, 38(6), 886-901.

Fried, E. I., Nesse, R. M., Guille, C., & Sen, S. (2015). The differential influence of life stress on
individual symptoms of depression. Acta Psychiatrica Scandinavica, 131(6), 465- 471.
https://doi.org/10.1111/acps.12395

Hyde, J. S., & Mezulis, A. H. (2020). Gender differences in depression: Biological, affective,
cognitive, and sociocultural factors. Harvard Review of Psychiatry, 28(1), 4- 13.
https://doi.org/10.1097/HRP.0000000000000230

16

82 C a n a d i a n J o u r n a l o f C o u n s e l l i n g / R e v u e canadienne de c o u n s e l i n g / 1 9 9 7 , V o l . 31:1

Norms and Construct Validity of the Rosenberg
Self-Esteem Scale i n Canadian H i g h School Populations:
Implications for Counselling

Christopher Bagley
University of Calgary
Floyd Bolitho
University of South Australia
Lome Bertrand
Canadian Research Institute for Law and the Family

Abstract

A r a n d o m sample of A l b e r t a h i g h schools yielded data for 1,084 males a n d 1,024 females.
Measures completed i n c l u d e d the Rosenberg Self-Esteem Scale. W i t h i n each age a n d sex group
a strong, unrotated factor was invariant. T h e r e was significant variation o f mean scores across
age-groups within female students. In each age group females had significantly lower self-
esteem than males. Evidence o f construct validity is derived from correlations with behaviour
p r o b l e m categories; the M c M a s t e r measure o f family relationships; and self-completion meas-
ures o f school climate, a n d physical a n d sexual victimization i n school. Possible uses for self-
esteem measures i n school counselling are proposed.

R é s u m é

U n é c h a n t i l l o n p r é l e v é au hasard d ‘ é c o l e s secondaires en Alberta a fourni des d o n n é e s de 1,084
é t u d i a n t s et de 1,024 é t u d i a n t e s . Parmi les mesures accomplies figurait l ‘ E c h e l l e d u respect de
soi Rosenberg. A l ‘ i n t é r i e u r de chaque groupe é t a b l i selon l ‘ â g e et le sexe, u n facteur solide et
n o n a l t e r n é é t a i t invariant. Il y avait une variation significative des scores moyens entre les
tranches d ‘ â g e , à l ‘ i n t é r i e u r d u groupe des é t u d i a n t e s . Dans chaque tranche d ‘ â g e , les jeunes
femmes avaient u n respect de soi c o n s i d é r a b l e m e n t i n f é r i e u r à c e l u i des jeunes hommes. L a
preuve de la validité de construction de l ‘ E c h e l l e Rosenberg vient des c o r r é l a t i o n s entre les
c a t é g o r i e s de troubles d u comportement; la mesure M c M a s t e r des relations familiales; et des
mesures que les é t u d i a n t s ont r e m p l i e u x – m ê m e s au sujet d u climat scolaire et de l ‘ i n t i m i d a t i o n
physique et sexuelle à l ‘ é c o l e . O n propose des utilisations possibles des mesures d u respect de
soi p o u r le c o u n s e l i n g scolaire.

I N T R O D U C T I O N

Self-esteem, the m a n n e r i n w h i c h an i n d i v i d u a l evaluates self-
characteristics relative to the perceived characteristics o f peers, is a
crucial variable for u n d e r s t a n d i n g identity development, a n d u n d e r p i n s
the development o f m e n t a l health adjustment ( Y o u n g & Bagley, 1982;
Bagley & Y o u n g , 1990). Self-esteem a n d self-concept (salient self charac-
teristics w h i c h the i n d i v i d u a l considers worthy o f evaluation) can be
measured i n g l o b a l terms (by an affective construct by w h i c h many
aspects o f self-functioning a n d self-worth are evaluated), o r by m o r e
specific evaluations o f role performance (e.g. performance i n r e a d i n g ,
mathematics, i n sport etc.). By adolescence the y o u n g person w i l l have

Rosenberg Self-Esteem Scale 83

a c q u i r e d a stable set o f self-evaluations so that it becomes increasingly
l i k e l y that even specific tasks (e.g. academic learning) w i l l be e n h a n c e d
o r i n h i b i t e d by pre-existing self-esteem.

A l t h o u g h there are many different measures o f self-esteem available
for teachers a n d counsellors, a review o f available measures by Blascovich
a n d T o m a k a (1991) indicates that eleven measures o f the several d o z e n
available have g o o d evidence o f reliability a n d validity a n d widespread
use by researchers. T h e most frequently used o f these measures is the
R o s e n b e r g Self-Esteem Scale (RSES) (Blascovich & T o m a k a , 1991) fol-
lowed by the C o o p e r s m i t h scale (Bagley, 1989), the Piers-Harris Scale
(Bagley & M a l l i c k , 1978), a n d the Tennessee Scale ( R o i d & Fitts, 1988).
These four scales account for 60 percent o f the j o u r n a l citations o f
self-esteem a n d self-concept studies. Since the p r i n c i p l e measures o f
g l o b a l self-esteem usually have intercorrelations o f a r o u n d 0.6 to 0.7
(Blascovich & T o m a k a , 1991) the selection o f a particular measure may
rest o n criteria such as the brevity o f the scale, a n d the ease with w h i c h
i n d i v i d u a l s with relatively p o o r r e a d i n g skills can understand it. T h e
R o s e n b e r g Self-Esteem Scale i n its brevity (10 items), a n d its easy-to-
understand format is, i n A m e r i c a n research, the instrument o f choice for
use with adolescent populations. In the standard text o n psychological
measurements, Blascovich a n d T o m a k a (1991) observe:

The Rosenberg SES has enjoyed widespread use and utility as a unidimensional
measure of self-esteem. In fact, the SES is the standard against which new measures
are evaluated. Its ease of administration, scoring, and brevity underlie our recom-
mendation for the use of the SES as a straightforward estimate of positive or
negative feelings about the self. (p. 123)

Rosenberg’s SES was d e v e l o p e d for use i n state-wide U . S . studies o f the
adaptation o f y o u t h (Rosenberg, 1965; Rosenberg & S i m m o n s , 1972),
a n d was used by K a p l a n a n d P o k o r n e y (1976) a n d K a p l a n (1980) i n
seminally i m p o r t a n t work o n the predictive power o f self-esteem. In
K a p l a n ‘ s study a c o h o r t o f 4,694 c h i l d r e n e n t e r i n g j u n i o r h i g h schools i n
H o u s t o n , Texas i n 1969, c o m p l e t e d a 7-item version o f the RSES ( i n c l u d –
i n g the five negatively w o r d e d statements, a n d two positively w o r d e d
statements i n the RSES) to construct a scale o f “self-derogation.” Follow-
u p o f this c o h o r t into a d u l t h o o d f o u n d that low SES scores were strong
predictors o f d e l i n q u e n t behaviours, u n w e d pregnancy, d r u g use, a n d
suicidal behaviours. T h e clear i m p l i c a t i o n o f these findings is that i f
negative self-esteem can be d i m i n i s h e d i n the elementary o r j u n i o r h i g h
school years, many negative behaviours associated with i m p a i r e d self-
esteem m i g h t be avoided.

T h e i m p o r t a n c e o f the RSES i n p r e d i c t i n g d e l i n q u e n c y a n d depres-
sion has been r e p l i c a t e d i n studies o f large, n a t i o n a l cohorts o f adoles-
cents by R o s e n b e r g a n d R o s e n b e r g (1978) a n d Rosenberg, S c h o o l e r a n d
S c h o e n b a c h (1989). T h e m o d e l o f d e l i n q u e n c y causation using self-

84 C h r i s t o p h e r Bagley, Floyd B o l i t h o , L o m e Bertrand

esteem as a key, predictive variable has been replicated with H o n g K o n g
adolescents ( L e u n g & L a u , 1989), suggesting that the m o d e l has strong,
cross-cultural validity.

T h e RSES was devised as a unitary scale, a n d A m e r i c a n work has
consistently f o u n d a l p h a values i n excess o f 0.85 for the scale. However,
K a p l a n a n d P o k o m e y ( 1976) f o u n d that while there was a strong general
factor i n the scale, rotation o f factors d i d demonstrate a two-factor
solution with positively w o r d e d items (e.g. “I take a positive attitude
towards myself’) l o a d i n g o n the first c o m p o n e n t , a n d negatively w o r d e d
items (e.g. “I certainly feel useless at times”) l o a d i n g o n the second.
T h e two c o m p o n e n t s were described by K a p l a n a n d P o k o r n e y (1976)
as “defense o f fndividual self-worth” a n d “self-derogation.” S h a h a n i ,
D i p b o y e a n d P h i l l i p s (1990) i n a study o f 1,762 U.S. adults c o n f i r m e d
this two-factor structure i n the scale, but also argue that the scale can
be regarded psychometrically as h a v i n g a coherent, u n i f i e d structure.
Vallieres a n d V a l l e r a n d (1990) using confirmatory factor analysis
t h r o u g h L I S R E L with data from Q u e b e c students c o m p l e t i n g a F r e n c h
version o f the RSES, also f o u n d that the RSES h a d a u n i f i e d factor
structure, with all items h a v i n g h i g h loadings o n a general factor.

C A N A D I A N S T U D I E S O F H I G H S C H O O L P O P U L A T I O N S
U S I N G T H E R O S E N B E R G S E S

A c o m p u t e r i z e d literature search i n d i c a t e d that over 1,000 A m e r i c a n
studies have used the R o s e n b e r g SES with h i g h school a n d j u n i o r college
populations. P u b l i s h e d C a n a d i a n studies using this scale, however, are
remarkably few. T h e f o l l o w i n g studies are all those that can be located:
L a z u r e a n d Persinger, 1992; Beales a n d B r o o k , 1990; Vallieres a n d V a l –
l e r a n d , 1990; Byrne, 1983 a n d 1990; Byrne a n d Shavelson, 1986 a n d
1987). A somewhat larger n u m b e r o f C a n a d i a n studies have studied
o l d e r college student a n d adult populations.

T h e most comprehensive o f the C a n a d i a n studies o f Adolescents are
those o f O n t a r i o h i g h school populations (grades 9 to 13) carried out by
Byrne a n d colleagues (1983, 1986, 1987, 1990). T h i s work has e x p l o r e d ,
using c o m p l e x statistical m o d e l l i n g techniques, the reliability o f some
measures o f general self-esteem ( i n c l u d i n g the Rosenberg S E S ) , the
relationship o f these scales with measures o f academic self-concept, a n d
the degree to w h i c h the c o m p l e x factor structure o f several self-esteem
a n d self-concept measures are c o m p a r a b l e across sex groupings. Byrne
(1983) showed that the RSES h a d adequate internal reliability, a n d test-
retest c o r r e l a t i o n o f 0.61 over a 7-month p e r i o d i n 929 O n t a r i o h i g h
school students i n grades 9 to 12. Byrne a n d Shavelson (1986, 1987) also
used the RSES scale total i n p r i n c i p a l c o m p o n e n t analysis w h i c h i n –
c l u d e d other measures o f self-esteem. T h e o n l y available C a n a d i a n study

Rosenberg Self-Esteem Scale 85

w h i c h carried out a factor analysis o f the scale items is that o f Vallieres and
V a l l e r a n d (1990), using a French-language version o f the R S E S .

B A C K G R O U N D O F T H E P R E S E N T S T U D Y

T h e measures r e p o r t e d i n the present research were c o l l e c t e d as part o f a
1993 study o f substance use by A l b e r t a adolescents ( B e r t r a n d , S m i t h ,
B o l i t h o & H o r n i c k , 1994). Stratified r a n d o m s a m p l i n g identified n i n e
s c h o o l districts with schools i n the p u b l i c a n d d e n o m i n a t i o n a l sectors,
representative o f u r b a n , small town a n d rural areas i n all parts o f the
Province o f A l b e r t a , y i e l d i n g a sample o f 1,084 male a n d 1,024 female
students. Besides the R o s e n b e r g S E S , students c o m p l e t e d the M c M a s t e r
measure o f family relationships; a description o f s c h o o l climate, a n d
personal v i c t i m i z a t i o n (physical a n d sexual) i n school; a n d measures o f
personal adjustment. T h e adjustment scales were those with established
reliability a n d validity i n the O n t a r i o C h i l d H e a l t h Study (Sanford,
O f f o r d , Boyle & Pearce, 1992).

T h e measure o f family relationships is the “general f u n c t i o n i n g sub-
scale” from the M c M a s t e r family assessment device (Epstein, B a l d w i n &
B i s h o p , 1983). T h i s 12-item scale contains items such as “There are a lot
o f b a d feelings i n o u r family” a n d “In times o f crisis we can turn to each
other for support.” T h i s scale has been used i n a n u m b e r o f previous
studies to measure the h e a l t h / p a t h o l o g y o f families with regard to
p a r e n t i n g styles, c o m m u n i c a t i o n , parental affection, a n d cohesiveness.
T h e a l p h a value for this scale i n the A l b e r t a sample was .90. T h e measure
o f s c h o o l climate was specially devised for this study, a n d contains items
such as ‘Teachers i n my s c h o o l care about students’ w o r k ” a n d ‘ T h e r e is a
lot o f fighting between students i n o r a r o u n d the s c h o o l . ” S c o r i n g o f
some items was reversed, so that a h i g h score o n this scale indicates
p e r c e p t i o n o f a positive s c h o o l e n v i r o n m e n t ; a l p h a for this scale is 0.78,
i n d i c a t i n g satisfactory i nt er nal reliability. T h e b r i e f measures o f physical
a n d sexual harassment i n s c h o o l were specially devised for this study, a n d
i n c l u d e items such as “Has someone slapped o r hit y o u i n anger?” a n d
“Has someone t o u c h e d the private parts o f your body w h e n y o u d i d n ‘ t
want them to?”

G i v e n the theoretical a n d substantive i m p o r t a n c e o f the A m e r i c a n
work o n the R o s e n b e r g SES a n d the dearth o f studies using this scale
i n C a n a d a , we present the means, standard deviations a n d a l p h a re-
liabilities for the scale by age a n d sex groups i n o u r A l b e r t a h i g h school
samples. P r i n c i p a l c o m p o n e n t analysis has been e m p l o y e d i n o r d e r to
e x a m i n e the structure a n d i n t e r n a l reliability o f the scale. C o r r e l a t i o n o f
the scale with measures o f family relationships, a n d s c h o o l e n v i r o n m e n t
have been used as cross-validators o n the assumption that quality o f
family life a n d s c h o o l e n v i r o n m e n t are i m p o r t a n t antecedents o f self-
esteem ( C o o p e r s m i t h , 1967; Bagley, Verma, M a l l i c k & Y o u n g , 1979).

86 C h r i s t o p h e r Bagley, Floyd B o l i t h o , L o m e Bertrand

T A B L E 1
General Factor of Rosenberg Self-Esteem Scale Items

Within Age Groups: Males

General Factor (loadings of 1st unrotated factor)
Scale Item* 12-19 18-19 16-17 14-15 12-13

1. Satisfied with self .69 .71 .65 .69 .73
2. I’m no good at all .67 .59 .65 .67 .72
3. Have good qualities .65 .73 .63 .65 .64
4. I can do things as well

as others .59 .61 .58 .63 .50
5. I don’t have ç i u c h to

be p r o u d o f .71 .65 .70 .72 .72
6. I feel useless .66 .63 .61 .68 .73
7. I am a person o f worth .55 .52 .62 .56 .48
8. I don’t respect myself .64 .50 .70 .60 .67
9. I ‘ m a failure .75 .73 .74 .74 .76

10. I have a positive self-
attitude .75 .73 .77 .76 .73

% o f variance 44.7% 41.8% 44.7% 45.2% 45.6%

N o . o f respondents 1084 98 338 428 220

Scale mean 31.36 31.59 30.88 31.60 31.50
Standard deviation 5.13 4.65 5.14 5.14 5.36
A l p h a .86 .85 .90 .86 .85

% with very poor self-
esteem (score < 21) 2.7% 1.0% 3.0% 2.1% 4.1%

* Scale items have been paraphrased a n d shortened. Since higher scores o n total scale indicate
better self-esteem, negatively w o r d e d items (2, 5, 6, 8, 9) are scored i n a reverse d i r e c t i o n .
M i n i m u m score o n scale is 10, m a x i m u m is 40.

Research u s i n g various tests o f self-esteem a n d self-evaluation have
consistently shown that females are m o r e l i k e l y to be self-critical
than males (e.g. M c D o n a l d & M c K i n n e y , 1994). Rather than i n d i c a t i n g
p o o r e r self-esteem, however, this may reflect a f e m i n i n e response style
o f b e i n g less self-declaratory o n the “powerful” aspects o f self, p l a c i n g
m o r e emphasis o n interpersonal aspects o f self-appraisal ( G i l l i g a n , 1982,
1990). We hypothesized, therefore, that females w o u l d have significantly
lower self-esteem scores than males.

R E S U L T S

P r i n c i p a l c o m p o n e n t analysis for each age a n d sex g r o u p i d e n t i f i e d a
powerful general factor, a c c o u n t i n g for between 40 a n d 50 percent o f the

Rosenberg Self-Esteem Scale 87

T A B L E 2
General Factor of Rosenberg Self-Esteem Scale Items

Within Age Groups: Females

General Factor (loadings of 1st unrotated factor)
Scale Item* 12-19 18-19 16-17 14-15 12-13

1. Satisfied with self .76 .66 .78 .75 .80
2. I’m no good at all .64 .70 .59 .67 .60
3. Have good qualities .73 .73 .75 .69 .77
4. I can do things as well

as others .65 .60 .63 .67 .69
5. I don’t have much to

be p r o u d of .74 .72 .72 .75 .72
6. I feel useless at times .63 .65 .60 .65 .64
7. I am a person of worth .67 .54 .73 .67 .61
8. I don’t respect myself .66 .56 .67 .65 .69
9. I’m a failure .77 .72 .77 .75 .81

10. I have a positive self-
attitude .80 .78 .79 .80 .79

% o f variance 50.2% 44.6% 51.1% 50.0% 51.9%

N o . o f students 1024 69 322 410 223

Scale mean 28.32 29.04 28.04 28.00 29.08
Standard deviation 5.49 5.11 5.42 5.48 5.81
A l p h a .89 .87 .89 .88 .89

% with very poor self-
esteem (score < 21) 7.5% 4.3% 7.1% 7.6% 8.5%

* Scale items have been paraphrased a n d shortened. Since higher scores o n total scale indicate
better self-esteem, negatively w o r d e d items (2, 5, 6, 8, 9) are scored i n a reverse direction.
M i n i m u m score o n scale is 10, m a x i m u m is 40.

total variance, a n d a second factor a c c o u n t i n g for some 12 percent o f
variance i n scale items. N o o t h e r p r i n c i p a l c o m p o n e n t h a d an E i g e n –
value greater than unity. V a r i m a x rotation o f the scale items does identify
the two factors d e s c r i b e d by K a p l a n a n d P o k o r n y (1969); however, given
the strength o f the g e n e r a l factor we present loadings for this unrotated,
general factor i n Tables 1 a n d 2. A l p h a values for the RSES range f r o m
0.85 to 0.90. T h i s is e x p l i c a b l e i n l i g h t o f the fact that RSES was c o n –
structed as a u n i d i m e n s i o n a l scale. M e a n values o f the SES are similar
across the male age g r o u p i n g s .

T h e r e are, however, some significant variations i n m e a n RSES scores
across the female age groups, with those aged 14 to 17 h a v i n g signifi-
cantly lower self-esteem than those aged 12 to 13, a n d those aged 18 to 19

88 C h r i s t o p h e r Bagley, Floyd B o l i t h o , L o m e Bertrand

T A B L E 3
Correlations of Rosenberg Self-Esteem Scale in Males

Variable 18-19 16-17 14-15 12-13

Ontario CHS Scales
Somatic problems -.29 -.44 -.42 -.43
Hyperactivity -.39 -.27 -.26 -.40
Conduct disorder -.26 -.29 -.20 -.26
Emotional disorder -.60 -.58 -.50 -.54

Other Scales
McMaster family relationships scale -.49 -.38 -.49 -.47

Positive school climate scale .32 .21 (.09) .32

Physical victimization i n school -.28 -.12* -.19 -.32

Sexual victimization i n school (-.07) (-.01) -.12* (-.08)

N o . o f respondents 98 338 428 220

(Correlations in brackets, not significant) * /><.05>.01 A l l other correlations p< .01.

T A B L E 4
Correlations of Rosenberg Self-Esteem Scale in Females

Variable 18-19 16-17 14-15 12-13

Ontario CHS Scales
Somatic problems -.37 -.36 -.44 -.49
Hyperactivity (-•09) -.37 -.41 -.34
Conduct disorder (-.20) -.27 -.38 -.34
Emotional disorder -.58 -.65 -.64 -.66

Other Scales
McMaster family relationships scale -.51 -.45 -.54 -.48

Positive school climate scale (•12) .18 .31 .45

Physical victimization in school (-.04) -.20 -.22 -.35

Sexual victimization in school (.03) -.21 -.17 -.27

N o . o f respondents 69 322 410 223

(Correlations i n brackets, not significant) * /><.05>.01 A l l other correlations p< .01.

(t-tests, /><.05). T h e reasons for these variations are not clear. W i t h i n
each age g r o u p females have significantly lower self-esteem than males o f
similar age (all t-tests p< .01 ). Females were at least twice as l i k e l y as males
to have “very p o o r self-esteem”: this i n d i c a t e d responses o f “agree” o r

Rosenberg Self-Esteem Scale 89

“strongly agree” i n a d i r e c t i o n i m p l y i n g low self-esteem o n seven o r m o r e
o f the items i n the 10-item scale.

Tables 3 a n d 4 present correlations g i v i n g evidence o f construct valid-
ity for the R S E S . F o r b o t h males a n d females o f all ages, g o o d self-esteem
is negatively correlated with the four sub-scales o f e m o t i o n a l a n d behav-
i o u r disorders (somatic p r o b l e m s , c o n d u c t disorder, a n d e m o t i o n a l dis-
o r d e r ) , with one e x c e p t i o n , that o f female students aged 18 a n d 19 for
w h o m hyperactivity a n d c o n d u c t disorder d i d not correlate significantly
with the R S E S . F o r a l l groups the e m o t i o n a l disorder scale has correla-
tions i n excess o f -0.50 (p< .001 ). T h e e m o t i o n a l disorder scale includes
items measuring depression, anxiety a n d suicidal feelings.

T h e M c M a s t e r measure o f family relationships also indicates evidence
o f construct validity for the R S E S : correlations (all jfx.001) range f r o m
-.38 to -.51 across the age a n d sex groups. Perceptions o f positive school
climate provide less consistent evidence o f construct validity; neverthe-
less, i n six o f the eight age a n d sex groups school climate is significantly
related to RSES (p< .001 i n the six groups). F o r boys, physical victimiza-
t i o n is l i n k e d to p o o r self-esteem, while for girls b o t h sexual a n d physical
v i c t i m i z a t i o n is l i n k e d to p o o r self-esteem, with the exception o f students
aged 18 a n d 19.

D I S C U S S I O N A N D C O N C L U S I O N

These findings indicate that the RSES is a reliable a n d potentially v a l i d
scale for use with C a n a d i a n h i g h school students o f all ages. T h e R S E S
means are close to those o b t a i n e d by Byrne ( 1990) with o l d e r h i g h school
students i n O n t a r i o ; similar to those o b t a i n e d by Vallieres a n d V a l l e r a n d
(1990) i n Q u e b e c , u s i n g a French-langauge version o f the RSES; a n d
similar to those o b t a i n e d i n A m e r i c a n studies o f h i g h school students
( M c D o n a l d & M c K i n n e y , 1994). L i k e A m e r i c a n researchers, we f o u n d
that males have h i g h e r scores o n the RSES than females, i n each o f the
four age-group comparisons. We hesitate, however, before c o n c l u d i n g
that females have “poorer” self-esteem than males, since the response
style o f females to self-esteem questions may be less egoistic, a n d m o r e
self-deprecating i n ways w h i c h i m p l y that relationship styles rather than
self-aggrandizement are m o r e i m p o r t a n t 1 . W h a t is salient is that some
females have very p o o r self-esteem relative to other females. T h e apparent
d e c l i n e i n self-esteem i n females i n mid-adolescence i n o u r A l b e r t a
sample is interesting, a n d possible reasons for this (such as changes i n sex
role identification, o r psychosocial development) s h o u l d be e x p l o r e d i n
further research. It is o f note that the finding o f d e c l i n i n g self-esteem i n
females i n the m i d d l e years o f h i g h school has also been r e p o r t e d by
Abernathy, Massad a n d Romano-Dwyer (1995) i n a T o r o n t o study o f
several thousand students i n grades 6 t h o u g h 9, using the self-esteem
scale d e v e l o p e d for the C a n a d a H e a l t h Attitudes a n d B e h a v i o u r Survey.

90 C h r i s t o p h e r Bagley, Floyd B o l i t h o , L o m e B e r t r a n d

A b e r n a t h y et al. (1995) i m p l y that this may be l i n k e d to the increase i n
the p r o p o r t i o n o f females who begin to smoke i n the m i d d l e years o f h i g h
school: females with low self-esteem i n grades 6 to 8 were about three
times as l i k e l y to smoke as those with h i g h self-esteem. However, o t h e r
psychosocial processes may u n d e r l i e b o t h b e g i n n i n g to smoke, a n d
d e v e l o p i n g p o o r e r self-esteem i n girls i n mid-adolescence.

O u r findings have i m p l i c a t i o n s for school counsellors, i n terms o f the
models p r o p o s e d i n o u r previous research (Bagley, 1976; Bagley et al.,
1979; Bagley, 1992). In an E n g l i s h study, we c o n d u c t e d a c o n t r o l l e d
intervention, r a n d o m l y a l l o c a t i n g h i g h school students who i n d i c a t e d
b o t h suicidal i d e a t i o n a n d devastated self-esteem (scores i n the lowest
5% c o m p a r e d with scores o n the C o o p e r s m i t h scale scores for the entire
c o h o r t o f 14 a n d 15-year o l d s ) . Intensive social interventions ( R o g e r i a n
c o u n s e l l i n g ; t u i t i o n for those with p o o r scholastic achievement; task-
linkage o f sociometrically isolated students with h i g h l y p o p u l a r students;
a n d social w o r k assistance for families o f students e x p e r i e n c i n g poverty,
family d i s r u p t i o n a n d divorce) was reflected a year later i n significant
improvements i n self-esteem, a n d eclipse o f suicidal feelings i n focus
students, i n c o m p a r i s o n with controls.

In a partial r e p l i c a t i o n o f this B r i t i s h work i n A l b e r t a , students c o m –
pleted a measure o f stressful events, self-esteem, depression a n d s u i c i d a l
ideas a n d b e h a v i o u r (Bagley, 1992). Results i n d i c a t e d that the strongest
predictors o f p o o r self-esteem, depression a n d suicidality were family
physical, sexual a n d e m o t i o n a l abuse, family d i s r u p t i o n , a n d parental
a l c o h o l i s m . Students were e n c o u r a g e d to contact school counsellors,
a n d a significant n u m b e r d i d so f o l l o w i n g the anonymous c o m p l e t i o n o f
the questionnaires. T h i s design (ensuring student anonymity) d i d not
p e r m i t a c o n t r o l l e d evaluation o f c o u n s e l l i n g . However, the results o f
that study, a n d the data presented i n the present study strongly support
the i d e a that i n every h i g h s c h o o l there is a g r o u p o f very distressed
students, with devastated self-esteem a n d various psychological prob-
lems, who m i g h t benefit f r o m c o u n s e l l i n g .

Note
1 Males who carry their “macho” sense o f self into a d u l t h o o d may fail, in Maslow’s (1954)

term, to self-actualize. T h e m a c h o male is aggressive a n d competitive, rejecting intimacy
a n d b o n d i n g relationships, seeking to express power over w o m e n , avoiding intimate b o n d s
with b o t h m e n a n d w o m e n . W o m e n by contrast may experience a m o r e complete self-
actualization t h r o u g h intimate b o n d s o f social support with the sisterhood o f w o m e n (Bagley
& Y o u n g , 1990). Excessively h i g h levels o f self-esteem i n males may in some respects represent
the narcissism o f “false self-esteem” w h i c h H w a n g (1995) categorizes as o n e o f the factors i n
“the failure o f A m e r i c a n schools.” A s counsellors attempting to enhance self-esteem in
adolescents we s h o u l d avoid strategies which m i g h t result in what H a u s e r (1971) describes as
“identity foreclosure.” A l l c o u n s e l l i n g s h o u l d attempt not only to e n h a n c e self-esteem, but
also to aid self-actualization (Bagley Sc Y o u n g , 1990).

Rosenberg Self-Esteem Scale 91

References
Abemathy, T., Massad, L . , & Romano-Dwyer, L . (1995). T h e relationship between smoking a n d

self-esteem. Adolescence, 30, 899-907.
Bagley, C . (1976). Suicidal behaviour a n d suicidal ideation i n adolescents: a p r o b l e m for

counsellors i n education. British Journal of Guidance and Counselling, 3, 190-208.
. ( 1989). Development of a short self-esteem measure for use with adults i n c o m m u n i t y

mental health surveys. Psychological Reports, 65, 13-14.
. (1992). Development o f an adolescent stress scale for use by school counsellors:

construct validity i n terms o f depression, self-esteem a n d suicidal ideation. School Psychology
International, 13, 31-49.

Bagley, C , & M a l l i c k , K . (1978). Development o f a short form o f the Piers Harris self-concept
scale. Educational Review, 30, 265-68.

Bagley, C , Verma, G . , M a l l i c k , K . , & Young, L . (1979). Personality, self-esteem and prejudice.
B r o o k f i e l d , V T : Avebury.

Bagley, C , & Y o u n g , L . (1990). Depression, self-esteem a n d suicidal behaviour as sequels
o f sexual abuse i n c h i l d h o o d : research a n d therapy In M . Rothery & G . C a m e r o n (Eds.).
Child maltreatment: Expanding our concepts of helping (pp. 183-219). H i l l s d a l e , N J : Lawrence
E r l b a u m .

Beales, J . , & B r o o k , Z . (1990). T h e effects o f h i g h school d r a m a o n social maturity. School
Counsellor, 38, 46-51.

Bertrand, L . , Smith, R., B o l i t h o , F., & H o m i c k , J . (1994). Substance use among Alberta adolescents:
prevalence and related factors. Calgary: C a n a d i a n Research Institute for Law a n d the Family.

Blascovich, J . , & T o m a k a , J . (1991). Measures o f self-esteem. In J . R o b i n s o n , Shaver &
L . Wrightsman (Eds). Measures of personality and psychological attitudes. New York: Academic
Press.

Byrne, B . (1983). Investigating measures o f self-concept. Measurement and Evaluation in Gui-
dance, 16, 115-26.

. (1990). Self-concept a n d academic achievement: investigating their importance as
discriminators of academic track membership i n h i g h school. Canadian Journal of Education,
15, 173-82.

Byrne, B . , & Shavelson, R. (1986). O n the structure o f adolescent self-concept. Journal of
Educational Psychology, 7#, 474-81.

. (1987). Adolescent self-concept: testing the assumption of equivalent structure across
gender. American Educational Research Journal, 24, 365-85.

C o o p e r s m i t h , S. (1967). The antecedents of self-esteem. San Francisco, C A : Freeman.
Epstein, N . , Baldwin, L . , & B i s h o p , D . (1983). T h e McMaster Family Assessment Device. Journal

of Marital and Family Therapy, 9, 171-80.
G i l l i g a n , C . (1982). In a different voice: Psychological theory and women’s development. C a m b r i d g e ,

M A : H a r v a r d University Press.
. (1990). Teaching Shakespeare’s sister: notes from the u n d e r g r o u n d o f female adoles-

cence. In C . G i l l i g a n , N . Lyons, & T . H a m n e r (Eds.) Making connections: The relational worlds of
adolescent girls at Emma Willard School (pp. 6-29). C a m b r i d g e , M A : H a r v a r d University Press.

Hauser, S. (1971). Black and white identity formation. New York: Wiley.
K a p l a n , H . (1980). Self-attitudes and deviant behavior. Santa M o n i c a , C A : Goodyear.
K a p l a n , H . , & P o k o m y , A . (1969). Self-derogation a n d psychosocial adjustment. Journal of

Nervous and Mental Disease, 149, 421-34.
. (1976). Self-derogation a n d suicide. Social Sdence and Medicine, 10, 113-21.

Lazure, C , & Persinger, M . (1992). Right hemishpericity a n d low self-esteem i n h i g h school
students: a r e p l i c a t i o n . Perceptual and Motor Skills, 75, 1058.

L e u n g , K . , & L a u , S. (1989). Effects o f self-concept a n d perceived disapproval of

Research Proposal

Independent variable 1: Sex

Independent variable 2: anxiety

Dependent variable: Stress

Question #1

My first independent variable (sex) and my dependent variable (stress) are related. Men and

women tend to experience stress differently. Similarly, men and women react differently to

stress.

I expect women to score higher than men on the dependent variable. Women suffer more stress

compared to men. A 2010 study discovered that women are more likely to experience an

increase in stress levels as compared to men. Women are also more likely to report emotional

and physical symptoms of stress compared to men (APA, 2012). The stress gap between men

and women is because their stress response is different. Women have a different hormonal

system that usually causes them to react more emotionally and become more fatigued.

Similarly, women are exposed to more stress-related factors since they assume several roles in

their daily life.

Question #2

My second independent variable (anxiety) is related to my dependent variable (stress). Anxiety

and stress can both cause severe physical and mental health issues, such as depression, muscle

tension, substance abuse, personality disorders, and insomia (Powell & Enright, 2015). Both are

emotions and normal responses that can become disruptive and overwhelming to day-to-day

life. They can interfere with important aspects of life, such as work, relationships,

responsibilities, and school.

An increase in anxiety can increase stress levels. Research indicates that excessive anxiety can

lead to stress-related symptoms such as difficulty concentrating, insomnia, irritability, muscle

tension, and fatigue. Individuals can manage their anxiety and stress with relaxation techniques.

This includes breathing exercises, yoga, physical activity, art therapy, meditation, and massage.

References

APA. (2012). 2010 Stress in America: Gender and Stress. Retrieved from:

https://www.apa.org/news/press/releases/stress/2010/gender-stress

Powell, T., & Enright, S. (2015). Anxiety and stress management. Routledge.

The Depression Anxiety Stress Scales—21 (DASS-21): Further
Examination of Dimensions, Scale Reliability, and Correlates

Augustine Osman,1 Jane L. Wong,2 Courtney L. Bagge,3 Stacey Freedenthal,4
Peter M. Gutierrez,5 and Gregorio Lozano6

1 The University of Texas at San Antonio
2 Armstrong Atlantic State University
3 University of Mississippi Medical Center
4 University of Denver
5 Denver VA Medical Center, MIRECC
6 The University of Texas at San Antonio

Objectives: We conducted two studies to examine the dimensions, internal consistency reliability
estimates, and potential correlates of the Depression Anxiety Stress Scales—21 (DASS-21; Lovibond
& Lovibond, 1995). Method: Participants in Study 1 included 887 undergraduate students (363
men and 524 women, aged 18 to 35 years; mean [M] age = 19.46, standard deviation [SD] = 2.17)
recruited from two public universities to assess the specificity of the individual DASS-21 items and to
evaluate estimates of internal consistency reliability. Participants in a follow-up study (Study 2) included
410 students (168 men and 242 women, aged 18 to 47 years; M age = 19.65, SD = 2.88) recruited
from the same universities to further assess factorial validity and to evaluate potential correlates of
the original DASS-21 total and scale scores. Results: Item bifactor and confirmatory factor analyses
revealed that a general factor accounted for the greatest proportion of common variance in the DASS-21
item scores (Study 1). In Study 2, the fit statistics showed good fit for the bifactor model. In addition,
the DASS-21 total scale score correlated more highly with scores on a measure of mixed depression
and anxiety than with scores on the proposed specific scales of depression or anxiety. Coefficient
omega estimates for the DASS-21 scale scores were good. Conclusions: Further investigations
of the bifactor structure and psychometric properties of the DASS-21, specifically its incremental and
discriminant validity, using known clinical groups are needed. C© 2012 Wiley Periodicals, Inc. J. Clin.
Psychol. 68:1322–1338, 2012.

Keywords: depression anxiety stress; self-report inventory; bifactor IRT models; nonclinical

Theorists have posited that a common factor may underline depression and anxiety (cf. Clark
& Watson, 1991), accounting for both observations of high comorbidity between the mood and
anxiety disorders (e.g., Sanderson, Di Nardo, Rapee, & Barlow, 1990; Seligman & Ollendick,
1998) and the lack of specificity of measures designed to assess specific disorders (e.g., depression;
Clark & Watson, 1990). Clinicians have also suggested that a new diagnosis be created which
combines elements of both disorders into a single mixed anxiety-depressive disorder. Lovibond
and Lovibond (1995) developed the original Depression Anxiety Stress Scales-42 (DASS-42)
to maximize discrimination between self-reported anxiety and depression while assessing the
full range of these disorders’ core symptoms. Antony, Bieling, Cox, Enns, and Swinson (1998)
subsequently confirmed that both the original DASS-42 and a shorter version, the DASS-21,
distinguish “well between features of depression, physical arousal, and psychological tension
and agitation” (p. 176) in clinical and nonclinical groups.

The DASS is not intended to be used as a diagnostic instrument as the three subscales as-
sess dimensional components of the anxiety and depressive disorders (Psychology Foundation
of Australia, 2011). Additionally, it is not only a distress measure, but rather a measure of

Please address correspondence to: Augustine Osman, The University of Texas at San Antonio, Department
of Psychology, One UTSA Circle, San Antonio, Texas 78249-0652. E-mail: augustine.osman@utsa.edu

JOURNAL OF CLINICAL PSYCHOLOGY, Vol. 68(12), 1322–1338 (2012) C© 2012 Wiley Periodicals, Inc.
Published online in Wiley Online Library (wileyonlinelibrary.com/journal/jclp). DOI: 10.1002/jclp.21908

Depression Anxiety Stress 1323

the shared causes across anxiety, depression, and stress characterized as a distinct syndrome.
Because the measure is not diagnosis-specific, it is appropriate for use in clinical populations
broadly speaking. This makes it a measure appropriate in a wide range of clinical and research
settings where measures of the interplay of different forms of emotional distress are needed.
However, it is probably not appropriate as a measure of individuals’ state distress as the time-
frame assessed is more than the current moment (Psychology Foundation of Australia). It may
perform similarly to other existing measures such as the Hospital Anxiety and Depression
Scale (HADS; Zigmond & Snaith, 1983), as the two yielded similar results in a study of for-
mer intensive care unit patients (Sukantarat, Williamson, & Brett, 2007). However, Sukantarat
et al. cited the DASS’ ability to assess the additional dimension of stress not covered by the
HADS and superior internal consistency reliability as two advantages of the DASS-21 over the
HADS.

The DASS-21 has been used in a range of settings across age groups. For example, Page,
Hooke, and Morrison (2007) evaluated the psychometric properties of the DASS-21 in three
samples of mostly adult psychiatric patients with mood disorders, comprising over 1,400 patients
between the ages of 14 and 83. Their primary findings support the strong internal consistency
of the total and scale scores, the factor structure as determined by Crawford and Henry (i.e.,
allowing three items to cross-load and correlated error terms; 2003), and sensitivity to change
resulting from treatment. The authors also report a significant ceiling effect for the depression
subscale, less of a ceiling effect (although still notable) for the anxiety subscale, and no ceiling
effect for the stress scale. They were unable to eliminate the ceiling effect by adjusting the
response options, and they suggest that the revision of the items to better capture the most
severe symptoms of depression is in order.

The DASS-21 was found to be an appropriate measure to assess improvement as a result of
treatment for patients admitted to a private psychiatric hospital in Australia (Ng et al., 2007).
In this sample of 388, majority female (61%), adult (mean age = 52) patients, the measure was
found to be easy to administer, low cost, and effective at detecting change in a sample of patients
diagnosed primarily with depressive and anxiety disorders. Sukantarat et al. (2007) evaluated
the DASS for use with 51 adult patients (mean age = 57 years) who survived intensive care unit
admissions. Internal consistency reliability for the DASS was acceptable at both 3-month and 9-
month follow-up and scale scores correlated at acceptable levels with another validated measure
of depression and anxiety. Finally, Allen and Annells (2009) concluded after conducting a critical
review of the literature that one advantage of using the DASS-21 to assess elderly patients is
that the lack of items on somatic complaints allows for more accurate assessment in this age
group.

The current studies build on previous investigations with the DASS-21 by examining the
factor structure, scale reliability, and potential correlates of the scores in two large undergraduate
student samples. The DASS items are scored on a 4-point scale ranging from 0 (did not apply
to me at all) to 3 (applied to me very much, or most of the time). Higher scores indicate more
frequent symptomatology. Seven items comprise each of three scales: Depression (example item:
“I couldn’t seem to experience any positive feeling at all”), Anxiety (e.g., “I experienced breathing
difficulty”), and Stress (e.g., “I found it hard to wind down”).

Factor Structure

Previous studies investigating the factor structure of the DASS-21 have identified a two-factor
solution, a three-factor solution, or a second-order factor defined by three lower first-order
factor solutions. However, several modifications were made to each of these models to at-
tain good fit estimates. As an example, in studies confirming support for a two-factor solu-
tion, researchers grouped together the DASS-21 Stress and Anxiety scale items into a general
anxiety-stress factor to help improve the fit of the model (Daza, Novy, & Stanley, 2002; Duffy,
Cunningham, & Moore, 2005; Tully, Zajac, & Venning, 2009). In studies involving a three-
factor solution, researchers have allowed for (a) multiple correlated errors within domain-specific
scales or (b) cross loading of items to attain best-fitting models (e.g., see Antony et al., 1998;

1324 Journal of Clinical Psychology, December 2012

Brown, Chorpita, Korotitsch, & Barlow, 1997; Clara, Cox, & Enns, 2001; Henry & Crawford,
2005).1

To date, only one study has attempted to address empirically the unique contributions of the
individual DASS-21 items to the depression, anxiety, and stress constructs. Using a confirma-
tory factor-analytic approach, Henry and Crawford (2005) reported support for a four-factor
(i.e., quadripartite) model in samples of nonclinical adults in the United Kingdom. Specifically,
unlike previous studies with the DASS-21 items, the investigators constrained each item to load
on a general distress dimension, and also on only one of the three domain-specific dimen-
sions. However, the researchers allowed for several correlated errors between items within the
domain-specific factors (i.e., the Anxiety and Stress scale items) to help establish good fit of
the quadripartite model and did not discuss how to derive scores for the DASS-21 scales that
included correlated errors (see Smith, McCarthy, & Anderson, 2000). Importantly, by relying
on model fit statistics to evaluate the adequacy of the study models, the authors did not evalu-
ate empirically the extent to which each individual DASS-21 item is related more strongly to a
proposed domain-specific dimension (i.e., specificity) than to a general distress (nonspecificity)
dimension (see Cook, Kallen, & Amtmann, 2009; Haynes, Richard, & Kubany, 1995).

Consequently, one of the goals of our Study 1 was to examine the relationships between each
individual DASS-21 item and a domain-specific dimension or scale. Specifically, we wanted to
determine the specificity and nonspecificity of the individual DASS-21 items (e.g., see Reise,
Morizot, & Hays, 2007; Simms, Grös, Watson, & O’Hara, 2008). As an example, if the DASS-
21 items are linked strongly with both the general and domain-specific dimensions, the items
should be conceptualized as multidimensional (i.e., nonspecific to the underlying dimension),
rather than specific, in nature. Accordingly, we conducted several item-level bifactor analyses to
address the current study goals. First, however, we conducted an exploratory Schmid-Leiman
(1957) transformation analysis to examine the (a) dimensions and (b) amount of common
variance accounted for by the DASS-21 scale scores. This approach is exploratory because the
structural dimensions of the DASS-21 items have not been examined in a large sample (e.g.,
N ≥ 500) of U.S. nonclinical college-age samples.

Scale Reliability

Beyond investigations of factor structure or dimensionality, studies have reported good estimates
of internal consistency reliability for the original scale scores (range = .82 to .97) of the DASS-21
in clinical and nonclinical samples (e.g., Henry & Crawford, 2005; Lovibond & Lovibond, 1995).
Because the coefficient-α estimation procedure has been observed to underestimate or overesti-
mate internal consistency reliability for multidimensional instruments (e.g., between-item), we
also employed an alternate internal consistency reliability estimation method, the McDonald
coefficient-omega (coefficient-ω; McDonald, 1999), to calculate internal consistency reliability
for scores on the DASS-21 scales (see Raykov, 2004; Sijtsma, 2009; Zumbo, Gadermann, &
Zeisser, 2007).

Confirmatory Analyses and Concurrent Validity

In Study 2, we used confirmatory factor analysis to evaluate fit estimates of the bifactor model
against two competing solutions that have been reported in the extant literature for the DASS-21
(e.g., Henry & Crawford, 2005; Tully et al., 2009). Specifically, unlike Study 1, the goal of this
study was to extend our understanding of the structure and nature of the relationships among
the proposed dimensions of the DASS-21. Accordingly, we hypothesized that the bifactor model
(i.e., a model comprising a general distress dimension plus three domain-specific dimensions)
would attain a better fit to the sample data than any of the alternative models (in particular, the
original oblique three-factor solution).

1Although some researchers tend to remove the DASS-21 Stress scale items from validation studies, we did
not modify the DASS-21 items given that the goal of the current study was not to revise the instrument (see
Clara et al., 2001; Smith & McCarthy, 1995).

Depression Anxiety Stress 1325

In Study 2, we also examined potential correlates of the DASS-21. The concurrent validity of
the DASS-21 scale scores has been supported by moderate to high correlations (rs = .40 to .65)
with related measures of depression and anxiety (e.g., Antony et al., 1998; Brown et al., 1997;
Crawford & Henry, 2003). Also, previous studies have demonstrated some evidence of discrim-
inant validity by showing that between-group specific correlations (e.g., DASS-21 depression
with a number of self-report measures of anxiety) were significantly lower than within-group
specific correlations (e.g., DASS-21 depression with other related measures of depression) among
clinical and nonclinical samples (e.g., Antony et al., 1998; Henry & Crawford, 2005; Gloster
et al., 2008). In this study, we further examined estimates of concurrent validity for scores on
the DASS-21.

Study 1

First, we conducted an exploratory factor analysis to evaluate the amount of common variance
in the DASS-21 item scores. Second, because our focus was on evaluating the characteristics
of the individual DASS-21 items, we conducted an item bifactor analysis. Third, we conducted
estimates of internal consistency reliability for the original DASS-21 total and subscale scores.
We also computed descriptive statistics (mean and standard deviation) for the DASS-21 total
and scales. Only participants with complete data were included in the analyses.2

Method

Participants

The sample included undergraduate students from large midwestern and southwestern public
universities. Because participants from the separate universities did not differ significantly on
relevant demographic variables, including age, gender, and marital status (all ps > .05), we
combined the data for all the subsequent analyses. The sample included 524 women (mean [M]
age = 19.38, standard deviation [SD] = 2.32 years) and 363 men (M age = 19.59, SD = 1.93
years); they did not differ significantly in age, t (885) = 1.42, p = .16. Of the sample, 652 (73.5%)
self-identified as Caucasian/White, 96 (10.8%) Hispanic American, 80 (9.0%) African American,
53 (6.0%) Asian American, and 6 (0.7%) as “other ethnic racial groups.” The majority of the
participants were single, never married (n = 785, 88.5%), 19 (2.1%) were married, 60 (6.8%)
were engaged, seven (0.8%) were separated, four (0.5%) were divorced, and 12 (1.3%) reported
“live-in” partner. The majority were freshmen (n = 705, 79.5%), 100 (11.3%) were sophomores,
52 (5.9%) were juniors, and 30 (3.3%) were seniors.

Measures and Procedure

Consistent with approvals from the institutional review boards, all the study participants volun-
tarily completed written informed consent forms, a brief demographic questionnaire (assessing
gender, age, ethnicity, year in college, and marital status), and the Depression Anxiety Stress
Scales—21. senior research undergraduates, who were trained and supervised by the first au-
thor, administered all questionnaire packets. All participants received partial course credit for
completing the questionnaire packet.

Results

Exploratory Factor (Schmid-Leiman) Analysis

To examine (a) the dimensions of the DASS-21 and (b) the amount of common variance ac-
counted for by the specific and general distress factors, we conducted an exploratory item

2To address concerns regarding missing data, responses to all individual questionnaire items were reviewed
for completeness at the time of data collection.

1326 Journal of Clinical Psychology, December 2012

Table 1
Exploratory Factor Analysis and Schmid-Leiman Analysis of the DASS-21 Items

Oblique three-factor solution Schmid-Leiman bifactor

Item F1 F2 F3 GF F1 F2 F3

Factor 1 (F1; Depression)
3. no positive feeling .74 −.03 .05 .58 .49 −.02 .03
5. no initiatives .32 .07 .24 .50 .21 .05 .13
10. had nothing .76 −.03 .01 .56 .51 −.02 .01
13. down-hearted .58 −.10 .15 .49 .38 −.07 .08
16. unable about things .60 −.03 .10 .52 .40 −.02 .06
17. wasn’t worth much .82 .02 −.12 .53 .54 .01 −.07
21. life is meaningless .79 .03 −.18 .47 .52 .02 −.10
Factor 2 (F2; Anxiety)
2. dryness of mouth −.09 .57 −.04 .32 −.06 .39 −.02
4. breathing problem −.06 .80 −.11 .45 −.04 .54 −.06
7. trembling .04 .57 .09 .52 .02 .39 .05
9. worried .11 .35 .27 .57 .08 .24 .15
15. close to panic .15 .49 .23 .66 .10 .33 .13
19. action of my heart −.01 .71 −.12 .41 −.01 .49 −.07
20. felt scared .06 .45 .23 .57 .04 .31 .13

Factor 3 (F3; Stress)
1. hard to wind down −.16 −.03 .83 .55 −.11 −.02 .46
6. over-react .11. −.10 .69 .59 .08 −.07 .38
8. nervous energy −.10 .05 .80 .63 −.07 .04 .44
11. getting agitated .06 −.03 .71 .62 .04 −.02 .39
12. difficult to relax −.08 −.03 .85 .63 −.05 −.02 .47
14. intolerant .07 .10 .53 .57 .04 .07 .29
18. rather touchy .16 −.11 .66 .59 .11 −.07 .36

Note. DASS-21 = Depression Anxiety Stress Scale – 21; GF = general distress factor.
Factor loadings ≥ .40 are set in bold.

bifactor analysis. The analysis was also designed to help provide additional information for de-
riving scores on the DASS-21. As in most exploratory factor analyses with the Schmid-Leiman
transformation, each item was allowed to load on a general distress factor and on one or more
of the domain-specific factors. A higher order structure would be indicated to the extent that the
general distress factor accounts for at least 20% of the composite variance (see Reckase, 1979).

Parallel analysis (95th percentile of random eigenvalues) of the matrix indicated that a three-
factor solution could be extracted. Results of the principal axis factoring (PAF) analysis with
promax rotation, followed by the Schmid-Leiman transformation are presented in Table 1. The
first five eigenvalues from the preliminary analyses were as follows: 8.06, 1.85, 1.67, 1.11, and
0.91. Regarding the three-factor oblique solution (columns 2–4), most of the items had good
loadings (i.e., values ≥ .40) on one of the domain-specific factors. However, the factor inter-
correlations were high: stress versus depression (.63), stress versus anxiety (.61), and depression
versus anxiety (.55).

Regarding results of the Schmid-Leiman transformation (columns 5–8), we found that the
general distress factor accounted for the greatest proportion of the common variance among
the DASS-21 item scores, 61.9%. The DASS-21 Depression factor (14.7%), the DASS-21 Stress
factor (12.3%), and the DASS-21 Anxiety factor (11.1%) accounted for smaller amounts of
the variance among the DASS-21 items. Using item-factor loadings ≥ .40 as salient, 20 of
the 21 (95.2%) DASS-21 items loaded more substantially on the general distress factor than
on the domain-specific factors that were identified in the exploratory factor analysis (EFA).
Further analyses showed that the domain-specific DASS-21 Stress factor correlated with the
general distress factor at .84. The domain-specific DASS-21 Depression factor correlated with

Depression Anxiety Stress 1327

the general distress factor at .75, and the domain-specific DASS-21 Anxiety factor correlated
with the general distress factor at .73. The obtained (hierarchical) coefficient-ωH for the general
distress factor of .87 was also high, suggesting that 87% of the variance for the composite
DASS-21 score was due to the variance on the general distress factor.

Confirmatory Item Bifactor Analysis: Analysis of Item Responses

We used the item response theory (IRT) for Patient-Reported Outcomes (IRTPRO) 2.1 statistical
program (Cai, du Toit, & Thissen, 2012) to conduct the analysis.3 As an IRT modeling program,
IRTPRO 2.1 also allows for the evaluation of several characteristics of an item such as (a) the
degree to which an individual item is specific or related to a specified construct (i.e., a slope
parameter) and (b) the intercept of the item (i.e., c parameter). In the current study, however,
we addressed the specific aims by focusing on (a) the magnitudes of the item-factor/dimension
loadings and (b) the slope parameters (a-parameters). We considered items with moderate to
high standardized loadings (i.e., values ≥ .40) and high slopes (values ≥ 1.50) to be strongly
linked with (i.e., most discriminating) the specified dimension (see also, Kim & Pilknois, 1999;
Reise & Waller, 1990).

In the analysis involving the item-bifactor modeling, we constrained each DASS-21 item to
load onto a general distress factor and on only a domain-specific factor (see Table 2). The
three domain-specific factors or dimensions were orthogonal to each other and to the general
distress factor. As an example, for the domain-specific DASS-21 Depression dimension items,
we constrained items 3, 5, 10, 13, 16, 17, and 21 to load on the DASS-21 Depression dimension
(specificity) and the general distress factor shared by all the DASS-21 items, including the
domain-specific DASS-21 Anxiety and DASS-21 Stress scale items (nonspecificity).

We used the Bock and Aitkin (BAEM; 1981) algorithm in IRTPRO 2.0 to conduct the analysis.
Convergence criterion was attained with the maximum number of E-steps set at 500, and the
quadrature (Q = 21) points were between -6.0 and 6.0. The result of the item bifactor analysis is
presented in Table 2. Of note is that for multidimensional instruments such as the DASS-21, the
threshold (b) parameters do not yield substantive interpretations. Thus, these parameters were
not discussed further. As shown in the table, items 3, 10, 17, and 21 had the highest slopes and
standardized loadings on the domain-specific factor, and are therefore most associated with the
DASS-21 Depression dimension. However, these items were also strongly associated with the
general distress (mixed) factor. Item 5 was least associated with the domain-specific DASS-21
Depression dimension; this item also had the lowest loading on this dimension. Taken together,
item 21 provided the most information about the DASS-21 Depression dimension.

Examination of the slope parameters for the DASS-21 Anxiety dimension showed that items
4 and 19 were most strongly associated with the domain-specific Anxiety dimension. Items 7,
9, 15, and 20 were more strongly associated with the general distress dimension than with the
domain-specific anxiety dimension. In particular, items 9 and 20 were least associated with
the domain-specific DASS-21 Anxiety dimension; these items also had the lowest loadings on
the domain-specific dimension. Item 4 provided the most important information about the
domain-specific anxiety dimension. For the domain-specific DASS-21 Stress dimension, we
found that items 1 and 12 were most strongly associated with this dimension. Each DASS-
21 Stress dimension item, however, was associated strongly with the general distress (mixed)
factor. Items 6, 11, 14, and 18 were least associated with the domain-specific DASS-21 Stress
dimension. In general, most of the Anxiety and Stress subscale items were strongly associated
with the general distress dimension.

3We obtained similar results when we used the Metropolis Robbins-Monro (MH-RM) algorithm in IRTPRO.
This is a new computer program that is designed to estimate unidimensional and multidimensional item
response models (and related statistics). The program manual provides comprehensive information regarding
several item parameter estimation methods (e.g., the Bock-Aitkin, 1981 for bifactor modeling), test statistics
(e.g., RMSEA), and scale score computation strategies.

1328 Journal of Clinical Psychology, December 2012

Table 2
Bifactor (Confirmatory) Analyses of the DASS-21 Items (N = 887)

Standardized factor loading (slope parameter estimates)

General Depression Anxiety Stress
Item Abbreviation factor (a1) factor (a2) factor (a3) factor (a4)

Factor 1 (Depression)
3. no positive feeling .70 (2.39) .51 (1.75)
5. no initiatives .64 (1.47) .19 (0.42)
10. had nothing .68 (2.44) .57 (2.05)
13. down-hearted .63 (1.70) .46 (1.23)
16. unable about things .63 (1.70) .45 (1.20)
17. wasn’t worth much .63 (2.16) .60 (2.06)
21. life is meaningless .65 (2.51) .62 (2.38)

Factor 2 (Anxiety)
2. dryness of mouth .32 (0.72) .56 (1.23)
4. breathing problem .51 (1.66) .69 (2.26)
7. trembling .63 (1.74) .47 (1.28)
9. worried .73 (1.87) .19 (0.50)
15. close to panic .83 (3.04) .30 (1.09)
19. action of my heart .46 (1.24) .62 (1.68)
20. felt scared .76 (2.20) .30 (0.87)

Factor 3 (Stress)
1. hard to wind down .59 (2.71) .71 (3.26)
6. over-react .80 (2.28) .12 (0.34)
8. nervous energy .75 (2.50) .43 (1.44)
11. getting agitated .81 (2.45) .17 (0.52)
12. difficult to relax .68 (3.45) .66 (3.37)
14. intolerant .77 (2.04) .08 (0.21)
18. rather touchy .81 (2.36) .07 (0.19)

Note. DASS-21 = Depression Anxiety Stress Scales.
Estimates ≥ .40 are set in bold. Slope (a1-a4) parameter estimates of the latent dimensions are in parenthesis.

Scale Reliability Analysis and Descriptive Statistics (Means and Standard Deviations)

We interpreted reliability estimates of .70 to .79 as moderate, and estimates of .80 or greater
as high for use of the scale scores in research settings (see Clark & Watson, 1995; Cicchetti,
1994). For the current study sample, the traditional coefficient-α estimates for the DASS-21
scale scores were as follows: DASS-21 Depression (α = .85; 95% confidence interval [CI], .83-
.87, average inter-item correlation [AIC], .47); DASS-21 Anxiety (α = .81; 95% CI, .79-.84, AIC,
.40); and DASS-21 Stress (α = .88; 95% CI, .87-.89, AIC, .52). The McDonald’s coefficient-ω
for the DASS-21 Depression scale score was .86 (M = 4.18, SD = 3.60). The coefficient-ω for
the DASS-21 Anxiety scale score was .82 (M = 2.93, SD = 3.38), and the coefficient-ω for the
DASS-21 Stress scale score was .88 (M = 5.29, SD = 4.57). As expected, each scale reliability
estimate was ≥ .80.

Study 2

First, we used confirmatory factor analysis to evaluate fit estimates of the bifactor (target)
model against alternative models to the sample data. Second, we used both the coefficient-ω and
the traditional Cronbach (coefficient-α) reliability analytic procedures to re-examine internal
consistency reliability estimates of the DASS-21 scale scores in an independent sample. Third,
based on empirical support obtained for the best-fitting model, we examined potential correlates
for the DASS-21.

Depression Anxiety Stress 1329

Method

Participants and Procedure

Following the analyses in Study 1, we collected data from an independent sample of 410 un-
dergraduate students in Midwestern and Southwestern state universities. The combined sample,
with no statistically significant differences in demographic variables, comprised 242 women (M
age = 19.95, SD = 3.15; range = 18–47 years) and 168 men (M age = 19.45, SD = 2.66; range =
18–39 years). The mean age of the sample was 19.65 years (SD = 2.88; range = 18–47); 259
(63.2%) self-reported as Caucasian/White, 90 (21.9%) as Hispanic/Latino American, 32 (7.8%)
African American, 24 (5.9%) Asian American, and 5 (1.2%) indicated “other” ethnicities. The
total sample comprised 305 (74.4%) freshmen, 52 (12.7%) sophomores, 33 (8.0%) juniors, and
20 (4.9%) seniors. After obtaining informed consent, participants completed the questionnaire
packets in small groups for partial course credit.

Confirmatory factor analytic (CFA) sample. All the study participants (N = 410)
provided data for inclusion in the descriptive statistics (mean and standard deviation) and CFA.

Concurrent validation subsample4. We included concurrent validation self-report ques-
tionnaires in 223 (149 women and 74 men, aged 18–25 years) of the 410 (approximately 54%)
study packets (see the Measures section). These measures have good psychometric properties
for assessing dimensions of constructs that are related to those evaluated with the DASS-21:

� Anxiety: Beck Anxiety Inventory total score (Beck & Steer, 1990), the Mood and Anxiety
Symptom Questionnaire-90 (MASQ-90; Watson, Clark et al., 1995), and Anxious Arousal
scale score.

� Depression: Beck Depression Inventory-II total score (BDI-II; Beck, Steer, & Brown, 1996)
and the MASQ-90 Anhedonic Depression scale score.

� Perceived stress: The Perceived Stress Scale total score (PSS; Cohen, Kamarck, & Mermel-
stein, 1983).

Measures

Besides a brief background questionnaire and the DASS-21, all participants completed the
Mood and Anxiety Symptom Questionnaire-90 (MASQ-90; Watson, Clark et al., 1995). The
subsample (n = 223) completed the following concurrent validation self-report measures.

The Mood and Anxiety Symptom Questionnaire-90 (MASQ-90; Watson, Clark
et al., 1995). We used the MASQ-90 as a concurrent validity measure of anxiety, depression,
and general distress (mixed depression and anxiety symptoms). Participants rated the instrument
items from 1 (not at all) to 5 (extremely). An example MASQ-90 item is “felt discouraged.” The
MASQ-90 has been shown to have adequate estimates of reliability and concurrent validity (see
Keogh & Reidy, 2000; Watson, Clark et al., 1995; Watson, Weber et al., 1995). As an example,
in the concurrent validation subsample (n = 223), estimates of internal consistency for scores on
the anhedonic depression (low positive affect; coefficient-α for 14 items = .94, 95% CI, .93-.95;
AIC, .53), anxiety (anxious arousal, coefficient-α for 17 items = .87, 95% CI, .81-.90; AIC,
.28), general distress depression (coefficient-α for 12 items = .91, 95% CI, .89-.93; AIC, .46),
and general distress anxiety (coefficient-α for 11 items = .81, 95% CI, .76-.85; AIC, .28) were
good.

4To ensure that participants in Study 2 did not participate in the Study 1 data collection process, we included
a checklist assessing familiarity with studies conducted in our laboratory during the Study 1 data collection
phase. The measures were also included randomly in each packet to control for order effects.

1330 Journal of Clinical Psychology, December 2012

The Perceived Stress Scale (PSS; Cohen, Kamarck, & Mermelstein, 1983)5. We
used scores on the PSS, a 10-item self-report instrument, to assess perceptions of the individual
to stress (negative affectivity) “in the past month.” Each PSS item is scored on a 5-point scale
ranging from 0 (never) to 4 (very often). Four of the items are reverse-scored and the ratings
summed on all the items to obtain a total scale score. Example PSS items are, “felt like you were
on top of things” and “felt nervous and stressed.” Satisfactory estimates of internal consistency
and concurrent validity have been established for scores on the PSS (see Hughes, 2007; Kohn
& Gurevich, 1993). Cronbach coefficient-α estimate of the PSS total scale score for the current
subsample (n = 223) was satisfactory at .87 (95% CI, .84-.89; AIC, .40). It is important to
note that scores on stress-related measures have been useful in assessing negative affectivity (see
Lucas, Michalopoulou, Falzarano, Menon, & Cunningham, 2008; Shewchuk, Elliott, MacNair-
Semands, & Harkin

Journal of Counseling & Development ■ Winter 2006 ■ Volume 84 61

Efforts to evaluate the tenets of attachment theory have con-
tributed to a growing body of research documenting the con-
tributions of the parent–child relationship to emotional well-
being and social competence across the life span. Although
attachment research focused initially on the observation of
mother–child attachment relationships in early childhood,
Bowlby (1982) maintained that attachment processes were
central to personality functioning from “cradle to grave”
(p. 172). Over the past decade, attachment researchers have
increased their attention to articulating the distinct qualities
of adult–child and adult–adult attachment and delineating
the cognitive and affective processes underlying attachment
across the life span (Crowell & Treboux, 1995).

One way in which attachment relationships are theorized
to affect well-being across the life span is by providing a
secure base of support. Through their availability as a source
of support, secure attachments can reduce anxiety, increase
environmental exploration, and contribute to competence in
interacting with the world (Kobak & Sceery, 1988). Among
young children, for example, the responsive and sensitive
caretaker is believed to contribute to child feelings of secu-
rity, confidence in exploring the environment, and the devel-
opment of instrumental competence (Bowlby, 1988; Kobak
& Sceery, 1988). For the late adolescent leaving home for
college, Kenny (1987) suggested that secure parental attach-
ments provide a secure base by supporting student explora-
tion and mastery of the college environment and by remain-
ing available as a source of advice and comfort when needed.
Main (1999) proposed that proximity seeking or secure base
behavior often increases in the later stages of the life span

when older adults are less able to care for and protect them-
selves and experience heightened feelings of vulnerability.

Beyond the role as a source of actual assistance, parental
attachments are also theorized to exert an enduring influence
on development through the formation of internal working
models. Sensitive and consistently available caretaking may
contribute to an internal working model of self as worthy of
love and a model of others as trustworthy and predictable.
Conversely, insensitive and unreliable caretaking may result
in a view of self as unworthy and a view of others as untrust-
worthy (Bowlby, 1982). These internal working models of self
and others are believed to serve as cognitive filters through
which current experiences are interpreted and ongoing ex-
pectations of self and others are formulated (Bowlby, 1982;
Bretherton, 1985, 1992). Positive internal working models
are theorized to enhance an individual’s ability to adapt to
stress over time, because the individual has confidence in the
self and trust in reaching out to others for help. Conversely,
and consistent with Beck’s (1967) model of depression, a nega-
tive view of self and others, associated with insecure attach-
ment, may increase vulnerability to depression (Carnelley,
Pietromonaco, & Jaffe, 1994; Kenny, Moilanen, Lomax, &
Brabeck, 1993; Kernis, Grannemann, & Mathis, 1991).

Characteristics of secure parental attachment have been
associated empirically with indices of adaptive social and
psychological functioning across a variety of developmen-
tal periods (Kenny & Barton, 2002). For the adolescent, se-
cure parental attachments have been conceptualized as pro-
viding a source of security and support as the adolescent
negotiates the numerous transitions and challenges of this

Maureen E. Kenny and Selcuk R. Sirin, Department of Counseling, Developmental, and Educational Psychology, Boston
College. Selcuk R. Sirin is now in the Department of Applied Psychological Studies, The Steinhardt School of Education, New York
University. The authors wish to thank Penny Hauser-Cram, Martha Bronson, Judith Palfry, Matthew Jans, Janet Morganelli,
Lauren Rogers-Sirin, and Ted Wall for their contributions in the development of the project and Penny Hauser-Cram for helpful
comments on drafts of the manuscript. This research was supported in part by the Robert Wood Johnson Foundation, which
provided partial funding of a larger study from which these data were drawn. The data presented, the statements made, and the
views expressed are solely the responsibility of the authors. Correspondence concerning this article should be addressed to
Maureen E. Kenny, Department of Counseling, Developmental, and Educational Psychology, Campion Hall 307, Boston College,
Chestnut Hill, MA 02467 (e-mail: kennym@bc.edu).

Parental Attachment, Self-Worth,
and Depressive Symptoms
Among Emerging Adults
Maureen E. Kenny and Selcuk R. Sirin

The characteristics of parental attachment were assessed for a sample of 81 emerging adults (ages 22–28 years)
and their mothers. Emerging adults’ reports of self-worth were found to mediate the relationship between their
reports of parental attachment and depressive symptoms. The emerging adults’ unique perspectives of the attach-
ment relationship were more predictive of their self-reported self-worth and depressive symptoms than were the
mothers’ perspectives of the relationship.

© 2006 by the American Counseling Association. All rights reserved, pp. 61–71.

Journal of Counseling & Development ■ Winter 2006 ■ Volume 8462

Kenny & Sirin

developmental period. Among early and middle adolescents,
secure parental attachments have been found to buffer life
stress and to be associated with positive self-worth and low
levels of depressive symptoms (Armsden, McCauley,
Greenberg, Burke, & Mitchell, 1990; Kenny et al., 1993;
Kobak, Sudler, & Gamble, 1991; Papini & Roggman, 1992).
Among college students, secure parental attachments have
been positively associated with college adjustment (Larose
& Boivin, 1997; Rice, FitzGerald, Whaley, & Gibbs, 1995),
assertiveness in social relationships (Kenny, 1987), enhanced
resources for coping with stress (Brack, Gay, & Matheny,
1993), and career exploration and commitment (Blustein,
Walbridge, Friedlander, & Palladino, 1991).

Despite the substantive body of research generated in
recent years focusing on parental attachments among ado-
lescents and college students, research has not assessed the
role of parental attachments among young people beyond
college age. Arnett (2000a, 200b) defined the years between
late adolescence (ending at age 18) and young adulthood
(beginning at age 30) as encompassing emerging adulthood,
a period of experimentation, instability, and diverse possi-
bilities in love and work. An increase in the years of formal
schooling and a delay in the age of marriage and parent-
hood, which are common in industrialized societies, have
contributed to extended personal and career exploration,
with enduring commitments to career and relationships of-
ten being postponed until the late 20s (Arnett, 2000a). Arnett
(1998) noted that although studies of college students are
plentiful, research has largely ignored the decade of the 20s,
despite recent evidence suggesting that the years from 21 to
28 are characterized by considerable stress and engagement
in high-risk behaviors. After college graduation, stress may
accrue as relationships with friends and family change and
opportunities for participation in peer and educational ac-
tivities shift from school to the community and workplace
(Bemporad, Ratey, & Hallowell, 1986). Culturally prescribed
pressures for career choice and entry and financial indepen-
dence also begin to increase (Gould, 1978).

In conjunction with the transitional stressors of the emerg-
ing adult years, feelings of loneliness and depression can arise
(Hallowell, Bemporad, & Ratey, 1989), with the median age of
onset for depression occurring during the 20s (Weiner, 1992).
Approximately 25% of late adolescents and young adults have
experienced a major depressive episode (Hart, Craighead, &
Craighead, 2001), with a 1-year prevalence rate of approxi-
mately 3% occurring across the 18- to-29-year-old age group
(Weissman, Bruce, Leaf, Florio, & Holzer, 1991). This is of par-
ticular concern because the experience of a major depressive
episode during the emerging adult years presents a substantial
risk for the recurrence of depression in later years (Hart et al.,
2001). Although developmental researchers have largely
neglected the decade when individuals are in their 20s, it is
a period of substantive stress and risk for depression. Increased
knowledge of the relevance of parental attachment to psycho-

logical distress during these transitional years could have im-
portance for prevention and intervention.

In contrast with traditional developmental (e.g., Blos,
Erikson) models that emphasize autonomy as the primary
criterion for successful adult status, contemporary research
suggests that separation from the family of origin occurs
most adaptively within the context of a caretaker relation-
ship that is transformed, rather than broken (Allen, Moore, &
Kuperminc, 1997; Hill & Holmbeck, 1986). Although con-
temporary perspectives suggest that continuity in the valu-
ing of parental attachment while gaining in autonomy is a
desirable developmental outcome (Ainsworth, 1989), the
relative importance of parental attachments as a secure base
of support or as a source of internal working models during
the emerging adult period has not been investigated. For the
22- to 29-year-old, the importance of parental attachments
as a source of internal working models may surpass the im-
portance of the attachments as a direct source of emotional
or financial support. A hypothesized adaptive outcome of
secure attachment is the development of effective coping
skills, such that, with increasing maturation, the child be-
comes less directly reliant on the attachment figure as a
source of help (Rice & Cummins, 1996). For example, the
emerging adult, equipped with cognitive maturity, problem-
solving abilities, and social relationships beyond the family,
may not seek direct contact with the attachment figure for
comfort, yet might still experience the ongoing effects of
the relationship through the internal working models. Al-
though research to date has not examined the role of paren-
tal attachments as a source of working models of self among
emerging adults who are beyond traditional college age,
this knowledge may inform intervention. Such knowledge
may have implications, for example, as to whether counse-
lors should focus on current patterns of interaction with par-
ents or on cognitive internal working models. Knowledge of
whether parental attachment is associated with secure base
characteristics, such as frequency of contact and financial
support, may have implications for the ways that parents
can be most helpful in supporting their children during the
often stressful transitional decade of the 20s.

In order to evaluate the role of the attachment figure as a
source of internal working models, researchers (e.g.,
Carnelley et al., 1994; Kenny et al., 1993) have compared
models of direct and mediated effects. Mediators “explain
how external physical events take on internal psychologi-
cal significance” (Baron & Kenny, 1986, p. 1176). Findings
have provided support for the hypothesized role of internal
working models as mediators of parental attachment and the
social and emotional functioning of offspring. In a study of
high school seniors, for example, Davila, Hammen, Burge,
Daley, and Paley (1996) found that self-worth mediated the
relation between attachment and interpersonal problem solv-
ing. Self-worth is conceptually related to the internal work-
ing model, because models of self developed in the context

Journal of Counseling & Development ■ Winter 2006 ■ Volume 84 63

Parental Attachment of Emerging Adults

of secure attachment relationships are theorized to contribute
to feelings of positive self-worth and self-confidence (Bowlby,
1980; Simons, Paternite & Shore, 2001). Among early adoles-
cents, view of self was found to partially mediate the relation
between perceived parental attachment and level of depres-
sive symptoms (Kenny et al., 1993). Although low self-worth
has been conceptualized as both a cause and consequence of
depression, it is not synonymous with depression, which in-
cludes a variety of other features (Harter, 1999). In research
assessing depressive affect among college students, self-worth
and dysfunctional attitudes were identified as mediators of
attachment and dysphoria (Roberts, Gotlib, & Kassel, 1996).

Prior research has also evaluated internal working models by
comparing the perspective of the attachment relationship as re-
ported by the parents and offspring. To the extent that internal
working models are a more powerful determinant of the well-
being of offspring than are current parental interactions, their
perspective is likely to be more important than the parents’ per-
spective in explaining well-being. In one of the few existing
studies to include reports from both parents and offspring, Rice
and Cummins (1996) found that the unique perspectives of the
attachment relationship as held by late adolescent college stu-
dents were more important in predicting their current view of self
than were the unique parental perspectives. The importance of
the late adolescents’ unique perception of the attachment rela-
tionship was interpreted as evidence of the importance of the
internal working model. The attachment ratings provided both
by students and parents in the Rice and Cummins study were,
however, retrospective, based on recollections of the first 16 years
of life, and thus did not assess the current attachment relation-
ship. The relative importance of parents’ and emerging adults’
perceptions of current attachment relationships has not been
examined to date, perhaps because of assumptions that rela-
tionships with family of origin have little importance after the
leaving-home transition as well as because of the challenges in
obtaining data from parents when children are no longer living
at home. Reports by parents, however, could help to identify the
unique importance of parents’ and offsprings’ perceptions of
attachment and help to assess the importance of internal
working models among emerging adults.

The current study seeks to increase knowledge of the char-
acteristics of the parental attachment of emerging adults
and to explore the role of parental attachment as a source of
internal working models during the transitional years of
emerging adulthood. With regard to the role of the attach-
ment figure as a secure base among emerging adults, we
assess whether financial support from parents and frequency
of communication with them are associated with perceptions
of parental attachment and emerging adults’ self-reports of
self-worth and depressive symptoms. If parents continue to
provide a secure base, they may be sought out to provide
emotional and f inancial support when needed and that
support may be positively associated with self-worth and
negatively associated with depressive symptoms. With

regard to internal working models, we next examine self-
worth as an internal factor that may mediate the relationship
between parental attachment and the experience of depressive
symptoms. Consistent with evidence that self-worth, as a com-
ponent of the working model of self, mediates the relationship
between parental attachment and well-being among early and
late adolescents, we expected that self-worth would mediate
the relationship between parental attachment and depressive
symptoms among emerging adults. Finally, by including the
parent perspective, we evaluated the relative importance of
parent and emerging adult perceptions of attachment in ex-
plaining self-esteem and depressive symptoms. To the extent
that working models of self are more salient than direct interac-
tion and support from parents at this age, we expected that
emerging adult perceptions would add to the explanation of
their functioning after accounting for the parent perspective.

Method
Participants

The participants in this study were 81 pairs of emerging
adults and their mothers. The 81 emerging adults (49 men
and 32 women) ranged in age from 22 to 29 years (M = 25.98
years, SD = 1.42). In terms of self-reported ethnicity, 64.2%
were Euro-American, 14.8% were African American, 7.4%
were Asian American, 6.2% were Hispanic/Latino, and 7.4%
reported other ethnic background. In terms of highest level
of education, 1% had completed some high school, 6.5%
had a high school degree, 21.3% had completed some col-
lege or technical school, 2.5% had completed an associate’s
degree, 55% had completed college, 10% had completed a
master’s degree, and 3.8% had an advanced graduate degree
beyond the master’s degree. Forty-six percent of the emerg-
ing adults were currently attending some type of educa-
tional or training program, and 76.5% were employed at
least part-time, with incomes ranging from less than $10,000
to over $60,000 per year. Sixty percent earned $20,000 or
less. Forty percent reported receiving some contributions
from parents toward their living expenses, with 60% report-
ing no financial assistance from parents. Twenty-three per-
cent lived with a parent or stepparent, with the remainder
living apart from their family of origin. Seventy-two percent
of the emerging adult participants were single and not liv-
ing with a partner, 9% were married, 18% were living with a
partner, and 1% were separated from spouse; 11% indicated
that they had children of their own. Twenty (24.7%) emerg-
ing adults reported that their parents were divorced; 13% of
their fathers and 11% of their mother had remarried.

The sample of mothers, composed of the 81 biological
mothers of the emerging adults, had a mean age of 54.09
years (SD = 5.25). Sixty-six percent self-reported as Euro-
American, 12.4% as African American, 6.7% as Hispanic,
3.4% as Asian American, and 11.5% reported other ethnic
background. In terms of achieved education, 4.9% had com-

Journal of Counseling & Development ■ Winter 2006 ■ Volume 8464

Kenny & Sirin

pleted grade school, 3.7% had some high school, 6.2 % had a
high school degree, 7.4% had completed some college or techni-
cal school, 12.3% had an associate degree, 24.7% had com-
pleted college, 33.3% had a master’s degree, and 7.4% had com-
pleted an advanced graduate degree beyond the master’s degree.

Procedure

Volunteers were recruited by a mail solicitation of individu-
als who had attended (but not necessarily graduated from)
high schools in an urban community and an adjacent subur-
ban community from 1990 through 1993 and by responses
to signs posted in local health centers. A total of 182 emerg-
ing adults volunteered and were paid $50 for their participa-
tion. If agreeable to the participant, their parents were con-
tacted by mail and invited to participate, and the parents
also received $50 for their participation. Out of 169 parents
who were contacted, 92 (89 mothers, 3 fathers) returned ques-
tionnaires. The sample presented here consists of the adult–
mother pairs for whom complete data were available. Al-
though we believe research needs to more actively focus on
the perspectives of fathers, we focused only on mothers in
this study because of the small number of fathers who re-
sponded. In order to determine whether the 89 emerging
adults whose mothers completed questionnaires were differ-
ent from the 93 emerging adults in our sample whose mothers
did not complete questionnaires, t tests were completed that
compared the two samples on emerging adult measures of
parental attachment, self-worth, and depressive symptoms, as
well as frequency of communication with mother and father
and mothers’ level of education. No significant differences
were found with the exception of mothers’ level of education,
which was higher for the group with maternal data (t = 2.8,
p < .01), suggesting that the sample with maternal data was
similar to our overall sample in most apparent ways.

Measures

The emerging adults who participated in this study com-
pleted a demographic questionnaire, the General Self-Worth
scale of the Adult Self-Perception Profile (ASPP; Messer &
Harter, 1986), the Center for Epidemiological Studies De-
pression Scale (CES-D; Radloff, 1977), and the Parental At-
tachment Questionnaire (PAQ; Kenny, 1987). Mothers com-
pleted a demographic questionnaire and a parent version of
the PAQ (PAQ-P), as well as the Family Cohesion subscale of
the Family Adaptability and Cohesion Evaluation Scale
(FACES III; Olson, Portner, & Lavee, 1985) and the Parental
Burden Scale (M. Windle, personal communication, Febru-
ary 15, 2002), both of which contributed to the validation of
the PAQ-P. The measures are described below.

The demographic questionnaires elicited background
information concerning the participants. In addition, some
information that was obtained from answers to the demo-
graphic questionnaire was included in the preliminary and
main analyses. For example, emerging adults and their

mothers were asked to indicate the highest level of educa-
tion that they had completed by checking one of eight cat-
egories. For purposes of data analysis, the eight categories
were coded in the following manner: grade school/less than
9th grade = 1, some high school/ less than 12th grade = 2,
high school degree or equivalent = 3, some college or
technical school = 4, 2-year college/associate’s degree = 5,
4-year college degree = 6, master’s degree = 7, advanced
degree (e.g., PhD, MD) = 8. Emerging adults were also asked
to respond to the question, “How often do you usually com-
municate with your mother?” by checking one of the fol-
lowing response options: every day, every week, every month,
a few times a year, once a year, every few years, never. The
responses were coded for purposes of data analysis on a 7-
point, Likert-type scale, ranging from never = 1 to every day
= 7. Emerging adults were asked to check yes or no to the
question, “Did your parents (or the people who raised you)
contribute to your living expenses in the current year?”

The CES-D (Radloff, 1977) is a 20-item instrument that
assesses the frequency of depressive symptomatology during
the previous 7-day period. Response options for each item
range from rarely or none of the time (scored as 0) to most or
all of the time (scored as 3). The reported reliability and valid-
ity of the CES-D are strong (Weissman, Sholomskas, Pottenger,
Prusoff, & Locke, 1977). For a community sample, the CES-D
was found to correlate with diagnosis of depression based
upon a structured clinical interview (Roberts & Vernon, 1983).
In this study, an internal consistency analysis yielded a
Cronbach’s alpha of .91.

The ASPP (Messer & Harter, 1986) was designed to mea-
sure several dimensions of perceived competence as well as
general self-worth. For this study, we used the six-item Gen-
eral Self-Worth scale. Participants respond to each item by
first selecting which of two given descriptions is more like
them and then rating whether that description is “really true
of me” or “sort of true of me.” Each item is scored on a scale
of 1 to 4, with a 1 indicating low self-worth and a 4 indicat-
ing high self-worth. Internal consistency as measured by
Cronbach’s coefficient alpha was reported by Harter (1990)
as ranging across varied samples from .87 to .92. For the
current sample, Cronbach’s alpha was .85.

The PAQ (Kenny, 1987) was designed to assess security
of attachment by adapting Ainsworth, Blehar, Waters, and
Wall’s (1978) conceptualization of attachment for use in a
self-report format with adolescents and young adults. Par-
ticipants are asked to rate their perceptions of their parents,
their relationships with their parents, and their experiences
and feeling about their parents using a 5-point, Likert-type
scale ranging from 1 (not at all) to 5 (very much).

As in previous research (Hinderlie & Kenny, 2002), two
subscales of the PAQ, Affective Quality of Attachment and
Parental Fostering of Autonomy, were combined to yield a
single score based on the strong correlation between the
two subscales and the conceptual association between the

Journal of Counseling & Development ■ Winter 2006 ■ Volume 84 65

Parental Attachment of Emerging Adults

parental-affect and autonomy-facilitating components of at-
tachment (Ainsworth et al., 1978). The .65 correlation found
between the two subscales in this study is consistent with
previous findings (e.g., Hinderlie & Kenny, 2002; Holmbeck
& Wandrei, 1993; Holtzen, Kenny, & Mahalik, 1995). Kenny
(1987) reported test–retest reliability of the PAQ over a
2-week interval of .92 for the measure as a whole. Exami-
nation of the validity of the PAQ in other research (Holmbeck
& Wandrei, 1993; Kenny & Donaldson, 1991) revealed mod-
erate correlations with conceptually relevant subscales of the
Family Environment Scale (Moos & Moos, 1986), the Family
Cohesion subscale of the FACES III (Olson et al., 1985), and
the Parental Scale of the Inventory for Peer and Parental
Attachment (Armsden & Greenberg, 1987). The internal con-
sistency (Cronbach’s alpha) for the current study was .94.

The PAQ-P is a newly developed 16-item scale that mea-
sures parents’ perceptions of their attachment relationship
with one of their children. In developing the parent version,
items that had the highest factor loadings on the Affective
Quality of Attachment and Parental Fostering of Autonomy
subscales of the PAQ were selected and reworded to be an-
swered from the parent’s perspective. Each mother was asked
to describe her relationship with the focus child by rating
each item using a 5-point, Likert-type scale ranging from 1
(not at all) to 5 (very much). Examples of items are “In gen-
eral as a parent, I let my child try out new things and learn on
his/her own,” reflecting the Parental Fostering of Autonomy
subscale, and “During recent visits or time spent together, I
felt relaxed and comfortable with my child,” reflecting the
Affective Quality of Attachment subscale.

Because the PAQ-P is a new measure, we examined its va-
lidity by assessing relationships with two other scales that are
designed to assess relevant aspects of family relationships.
We expected the Family Cohesion scale of the FACES III (Olson
et al., 1985), which assesses emotional bonding and support
among family members, to be positively associated with ma-
ternal ratings of attachment. We expected the Parental Bur-
den Scale (M. Windle, personal communication, February 15,
2002), which assesses the parent’s view of the child as overly
dependent and lacking in self-reliance, to be negatively cor-
related with PAQ-P scores. The PAQ-P is positively correlated
with scores on the Family Cohesion subscale (r = . 34, p < .01)
and negatively correlated with scores on the Parental Burden
Scale (r = .55, p < .001), providing support for the convergent
validity of the PAQ-P as assessing related, but not identical,
constructs as measured by the Family Cohesion subscale and
the Parental Burden Scale. The internal consistency
(Cronbach’s alpha) for the PAQ-P in the current study was .82.

Results
Preliminary Analyses

Table 1 presents the means and standard deviations of the
study variables, and Table 2 presents Pearson correlation

coefficients. Overall, scores on the emerging adult measures
indicate positive perceptions of parental attachment (item
M = 3.93 on a 5-point scale), positive self-worth (item M =
3.28 on a 4-point scale), and, as would be expected for a
community sample, overall levels of depressive symptoms
(M = 11.73) well below the cutoff score of 16, which is in-
dicative of clinical depression. Mothers’ reports of attach-
ment were also highly positive (item M = 4.44).

On the basis of evidence that education is related to
emotional well-being (Johnson, Cohen, Dohrenwend, Link,
& Brook, 1999) and knowledge that the mothers who par-
ticipated in our study were more highly educated than
mothers of emerging adults who did not participate, we
examined the relation between emerging adults’ and moth-
ers’ education levels and study variables. As presented in
Table 2, both emerging adults’ and mothers’ levels of edu-
cation were signif icantly related to the well-being of the
emerging adults and parental attachment. On the basis of
these f indings, the education levels of the mothers and
emerging adults were considered in the main analyses of
this study. In addition, because research has commonly
noted gender differences in levels of self-worth and de-
pressive symptoms (Nolen-Hoeksema, 1987) and in close-
ness to family (Josselson, 1988; Kenny & Donaldson, 1991),
we examined gender differences for the study variables.
The results of t tests revealed no signif icant gender differ-
ences for the measures completed by the emerging adults
or for the mothers’ ratings of attachment. As a result, gen-
der was not considered in subsequent analyses.

Secure Base Characteristics of
Emerging Adult Attachment

The descriptive data indicate that among the emerging adults
in this sample, proximity seeking and secure base behavior
with mothers were high, with emerging adults reporting fre-
quent contact and communication with their mothers. Most
participants (54%) reported that they communicated with
their mothers once a week, with many (33%) reporting daily
contact. Forty-one percent reported receiving at least some

Measure M

Emerging adults
General Self-Worth Scale
CES-D
PAQ

Mothers
PAQ-P
Parental Burden Scale
Family Cohesion subscale

SD Range

3.28
11.73

3.93

4.44
10.66
37.99

0.67
9.85
0.58

0.45
4.79
6.48

1–4
0–52

2.45–4.47

3.13–5.00
8–32
18–50

TABLE 1

Means and Standard Deviations
for the Study Measures

Note. CES-D = Center for Epidemiological Studies Depression Scale
(Radloff, 1977); PAQ = Parental Attachment Questionnaire (Kenny,
1987); PAQ-P = parent version of PAQ.

Journal of Counseling & Development ■ Winter 2006 ■ Volume 8466

Kenny & Sirin

contribution from their parents toward their living expenses
in the last year. These characteristics of proximity seeking
and support were found to be unrelated to perceptions of
parental attachment, however. Correlational analyses re-
vealed no significant associations between emerging adults’
self-reports of frequency of communication with mothers
and emerging adults’ reports of parental attachment, depres-
sion, or self-worth, although mothers’ perceptions of attach-
ment were positively associated with the frequency of com-
munication (see Table 2). The results of t tests revealed no
significant relationship between emerging adults’ financial
independence and their self-reports and mothers’ reports of
attachment, their self-worth, or their depressive symptoms.

Self-Worth as a Mediator of Parental Attachment
and Depressive Symptoms

In the next series of analyses, we examined whether or not
global self-worth mediates the relationship between parental
attachment and the experience of depressive symptoms. Ac-
cording to Baron and Kenny (1986), mediation is established
when (a) the independent variable significantly predicts the
dependent variable, (b) the independent variable significantly
predicts the mediator variable, and (c) the effect of the indepen-
dent variable on the dependent variable is diminished when
the mediator variable is controlled for (entered simultaneously
in the regression equation). More specifically, mediation is
said to occur if the independent variable in the third step is
either no longer significant (full mediation) or is reduced in
comparison with the first equation (partial mediation).

We conducted three two-step hierarchical regression equa-
tions, following the Baron and Kenny (1986) procedure, in
order to examine whether or not self-w

RESEARCH

A Preliminary Construct Validation of
the Multidimensional Measurement of
Religiousness/Spirituality Instrument:

A Study of Southern USA Samples

Chris Stewart and Gary F. Koeske
School of Social Work

University of Pittsburgh

The Multidimensional Measurement of Religiousness/Spirituality (MMRS) survey
was administered to 515 respondents from the southeastern United States, 355 gradu-
ate and undergraduate students in Social Work, and 160 clients undergoing alcohol-
ism or drug treatment. Exploratory factor analysis and reliability analyses resulted in
retaining 59 of 81 MMRS items measuring 3 primary factors (Meaning, Spirituality,
Religious Practices and Organized Religiousness) and 2 secondary factors (Guilt vs.
God’s Grace, and Loving/Forgiving God). The primary factors were highly internally
consistent and showed acceptable stability reliability for a subsample of clients; they
also largely replicated intended dimensions of the MMRS. Convergent, discriminant,
and theoretical construct validity were generally supported for the factors in
correlational analyses, but the Guilt dimension, in particular, should be reevaluated
in future research. Various limitations were discussed. It was suggested that the 3 pri-
mary factors could be reliably assessed with 10 items, and that this 30-item scale
could be a very useful multidimensional research tool.

There appears to be an increasing interest in the subject of religiosity and spiritual-
ity in human behavioral research. Numerous articles investigate the possible bene-
fits of religious and spiritual domains for health and mental health (Cooper-Patrick

THE INTERNATIONAL JOURNAL FOR THE PSYCHOLOGY OF RELIGION, 16(3), 181–196
Copyright © 2006, Lawrence Erlbaum Associates, Inc.

Correspondence should be sent to Chris Stewart, University of Pittsburgh, School of Social Work,
2117 Cathedral of Learning, Pittsburgh, PA 15260. E-mail: jcs24@pitt.edu

et al., 1997; Ming, Williams, Simpson, & Lyons, 2002; Russinova, Wiewiorski, &
Cash, 2002; Thompson, 2002; Thoreson & Harris, 2002). Similarly, the contribu-
tion of religiosity and spirituality has been the focus of work in the addiction treat-
ment arena (Cancellaro, Larson, & Wilson, 1982; Connors, Tonigan, & Miller,
1996; Koenig, George, Meador, Blazer, & Ford, 1994; Miller 1998). Despite the
lack of conclusive results, sufficient evidence suggests that these constructs con-
tribute to positive outcomes (Hackney & Sanders, 2003).

The psychology of religion has made great advances in conceptualizing and
operationalizing many dimensions of religiosity and spirituality so that, currently,
the creation of new instruments is generally discouraged (Gorsuch, 1984, 1990). In
one volume alone Hill and Hood (1999) reviewed over 120 measures of religiosity,
including often-utilized measures, such as the Spiritual Well-Being Scale (SWB)
and the Religious Orientation Scale (ROS).

Despite the vast array of measures, there is room for improvement in measuring
both religiosity and spirituality, particularly in relation to health and mental health
outcomes (Hill & Pargament, 2003). One suggestion from the literature is to in-
clude both religiosity and spirituality domains in a single measure that might be
conceptualized as an individual’s global spirituality (Miller & Thoreson, 1999).

Spirituality, as thus defined, is a complex construct generally thought to be best
defined as multidimensional, encompassing every individual’s beliefs concerning
reality beyond the sensory, material world (Larson, Swyers, & McCullough, 1998;
Miller 2003). The exact number and nature of dimensions has been the subject of
some debate in the literature, although some have suggested components, includ-
ing self-identification or affiliation, formal practice, private practice, recognition
of importance or centrality, belief, and spiritual experience (Hill & Hood, 1999;
Hood, Spilka, Hunsberger, & Gorsuch, 1996; Miller & Thoreson, 1999). Addi-
tionally, there is general agreement that religiosity and spirituality are distinct but
related constructs that should be considered collectively, particularly in assessing
the impact of global spirituality on outcomes of interest (Hill & Pargament, 2003;
Miller, 1998, 2003).

Some proven and accepted instruments may tap particular dimensions of spiri-
tuality as exemplified by the Spiritual Assessment Inventory (SAI) and SWB;
many measures tend to focus primarily on aspects of religiosity, such as the ROS
(Hall & Edwards, 1996; Hill & Hood, 1999). Because the domains are signifi-
cantly related it seems important to try and capture an overall spiritual construct by
including both domains in one measure. A further concern is the inclusion of spiri-
tual expressions that are outside the Judeo–Christian tradition as many current
measures were normed without such considerations (Hill & Pargament, 2003).

A consortium of religiosity, spirituality and health specialists (Fetzer Institute,
1999) collaborated to compile a multidimensional measure of religiousness and
spirituality (MMRS) that included multiple domains of both religiosity and spiri-
tuality related to an individual’s overall or global spirituality that might be used in

182 STEWART AND KOESKE

health and mental health research. Conceptual and methodological issues in re-
searching spirituality, religiosity, health, and aging were explored (Underwood &
Teresi, 2002). This ambitious exercise resulted in a full-length, 88-item measure
and a short form with domains, including daily spiritual experiences, religious be-
liefs, life meaning, religious practice, forgiveness, and other elements. This elabo-
rate instrument may have utility for studies dedicated to the examination of religi-
osity and spirituality as preeminent variables and attempts to holistically capture
global spirituality.

As of September 2004 the Fetzer Institute reported on its Web site that 23 pub-
lished studies have utilized either the complete measure or the brief form of the
MMRS and provides a list of 58 publications that cite the measure. The first author
has utilized both forms of the MMRS in published research involving student atti-
tudes concerning religiosity in social work practice and treatment of addictions
(Stewart & Koeske, in press-a, in press-b). Despite the interest in the measure and an
increase in its use, little psychometric evidence is currently available for the MMRS.

Some of the longer subscales have psychometric evaluation. For example,
Underwood and Teresi (2002) reported on the reliability and validity of the Daily
Spiritual Experiences Scale. This scale has undergone significant psychometric
study with three different samples demonstrating good reliability and internal con-
sistency greater than .90. Similarly, the domain tapping religious coping has been
extensively tested (Pargament, 1999; Pargament, Koenig, & Perez, 2000). The full
RCOPE has demonstrated good validity and reliability with two different samples.
Further, factor analysis was largely consistent with a priori conceptualization
(Pargament, Koenig, & Perez, 2000). The Brief RCOPE, utilized by the MMRS
also demonstrated strong internal consistency and validity (Pargament, 1999).

This project is a beginning effort into validating the complete MMRS instru-
ment as a comprehensive measure of global spirituality. Exploratory factor analy-
sis was used to examine the structure of the MMRS and to compare the derived fac-
tors to the a priori or intended item clusters. Homogeneity and test–retest
reliabilities are reported for all subscales. Construct validity evaluations are pro-
vided, including convergent, discriminant, and theoretical validity.

METHODS

Samples

Two samples responded to the MMRS and related instruments. The first sample
was composed of 355 undergraduate and graduate social work students from three
universities in the southeastern United States. The second sample was composed
of 160 clients receiving treatment for alcoholism at centers also located in the
southeast. The clients received traditional 12-step treatment in a 28-day inpatient
setting. Seventy of the treatment clients responded to the MMRS and related scales

VALIDATION OF THE MMRS 183

a second time, approximately one month after the first testing. For the exploratory
factor analysis, the student and client samples were combined into one larger sam-
ple, N = 515. The students responded anonymously in classroom groups of various
sizes; the clients responded to the confidential survey shortly after intake and (for
those responding to the posttest) about one month later after completing treatment.

In the Student sample, 317 (89%) were females, the median age was 25 years,
and 54% were White, 21% African American, 13% Hispanic American, and the
remainder were “other” racial categories of 1% or lower representation. Their reli-
gious identification was predominantly Protestant (47%) and Catholic (22%), with
other affiliations accounting for 18% of the sample. Of this 18% slightly less than
3% were Jewish, and slightly greater than 7% was nonreligious, Buddhist, or exis-
tentialist. The Client sample was older (median age = 38), largely male (67.5%),
and predominantly African American (52%) or White (44%), with only 2% His-
panic. One half had experienced previous treatment. The primary drugs of choice
were alcohol (34%), marijuana (24%), and cocaine or crack cocaine (27%).

Measures

The full-length MMRS was administered with standard instructions to both sam-
ples. The 88 MMRS questions are organized into 12 labeled subsections with k
items per section: Daily spiritual experiences (k = 16), Meaning (k = 20), Beliefs (k
= 7), Forgiveness (k = 10), Private religious practices (k = 4), Religious/spiritual
coping (k = 11), Religious/spiritual history (k =5), Organizational religiousness (k
= 7), Commitment (2), Religious preference (k = 1), Values (k = 3), and Overall
ranking (k = 2). Our psychometric evaluation of the measure excluded the religious
preference and history sections as well as the monthly or yearly dollar contribution
item. This left 81 items, which were rated on 3-, 4-, 5-, 6-, 8-, and 9-step ordinal
metrics. Each a priori labeled subsection with three or more items was assessed as
a separate subscale and potential construct.

In addition to the MMRS, the Student sample responded to a number of mea-
sures of religious/spirituality and social work practice developed by Sheridan
(1992, 1999). These included the following:

1. A 19-item Likert-type scale measuring the attitude toward infusing a reli-
gious or spiritual element in social work practice.

2. A 15-item (yes or no) scale for perceived appropriateness of various reli-
gion-relevant practice behaviors.

3. A summed rating of whether the respondent had personally used the listed
practice behavior.

4. A single-item 8-step report of frequency of current participation “in reli-
gious services.”

184 STEWART AND KOESKE

5. A single-item 6-step report of current participation in “religious or spiritual
practices.”

6. A single-item 5-step report of “present relationship to an organized reli-
gion or spiritual group,” from “active participation, high level of involve-
ment” to “disdain and negative reaction to religion or spiritual tradition.”

7. A reported choice among six religious ideological positions, which we
scored as “1” for endorsement of “There is a personal God or transcendent
existence and power whose purpose will ultimately be worked out in his-
tory” and “0” for endorsement of any of the other 5 positions.

8. Reports of having had graduate training in religion or theology and participa-
tion in workshops or conferences on religion or spirituality in the last 5 years.

These measures were used to assess convergent and theoretical validity of the
MMRS.

The Client sample respondents completed, along with the MMRS, two addi-
tional multi-item measures of religiosity: the Allport and Ross (1967) 26-item Re-
ligious Orientation Scale (ROS) and spirituality, the 20-item Spiritual Well-being
Scale (SWB; Ellison 1983). These scales are scored by summation of the items
rated on 1–5 (ROS) and 1–6 (SWB) agree–disagree scales. These measures were
also used to assess convergent and theoretical validity of the MMRS.

RESULTS

Internal Consistency of the MMRS “A Priori” Section
Subscales

Prior to exploring the factor structure of the MMRS, we estimated the homogene-
ity of the eight item sets of “a priori” subscales that had three or more items. The
coefficient alphas are reported in Table 1 for the eight subscales for the Student and

VALIDATION OF THE MMRS 185

TABLE 1
Homogeneity (Alpha) Reliabilities for MMRS “A Priori” Section Scales

Students (N = 355) Clients (N = 160) Merged (N = 515)

Daily Spiritual Experiences .93 .93 .93
Meaning .96 .95 .96
Beliefs .83 .78 .81
Forgiveness .56 .44 .50
Religious Practices .83 .81 .81
Religious/Spiritual Coping .79 .82 .79
Organizational Religiousness .84 .80 .83
Values .44 .36 .40

Client samples separately. The alphas are in the very good to acceptable range for 6
of the 8 subscales; they are, however, very low for the 3-item Values set and the
10-item Forgiveness set, for which none reach .60.

Exploratory Factor Analysis

Because exploratory factor analysis is a large sample procedure, the Student and
Client samples were merged into one large sample of 515 cases. Although the
81-item MMRS is relatively new and may not be appropriate for a confirmatory
factor analysis, we maintained the a priori subsections as one guide for the number
of factors to extract, along with the scree test, imputation of meaningfulness, and a
requirement of at least three items loading uniquely to a .40 criterion on a factor. A
principal components analysis followed by an oblique rotation of the factors ex-
tracted eight factors, while the scree test suggested five viable factors.

Table 2 shows the loadings for each of the five retained factors. Fifty-seven per-
cent of the scale variance was attributable to the 8 factors, 52% to the 5 retained
factors. The KMO diagnostic assessment of the item set’s suitability for analysis
was .96, far exceeding the .70 criterion. All 20 Meaning items of the MMRS
loaded uniquely on the first factor and accounted for 27% of the scale variance; all
but two loaded above .50. One item that loaded .44 on this factor, an item from the
Religious/Spiritual Coping section, was omitted because it was complex, loading
beyond .35 on two other factors. The next largest factor (labeled Factor 3; 21% of
variance) included 10 the 11 items from the Religious Practices and Organiza-
tional Religiousness sections plus the overall religiosity item (To what extent do
you consider yourself a religious person?), and the third largest factor (labeled
Factor 3; 18% of the scale variance) included 14 of the 16 Spiritual Experiences
items plus the overall spirituality item (To what extent do you consider yourself a
spiritual person?). So, these three empirical factors replicated fairly well their cor-
responding a priori item subsections.

Table 2 shows that the second factor after the oblique rotation (5.6% of the vari-
ance) included 4 items from the Forgiveness section of the MMRS and 2 items
from the Spiritual/Religious Coping section. The four former items each express a
sense of having done wrong, but do not directly refer to forgiveness. The latter two
items refer to receiving punishment from, and having been abandoned by, God.
When scored in this “negative” direction the items of this factor seem to express a
sense of guilt; it will, therefore, be labeled the Guilt dimension. The “positive” (or
“proreligious”) end of this dimension might be regarded to represent the notion of
God’s grace. This is the only dimension scored in a “nonreligious” direction, in the
sense that high scorers report feeling abandoned by God and not forgiven. Because
of possible ambiguity in this designation (i.e., high scorers may be implicitly ac-
knowledging that there is a God that reviews their behavior and evaluates their
lives), an exploratory analyses was conducted extracting between 5 and 9 factors to

186 STEWART AND KOESKE

187

TABLE 2
Factor Loadings for Five Retained Factors for 59 MMRS Survey Items

Factor

MMRS Item & Number 1 2 3 4 5

Factor 1: Meaning
34. My feelings of spirituality add meaning to the events in my life. .81 –.05 –.02 .14 .02
36. My spirituality helps define the goals I set for myself. .79 –.05 –.04 .11 .06
23. My spiritual beliefs give my life a sense of significance and purpose .78 –.09 –.02 .08 .00
31. Looking at the most troubling or confusing events from a spiritual

perspective adds meaning to my life
.76 –.10 –.01 .14 –.02

25. When I am disconnected from the spiritual dimension in my life, I
lose my sense of purpose.

.71 .10 –.07 –.01 –.07

19. Without a sense of spirituality, my daily life would be meaningless. .70 .01 –.04 .02 –.04
28. What I try to do in my day-to-day life is important to me from a

spiritual point of view.
.70 –.05 –.09 .18 .13

30. Knowing I am part of something greater than myself adds meaning to
my life.

.69 .02 –.02 .10 –.12

17. My spiritual beliefs give meaning to my life’s joys and sorrows. .66 –.09 –.06 .04 .02
22. When I lose touch with God, I have a harder time feeling there is

purpose and meaning in life.
.63 .15 –.09 –.12 –.08

21. My religious beliefs help me find a purpose in even the most painful
and confusing events in my life.

.62 –.09 –.15 –.07 –.17

18. The goals of my life grow out of my understanding of God. .61 –.01 –.13 .00 –.04
35. God plays a role in how I choose my path in life. .60 .03 –.15 .04 –.25
32. My purpose in life reflects what I believe God wants for me. .59 .12 –.14 –.02 –.26
26. My relationship with God helps me find meaning the ups and downs

of life.
.59 .02 –.10 .02 –.33

24. My mission in life is guided/shaped by my faith in God. .58 .05 –.19 .01 –.26
29. I am trying to fulfill my God-given purpose in life. .57 .06 –.08 .06 –.23
27. My life is significant because I am part of God’s plan. .52 .07 –.11 .01 –.39
33. Without my religious foundation, my life would be meaningless. .47 .26 –.23 –.02 –.18
20. The meaning in my life comes from feeling connected to other living

things.
.47 –.13 .23 .03 .14

Factor 2: Guilt
63. I feel that stressful situations are God’s way of punishing me for my

sins or lack of spirituality.
–.09 .73 .11 –.05 .07

47. I belief there are times when God punished me. –.13 .72 .07 .07 –.08
49. I often feel that no matter what I do now I will never make up for the

mistakes I have made in the past.
.04 .63 –.01 –.01 .06

64. I wonder whether God has abandoned me. –.01 .59 .04 –.12 .39
53. I often feel like I have failed to lead the right kind of life. .13 .55 .01 –.07 –.02
45. If I hear a sermon, I usually think about things I have done wrong. .07 .49 –.16 .01 –.23

Factor 3: Religious Practices & Organized Religiousness
78. I feel at home in this church/synagogue. –.01 –.07 .86 –.06 –.09
76. How well do you feel you fit in your church/synagogue? .01 –.09 .81 –.01 –.09
77. If I had to change churches/synagogues, I would feel a great sense of

loss.
–.06 –.06 .79 –.08 –.00

(Continued)

evaluate its stability. Though it is a “small” factor, it appeared in each exploratory
analysis performed.

The last (fifth) factor (accounting for 13% of the variance) included 4 items
from the Beliefs section, 1 from the Forgiveness section, and 2 from the Spiri-

188 STEWART AND KOESKE

TABLE 2 (Continued)

Factor

MMRS Item & Number 1 2 3 4 5

80. The church/synagogue matters a great deal to me. .06 –.03 .78 –.05 –.08
74. How often do you attend religious services? .07 –.01 .77 .05 .09
75. Besides religious services, how often do you take part in other

activities at a place of worship?
.12 –.05 .65 .12 .25

87. To what extent do you consider yourself a religious person? .08 .05 .56 .05 –.05
56. How often do you read the Bible or other religious literature? .10 .12 .56 .18 .15
55. How often do you watch or listen to religious programs on TV or radio? –.05 .37 .46 .16 .09
57. How often are prayers or grace said before or after meals in your home? .04 .22 .43 .21 –.10
81. I try hard to carry my religious beliefs over into all my other dealings

in life.
.26 .02 .41 .10 –.06

Factor 4: Spirituality
2. I experience a connection to all life. .09 –.07 .14 .70 .08
11. The beauty of creation spiritually touches me. .14 –.02 –.02 .67 .03
6. I feel deep inner peace or harmony. –.11 .01 –.04 .66 .05
12. I feel thankful for my blessings. .00 .07 .02 .64 –.25
13. I feel a selfless caring for others. –.08 –.08 –.10 .59 –.03
5. I find comfort in my religion or spirituality. .24 .00 –.14 .57 –.05
4. I find strength in my religion or spirituality. .25 –.01 –.15 .57 –.05
1. I feel God’s presence. .17 –.02 –.09 .53 –.21
14. I accept others even when they do things I think are wrong. –.10 .07 –.13 .50 .08
3. During worship, or at other times when connecting with God, I feel joy

which lifts me out of my daily concerns.
.11 .15 –.23 .47 –.19

7. I ask for God’s help in the midst of daily activities. .07 .10 –.24 .45 –.20
8. I feel guided by God in the midst of daily activities. .14 .14 –.18 .45 –.22
88. To what extent do you consider yourself a spiritual person? .26 –.25 –.05 .44 .26
10. I feel God’s love for me through others. .15 .06 –.16 .44 –.20
9. I feel God’s love for me directly. .18 .05 –.20 .41 –.30

Factor 5: Loving/Forgiving God
66. I question whether God really exists. .26 –.14 –.11 .06 .56
39. God’s goodness and love are greater than we can possibly imagine. .26 .09 –.18 .01 .53
41. When I am faced with a tragic event I try to remember that God still

loves me and that there is hope for the future.
.29 .12 –.05 .03 .52

65. I try to make sense of the situation and decide to do without relying
on God.

.10 –.23 –.10 .19 .48

43. I think that everything that happens has a purpose .12 .18 .06 .08 .44
42. I feel that it is important for my children to believe in God. .21 .15 –.28 –.06 .44
46. I believe that God has forgiven me for things I have done wrong. .13 –.01 –.11 .10 .40

tual/Religious Coping section. These items seem to reflect a firm belief in a loving
and forgiving God that provides purpose to one’s life and it was also relatively sta-
ble across exploratory analyses. This set includes the item “I question whether God
exists,” for which high scorers were more likely to answer “not at all.”

The remaining three factors in this solution were not regarded as viable. The
sixth factor included three items from different sections, and one of these items
was complex, having a loading to criterion on the fifth factor as well. No items
loaded to criterion on the seventh factor. The last factor included four items from
the Forgiveness section, which referred to interpersonal, but not God-granted, for-
giveness. It did not, therefore, appear to reflect a dimension of religiosity.

This central analysis supported the integrity of two of the a priori MMRS sec-
tions (Meaning and Spiritual Experiences) and folded two others (Religious Prac-
tices and Organizational Religiousness) into a single dimension, Organized Reli-
gion and Religious Practices. Fifty-nine items loaded uniquely on one of the five
factors; 22 items did not load on a viable factor, loaded below criterion, or loaded
on more than a single factor (i.e., they were complex items).

Factor Score Distributions

Each of the distributions was skewed in the direction of lower religiosity/spiritual-
ity scores (i.e., there were a small number of respondents with very low scores).
The skew was minimal for Organized Religion and Practices (sk = –.23) and not
excessive for Spirituality (sk = –.70) and Guilt scores (sk = .71). The skewness of
the Meaning scores exceeded 1.0 (sk = –1.09), but could be corrected to –0.46 with
a square transformation. The Loving and Forgiving God scores were highly
skewed (sk = –2.12) and nontransformable. It is likely that similar non-normality
will be found for Meaning and Loving God scores in other samples, including het-
erogeneous/representative samples. Therefore, transformations may be needed for
multivariate analysis of Meaning scores and scores on the Loving and Forgiving
God dimension may need to be dichotomized.

Factor Score Intercorrelations

The factor scores were generally not correlated providing support for the unique-
ness of each derived factor. The factor intercorrelations ranged between .45 for
Meaningfulness with Spirituality to 0.00 for God’s Grace with Spirituality.

Homogeneity and Test–Retest Reliabilities of the Empirical
Dimensions of Religiosity/Spirituality

We used the saved standardized factor scores to measure the five retained dimen-
sions, partly due to the several metric lengths used in the MMRS. It would be use-

VALIDATION OF THE MMRS 189

ful, however, to know the conventional alpha reliability estimates for the items that
are identified with each factor, based on the factor loadings exceeding the .40 crite-
rion. Table 3 shows these estimates for the merged sample of 515, as well as for the
separate Student and Client samples. The alphas reported are the standardized item
alphas, because the items may have been rated on scales of different length. These
alphas were quite high, ranging from the high 80s to the 90s, with the exception of
the Guilt set, which .73 in the merged file, .72 in the Student sample, and .56 in the
Client sample.

The sample of 70 clients that responded to the MMRS survey twice was used to
provide an estimate of stability reliability. Because these respondents were in alco-
holism treatment, their true religiosity/spirituality scores might be expected to
change at varying rates. Consequently, these estimates may underestimate the reli-
ability of the measured dimensions. Table 3 includes these stability coefficients.
Pearson coefficients were used for four of the tests; a Spearman rho was used for
the Loving/Forgiving God dimension, which was markedly negatively skewed.
Given that these clients had undergone an intervention, the estimates are quite high
for Meaning, Spirituality, and Religious Practices and Organized Religion, but
modest for Loving/Forgiving God (rho = .61) and Guilt (.57).

Convergent and Discriminant Validation of the MMRS
Empirical Dimensions

Four questions about religiousness and spirituality asked of the students might be
expected to relate to the factor scores derived from the MMRS. These relationship
tests reflect on convergent validity, albeit dimly, as these are exclusively self-report
measures. Table 4 presents the nonparametric intercorrelations of the factor scores
with endorsement of the statement “There is a personal God or transcendent exis-
tence and power whose purpose will ultimately be worked out in history,” and with
frequency of participation “in religious services,” frequency of participation in

190 STEWART AND KOESKE

TABLE 3
Alpha Reliabilities for Empirically-Derived Scales in Separate and Merged

Samples and Stability Coefficients for 70 Clients

Sample

Scale
Students

(N = 355)
Clients

(N = 160)
Merged

(N = 515)
Prepost Stability

(N = 70)

Meaning .96 .95 .96 .67
Spirituality .92 .93 .92 .78
Religious Practices & Organized Religion .92 .89 .91 .84
Guilt .72 .56 .73 .57
Loving/Forgiving God .88 .80 .87 .61

“personal religious or spiritual practices,” and extent of “relationship to an orga-
nized religion or spiritual group.” Excluding the Guilt dimension, we see that the
factor scores are significantly and often substantially related to these single-item
measures of religion and spirituality. Consistent with discriminant validity, the or-
ganized religion dimension was particularly highly related to frequency of partici-
pation in religious services and to report of a deeper involvement in organized reli-
gion, whereas the correlation of the frequency of spiritual practices item was
somewhat more highly related to the Spirituality dimension than to the organized
religion factor. Although the correlations with Guilt were very low, this factor was
significantly related to participation in religious services (rho = .19, p < .001) and
endorsement of personal God ideology (rho = .12, p = .04).

Table 4 also shows the correlations of the factor scores with the scores on the
Religious Orientation Scale and the Spiritual Well-Being Scale for the 160 respon-
dents in the Client sample. These correlations range between .40 and .64 for the
MMRS factors, excluding Guilt. Low Guilt, which might be construed as a belief
in God’s grace, was related more modestly, but significantly, to the two independ-
ent measures of religiosity and spirituality.

Theoretical Construct Validity

Beyond support for the rather obvious expectation that people who do not identify
themselves with religious group should score lower on dimensions of religiosity
and spirituality, the Student sample permits the assessment of construct validity
through the administration of the Sheridan (1999) measures of attitude of inclusion

VALIDATION OF THE MMRS 191

TABLE 4
Spearman Correlations of Factor Scores With Religious Involvement

Questions in Student Sample and With Religious Orientation Scale (ROS)
and Spiritual Well-Being (SWB) Scale in Client Sample

Item/Scale

Factor Scale

Participation
In Religious

Services

Personal
Spiritual
Practices

Personal
God

Ideology

Involvement
Organized
Religion ROS SWB

Meaning .53** .45** .38** .47** .51** .62**
Spirituality .52** .53** .27** .44** .40** .59**
Religious Practices
& Organized
Religion

.83** .47** .42** .72** .54** .55**

Guilt .19** –.05 .12* .07 –.23** –.32**
Loving/Forgiving
God

.51** .37** .41** .41** .55** .64**

*p < .05. **p < .01

of religious-based practices in social work practice, the appropriateness of specific
exemplary religious practices, and the actual use of such practices. Many of these
students, particularly the graduate students, had fieldwork or employment experi-
ence in which these practice behaviors may have been utilized. As would be ex-
pected (see Table 5), respondents with a more positive attitude, those who deemed
more religious-based practices “appropriate,” and those who had personally prac-
ticed more of these behaviors scored significantly higher on Meaning, Spirituality,
Religious Practices and Organized Religiousness. Guilt scores were not related to
any of the three measures, and the Loving/Forgiving God score was only signifi-
cantly related to Attitudes and use of practice behaviors.

Finally, based on the Client data, we might expect that valid scores on religi-
osity/spirituality would be sensitive to exposure to a 12-step treatment interven-
tion, which emphasizes relationship to a transcendent spiritual power. We would
predict, therefore, that clients undergoing such an intervention would score
higher on dimensions of religiosity following exposure to 12-step-oriented alco-
holism or drug treatment. Wilcoxin signed ranks tests showed that clients’ religi-
osity factor scores were higher at the posttest, compared to the pretest, for each
of the comparisons. The probabilities achieved were .09 for Meaning, .006 for
Spirituality, .001 for Organized Religiosity, .04 for Loving/Forgiving God,

Journal of Personality and Social Psychology
1994, Vol. 67, No. 6, 1063-1078

Copyright 1994 by the American Psychological Association, Inc.
0022-3514/94/S3.00

Distinguishing Optimism From Neuroticism (and Trait Anxiety,
Self-Mastery, and Self-Esteem): A Reevaluation of the

Life Orientation Test

Michael F. Scheier, Charles S. Carver, and Michael W. Bridges

Research on dispositional optimism as assessed by the Life Orientation Test (Scheier & Carver, 1985)
has been challenged on the grounds that effects attributed to optimism are indistinguishable from
those of unmeasured third variables, most notably, neuroticism. Data from 4,309 subjects show that
associations between optimism and both depression and aspects of coping remain significant even
when the effects of neuroticism, as well as the effects of trait anxiety, self-mastery, and self-esteem,
are statistically controlled. Thus, the Life Orientation Test does appear to possess adequate predictive
and discriminant validity. Examination of the scale on somewhat different grounds, however, does
suggest that future applications can benefit from its revision. Thus, we also describe a minor modi-
fication to the Life Orientation Test, along with data bearing on the revised scale’s psychometric
properties.

Accumulating evidence from a variety of sources suggests
that dispositional optimism is beneficial for physical and psy-
chological well-being. For example, Aspinwall and Taylor
(1992) have recently shown that optimistic persons adjust more
favorably to important life transitions than do persons who are
more pessimistic in outlook. In a similar vein, Litt, Tennen,
Affleck, and Klock (1992) have reported that optimistic women
who are unsuccessful at in vitro fertilization respond better to
the failure than do women who are more pessimistic. Concep-
tually similar results have also been reported by Scheier et al.
(1989). Their study tracked a group of men undergoing coro-
nary artery bypass surgery. Optimistic men evidenced a more
rapid physical recovery after their surgery and reported a higher
quality of life 6 months postoperatively than did the more pes-
simistic men in the sample. Nor are these the only beneficial
effects for dispositional optimism that have been reported in
the literature (for a more comprehensive review, see Scheier &
Carver, 1992).

Related research suggests that these differences in outcomes
derive partly from differences between optimists and pessimists
in the manner in which they cope with the challenges in their
lives. Optimists differ from pessimists in their stable coping

Michael F. Scheier and Michael W. Bridges, Department of Psychol-
ogy, Carnegie Mellon University; Charles S. Carver, Department of Psy-
chology, University of Miami.

Data collection for this article was supported by National Science
Foundation (NSF) Grants BNS87-17783 and BNS87-06271. Prepara-
tion of the article was facilitated by NSF Grants BNS90-11653 and
BNS90-10425, American Cancer Society Grant PBR-56, and National
Heart, Lung, and Blood Institute grant HL44436-01A1.

We are indebted to Joel Greenhouse for consulting on those aspects
of the data analysis pertaining to factor analysis, and to Jeffrey Thomas
for assisting in the preparation of data tables.

Correspondence concerning this article should be addressed to Mi-
chael F. Scheier, Department of Psychology, Carnegie Mellon Univer-
sity, Pittsburgh, Pennsylvania 15213.

tendencies (Carver, Scheier, & Weintraub, 1989) and in the
kinds of coping responses that they spontaneously generate
when given hypothetical coping situations (Scheier, Weintraub,
& Carver, 1986). Optimists also differ from pessimists in the
manner in which they cope with serious disease (Friedman et
al., 1992) and with concerns about specific health threats (e.g.,
Carver etal., 1993; Stanton& Snider, 1993; Taylor etal., 1992).
A general characterization of the findings of this research is that
optimists tend to use more problem-focused coping strategies
than do pessimists. When problem-focused coping is not a pos-
sibility, optimists turn to more adaptive emotion-focused cop-
ing strategies such as acceptance, use of humor, and positive
reframing of the situation. Pessimists tend to cope through overt
denial and by mentally and behaviorally disengaging from the
goals with which the stressor is interfering, regardless of
whether something can be done to solve the problem or not.

These findings regarding optimism are consistent with the
model of behavioral self-regulation from which our own work
on optimism grew (e.g., Carver & Scheier, 1981, 1990a; Scheier
& Carver, 1988). This is a model that has roots in the long tra-
dition of expectancy-value theories in psychology. In this
model, people are seen as remaining engaged in efforts to over-
come adversity to reach goals as long as their expectancies of
eventual success are sufficiently favorable. When their doubts
are too severe, people are more likely to give up on the threat-
ened goals. These differences in expectancies are also assumed
to be paralleled by variations in affective experience (for details,
see Carver & Scheier, 1990b). With enough movement toward
desired goals, affect is positive. If movement toward desired
goals is sufficiently impeded, affect is negative.

Although this viewpoint on behavior and affect can be ap-
plied in terms of situational variations in expectancies across
time or events, it can also be applied in terms of individual
differences. Optimists are people who tend to hold positive ex-
pectancies for their future; pessimists are people who tend to
hold more negative expectations for the future. Thus, our anal-

1063

1064 M. SCHEIER, C. CARVER, AND M. BRIDGES

ysis of how optimism versus pessimism leads to different re-
sponses to adversity is one application of a more general model
of the processes that underlie behavior, a model that is applica-
ble to a wide range of motivational issues and contexts.

Critique and Challenge

Much of the research on optimism and pessimism (although
certainly not all of it; e.g., Beck, Steer, Kovacs, & Garrison,
1985; Reker & Wong, 1983) has made use of the Life Orienta-
tion Test (LOT; Scheier & Carver, 1985) to assess individual
differences on this dimension. This scale has recently been crit-
icized by others. By implication, this criticism also undermines
the integrity of the optimism construct. The primary purpose
of this article is to address these issues.

Most of the criticism aimed at the scale involves the third
variable problem. Smith, Pope, Rhodewalt, and Poulton (1989)
were the first to raise this issue, doing so with respect to trait
anxiety. That is, they questioned whether effects attributable to
optimism might really be due to variance that optimism shared
with trait anxiety. Consistent with this view, Smith et al. (1989)
reported relatively high correlations between optimism and
trait anxiety across two independent samples. Smith et al. also
showed that it was possible to eliminate the significant negative
association that they found between optimism and reports of
physical symptoms by controlling for the effects of trait anxiety.
In contrast, the association between trait anxiety and reports of
physical symptoms remained significant even after the effects
of dispositional optimism were removed. Similarly, significant
zero-order correlations between optimism and different varie-
ties of coping were largely eliminated when the effects of trait
anxiety were controlled, whereas many of the associations be-
tween anxiety and coping remained strong after removing the
effects of dispositional optimism. In a similar vein, Marshall
and Lang (1990) have also raised the third variable problem,
but with respect to self-mastery rather than trait anxiety (see
also Robbins, Spence, & Clark, 1991).

We have several observations to make with respect to this
work. Our first point concerns the nature of the outcome vari-
ables that have been examined across studies. Both Smith et
al. (1989) and Robbins et al. (1991) examined the relationship
between optimism and health complaints, and neither found an
independent effect for optimism when variables such as trait
anxiety were controlled. Smith et al. also examined coping
strategies, however, and found that optimism was an indepen-
dent predictor of certain coping responses. Similarly, Robbins
et al. (1991) also examined health maintenance behaviors and
found optimism to be an independent predictor of these. The
point here is that optimism may be a stronger independent pre-
dictor of some outcomes than of others. Shared variance may
explain the association between optimism and symptom report-
ing, for example, but may not fully explain the link between
optimism and other outcomes of interest.

There is a second (though related) issue here, as well. Previous
researchers have used outcome measures that were somewhat
limited in scope. For example, Smith et al. (1989) relied on only
five coping categories to explore associations between disposi-
tional optimism and coping tendencies, controlling for trait anx-
iety. Yet a far greater number of coping responses can be identi-

fied and measured reliably (Carver et al., 1989). The possibility
thus remains that dispositional optimism may be uniquely re-
lated in important ways to outcome variables that went un-
measured in these earlier studies. In this regard, it is interesting
to note that dispositional optimism was associated with eight
different coping qualities in Carver et al.’s (1993) study of ad-
justment to breast cancer surgery. Of these eight coping re-
sponses, only half were measured in the Smith et al. (1989)
study.

Our third observation has to do with the nature of the pre-
dictors with which optimism has been compared. Most of the
concern to date has centered around the overlap between opti-
mism and neuroticism or negative affectivity, as indexed by one
or another measure of chronic anxiety. Note, however, that neu-
roticism is conventionally viewed as a multifaceted construct
that consists partly of the absence of optimism (i.e., pessimism).
Thus, there is a distinct conceptual link between constructs. On
the other hand, neuroticism also incorporates a host of other
factors, such as self-doubt, emotional lability, and worry. Com-
bining qualities in this way can create problems of interpreta-
tion (Carver, 1989) because it becomes very difficult to identify
which components of neuroticism underlie a given effect. As a
hypothetical example, it might be the pessimism facet of neu-
roticism that relates to such variables as active coping, plan-
ning, giving up, and positive reinterpretation. The emotional
lability component may not be as good a predictor of these vari-
ables, but may relate well to other variables such as the experi-
ence of physical symptoms.

The other constructs under consideration here also have con-
ceptual overlap with optimism, though in a different way. Self-
mastery is the perception that one exerts control over the events
in one’s life (Pearlin & Schooler, 1978). This construct thus in-
corporates a strong sense of positive expectancy for the future,
but weds to it a sense of personal responsibility for that expec-
tancy. Self-esteem shares ground with optimism in a more
diffuse way. Self-esteem represents a sense of self-worth, which
carries the implication that one will be accepted rather than
rejected by others, and that one is not a failure in one’s life.
These consequences, of course, involve positive versus negative
outcomes, thus linking self-esteem conceptually to optimism.
As with self-mastery, what seems to differentiate this concept
from optimism involves (at least in part) a kind of ascription to
the self. The ascription in this case, however, is not one of con-
trol but rather of an intrinsic tie between feelings of worth or
the self’s value and positive outcomes.

In sum, the alternative constructs being examined here all
have conceptual as well as empirical overlap with optimism.
Each, however, incorporates at least one additional quality that
takes it beyond optimism. In the case of neuroticism, there may
be several such qualities.

Our fourth and final observation is that results pertaining to
the discriminant validity of the LOT have not all been com-
pletely negative. As noted earlier, Smith et al. (1989) found that
the associations between self-blame and optimism (in both of
their studies) and seeking of social support and optimism (in
one of their studies) remained significant even when trait anxi-
ety was controlled. Robbins et al. (1991) found that the associ-
ation between optimism and health maintenance behaviors re-
mained intact, even when the effects of manifest anxiety, instru-

REEVALUATING THE LIFE ORIENTATION TEST 1065

mentality, anger, and alienation were simultaneously controlled.
Indeed, of all the variables studied, optimism was one of only
two that made significant independent contributions to the pre-
diction of health maintenance behaviors. More recently, Aspin-
wall and Taylor (1992) have reported that optimism predicts
adjustment to the first semester of college, independent of self-
esteem, locus of control, and desire for control. Finally, Mroc-
zek, Spiro, Aldwin, Ozer, and Bosse (1993) have found that op-
timism continues to predict psychological distress among mid-
dle-aged men, even after the distress measure is adjusted for
differences in neuroticism. Given these various considerations,
it seems premature to conclude that the LOT has no predictive
validity independent of other measures. In the same way, it
seems premature to close the book on optimism-pessimism as
an independent construct.

In this article we address this set of questions further. For the
past several years, we have been collecting information from
large groups of respondents on a variety of personality variables,
coping styles, and other outcome measures. The personality bat-
tery includes a measure of optimism-pessimism (the LOT) as
well as measures of self-esteem, trait anxiety, self-mastery, and
neuroticism. The measure of coping is broad in scope and cov-
ers a wide range of diverse coping tendencies. The other outcome
measures include both physical symptoms (number and inten-
sity) and a measure of depression.

The primary purpose of this article is to use the data we have
assembled to reexamine the predictive validity of the LOT by
using it and the other personality factors to predict variations in
coping, symptoms, and depression. Two general sets of predic-
tions are advanced. First, on the basis of findings reported by
Smith et al. (1989), Robbins et al. (1991), and Mroczek et al.
(1993), we expect that zero-order correlations between opti-
mism and physical symptoms will be substantially reduced
when the data are adjusted for trait anxiety and neuroticism.
Second, we expect that other associations involving optimism,
coping, and depression will remain strong, even after the data
are adjusted for the various personality factors that have been
measured.

Study 1: Reevaluating the Life Orientation Test

Method

Subjects and Procedure

A total of 4,309 undergraduates from Carnegie Mellon University and
the University of Miami participated in the research (1,846 women,
2,417 men, and 46 participants who did not indicate their gender). Par-
ticipation was in partial fulfillment of a psychology research require-
ment. AH subjects completed a number of scales as part of a larger pre-
testing packet. Packets were administered in large group testing sessions
across successive semesters from 1988 to 1990. Because of time con-
straints, not all groups received all scales. As a result, sample sizes for
analyses vary from analysis to analysis.

Measures

Optimism. Optimism was measured by using the LOT (Scheier &
Carver, 1985). The LOT is an eight-item self-report measure (plus four
filler items) assessing generalized expectancies for positive versus nega-
tive outcomes. Respondents were asked to indicate their degree of

agreement with statements such as “In uncertain times, I usually expect
the best,” and “I hardly ever expect things to go my way,” using a 5-
point response scale ranging from 0 (strongly disagree) to 4 (strongly
agree). Of the 8 scored items, 4 are worded in a positive direction and 4
are worded in a negative direction. After reversing the scoring for the
negatively worded items, item scores were totaled to yield an overall
optimism score with high scores representing greater optimism. In our
sample, scores ranged from 0 to 32. Cronbach’s alpha was .82.

Neuroticism. The Emotional Stability subscale of the Guilford-
Zimmerman Temperament Survey (GZTS; Guilford, Zimmerman, &
Guilford, 1976) was used to assess neuroticism. Participants were asked
to indicate if a series of 30 statements were true for them, using a 3-
point response scale (1 = yes, 2 = no,i = uncertain). Two sample items
are “You are sometimes bubbling over with energy and sometimes very
sluggish,” and “Your mood often changes from happiness to sadness, or
vice versa, without your knowing why.” Item responses are first recoded
as needed so that higher scores indicate higher levels of neuroticism. To
recode, responses receiving a score of 1 or 2 are reversed, responses
receiving a score of 3 are left the same. An overall neuroticism score is
then computed by totaling the number of responses receiving a score of
1 (responses receiving a 2 or 3 do not contribute to the overall score).
The Emotional Stability subscale of the GZTS has been recommended
as a good measure of neuroticism by others (e.g., Costa & McCrae,
1985). Cronbach’s alpha for the present sample was .85.

Self-mastery. Self-mastery was assessed by using Pearlin and
Schooler’s (1978) Self-Mastery Scale (SMS). This seven-item instru-
ment assesses the extent to which a person generally feels as though
he or she manifests personal mastery over life outcomes (e.g., “What
happens to me in the future mostly depends on me,” and “Sometimes
I feel that I am being pushed around in life”). Its basic psychometric
properties are well established (Pearlin & Schooler, 1978), and it has
been used successfully in the past with several different community-
based populations (e.g., Pearlin & Schooler, 1978). It was also the scale
that Marshall and Lang (1990) used in their study exploring the predic-
tive power of optimism and self-mastery with respect to depression.
Cronbach’s alpha for the SMS in the present sample was .75.

Self-esteem. Rosenberg’s (1965)10-item Self-Esteem Scale (or SES)
was used to assess self-esteem. The scale, which provides a convenient
measure of global attitudes about the self, has five negatively worded
items and five positively worded items. Participants were asked to indi-
cate their agreement on a scale of 1 (strongly disagree) to 4 (strongly
agree) with statements such as “I feel I have a number of good qualities,”
and “At times, I think I am no good at all.” This scale is one of the most
widely used measures of self-esteem and has displayed good reliability
and validity (Crandall, 1973; Rosenberg, 1965). In our sample, the scale
had an internal reliability of .88.

Trait anxiety. The trait form of the State-Trait Anxiety Inventory
(STAI; Spielberger, Gorsuch, & Lushene, 1974) was used to measure
trait anxiety. This scale is composed of 20 Likert items evaluating the
extent to which the respondents experience a variety of feelings such
as happiness, self-confidence, tension, and disappointment (e.g., “I feel
content,” and “I worry too much over something that really doesn’t
matter”). The scale has been used extensively in prior psychosocial re-
search and its psychometric properties have been well documented (see,
e.g., Watson & Clark, 1984). Cronbach’s alpha for the current sample
was .89.

Depression. The Beck Depression Inventory (BDI) short form
(Beck, Rial, & Rickels, 1974) was used to assess depression. The BDI
assesses attitudes and symptoms derived from clinical observations that
are typically observed in depressed psychiatric patients but not in non-
depressed psychiatric patients (Beck, Ward, Mendelson, Mock, & Er-
baugh, 1961). The scale has been widely used and has well-established
psychometric properties (Beck, Steer, & Garbin, 1988). The short form,
13-item version used in our study assessed attitudes and symptoms

1066 M. SCHEIER, C. CARVER, AND M. BRIDGES

across several different domains, including (but not limited to) mood,
sense of failure, lack of satisfaction, social withdrawal, and indecisive-
ness. For each item, respondents were asked to choose from a group of
four statements (rated 0 to 3 in depressive symptomatology) the state-
ment that best described the way they were feeling that day. Scores for
the present sample ranged from 0 to 39 (with higher scores indicating
greater depression). Cronbach’s alpha was .87.

Physical symptoms. Self-reports of physical symptoms were as-
sessed with the Cohen-Hoberman Inventory of Physical Symptoms
(CHIPS; Cohen & Hoberman, 1983). The CHIPS comprises a list of 38
commonly experienced physical symptoms (e.g., back pain, headache,
and stuffy head or nose). Symptoms more psychological in nature (e.g.,
feeling depressed or anxious) are explicitly excluded from the list. Sub-
jects were asked to indicate how much they had been bothered by each
of the symptoms in the past 4 weeks, using a 5-point scale (1 = not at all
to 5 = extremely). In past research (Cohen & Hoberman, 1983), the
CHIPS was found to predict use of student health facilities among two
separate samples of college students (Cohen & Hoberman, 1983),
thereby supporting the construct validity of the scale.

Responses to the CHIPS were scored in two ways in our research.
First, a measure of symptom intensity was computed by summing the
degree to which subjects reported being bothered by each symptom
across all 38 symptoms. Scores for symptom intensity ranged from 38
to 162. Second, a simple tally was made of the number of symptoms for
which subjects indicated they were bothered to some degree, ignoring
the extent to which they were bothered (i.e., a score of 1 was given for
each item that received a response of 2 or more). Scores for number of
symptoms ranged from 0 to 35 in the present sample.

Coping. Coping was measured by the COPE (Carver et al., 1989).
The COPE is a 60-item, multidimension coping instrument designed to
assess 15 conceptually distinct methods of coping. The 60 items repre-
sent a large range of coping responses including (but not limited to)
active coping, positive reinterpretation and growth, seeking of social
support for emotional reasons, denial, mental and behavioral disengage-
ment, and focusing on and venting of emotions (e.g., “I do what has to
be done, one step at a time,” “I turn to work or other substitute activities
to take my mind off things,” and “I talk to someone about how I feel”).
Participants were instructed to indicate how much they usually did each
of the things that the items reflected when they encountered difficulties
or problems in their lives, using a 4-point Likert scale ( 1 = 7 usually
don’t do this at all to 4 = I usually do this a lot). Cronbach’s alpha for
the 15 scales in the current study ranged from .37 (mental disengage-
ment) to .93 (use of religion). With the exception of mental disengage-
ment, the remainder of the alphas were all above .59, with the majority
of the scales above .70. The average alpha across the 15 scales was .73.

Results

Correlations Among Predictors

The zero-order correlations among the five predictor vari-
ables are shown in Table I.1 As can be seen, all of the intercor-
relations among the predictors were significant. It is also note-
worthy, however, that the magnitudes of the correlations be-
tween the LOT and the other predictors were only in the
moderate range (the average correlation between the LOT and
the other predictors was .54, which was the lowest average cor-
relation that was obtained). This generally suggests that the
LOT had less in common with the other predictors than the
other predictors had in common with each other, with the pos-
sible exception of self-mastery, which also had a relatively low
average correlation (.56).2

Table 1
Correlations Among Predictor Variables

Variable 1

1. Optimism
r —
N

2. Self-mastery
r
N

3. Trait anxiety
r
N

4. Neuroticism
r
N

5. Self-esteem
r
N

2

.55
1,883

—

3

– . 5 9
1,420

– . 6 9
572

—

4

– . 5 0
1,692

– . 4 8
569

.74
1,423

5

.54
595

.58
624

– . 7 2
181

– . 6 6
181

Note. All correlations specified here reached significance at p < .001,
two-tailed.

Is Optimism a Predictor of Outcomes?

Zero-order correlations between predictors and outcomes are
shown in Table 2. As can be seen, all of the predictors were
moderately correlated with depression to about the same de-
gree. Depression correlated highest with self-esteem and lowest
with neuroticism. In addition, all of the predictors were signifi-
cantly correlated with the two symptom measures. For symp-
tom intensity, the highest correlations involved neuroticism
and trait anxiety. The lowest correlation involved optimism. A
similar ordering occurred for number of symptoms.

As a group, the predictor variables were also substantially
correlated with different aspects of coping. All of the predictors
correlated significantly with active coping, planning, positive re-
interpretation and growth, denial, mental disengagement, and
behavioral disengagement. Correlations between predictors and
outcomes tended to be higher for the more negative coping ten-
dencies (e.g., mental and behavioral disengagement) than for
the more positive coping tendencies (e.g., active coping and
planning).

With the exception of dispositional optimism, associations
between the predictors and the remaining coping tendencies
were more sporadic. For example, whereas four of the five pre-

1 Portions of the analyses from Study 1 have been briefly described
elsewhere (in Scheier et al., 1989, and in Scheier and Carver, 1992).

2 Subsidiary analyses conducted separately by gender suggested that
the associations between the predictors tended to be somewhat higher
for women than for men, often significantly so, given the size of the
samples. This pattern corresponds to the gender differences reported by
Scheier and Carver (1985) on a similar set of data. Subsidiary analyses
were always conducted on the data to assess the effects of gender. Very
few other gender differences emerged, however, in either Study 1 or
Study 2. For example, of the 72 partial correlations reported in Table 2,
only one involved a significant difference due to gender. The few gender
differences that did emerge seemed random in nature and were basically
uninterpretable. As a result, gender is discussed only when there are
meaningful differences to report.

REEVALUATING THE LIFE ORIENTATION TEST 1067

Table 2
Correlations Between Predictor Variables and Outcomes

Outcomes

Depression
r
N

Number of symptoms
r
N

Intensity of symptoms
r
N

Active coping
r
N

Planning
r
N

Suppression of competing
activities

r
N

Restraint
r
N

Positive reinterpretation
and growth

r
N

Use of humor
r
N

Seeking instrumental social
support

r
N

Seeking emotional social
support

r
N

Turning to religion
r
N

Acceptance
r
N

Denial
r
N

Focusing on and venting of
emotions

r
N

Mental disengagement
r
N

Behavioral disengagement
r
N

Use of drugs or alcohol
r
N

Optimism

-.42***
1,900

– . 2 1 * * *
1,015

-.25***
1,015

.30***
813

.30***
813

.14***
815

.12***
814

.47***
815

.10**
815

.16***
814

.12***
815

.22***
816

.10**
816

-.17***
815

-.10**
815

-.18***
816

.33***
816

– . 1 1 * *
816

Self-mastery

-.43***
1,306

-.27***
443

-.28***
443

.32***
375

27***
375

.09
375

.08
375

.34***
375

.10
375

.10
375

.11*
375

.08
375

.11*
375

-.17***
375

– . 0 9
375

-.23***
375

– . 4 1 * * *
375

.16**
375

Trait anxiety

-.49***
547

.47***
315

4 7 *»*
315

-.28***
394

-.17***
394

.00
396

-.19***
395

-.23***
395

– . 1 3 * *
396

-.02
396

.02
396

– . 0 3
397

-.15**
397

.32***
397

.32***
396

.34***
397

.45***
397

.27***
397

Self-esteem

-.54***
604

-.26***
443

-.27***
443

.25***
375

.20***
375

.07
375

.19*
375

.33***
375

.12*
375

.08
375

.07
375

.06
375

.04
375

-.20***
375

– . 1 0
375

-.17***
375

-.38***
375

-.08
375

Neuroticism

.41***
545

.51***
591

.54***
591

-.20***
393

-.10*
393

.02
395

-.17*
394

-.20***
394

– . 1 0
395

.05
395

.11*
395

-.05
396

-.10*
396

.26***
396

.42***
395

.42***
396

37*. *
396

.19***
396

*p< .05, two-tailed. * * p < .01, two-tailed. ***p< .001, two-tailed.

1068 M. SCHEIER, C. CARVER, AND M. BRIDGES

dictors (self-mastery, trait anxiety, neuroticism, and optimism)
were all significantly associated with drinking and substance
abuse, only optimism correlated significantly with suppression
of competing activities, seeking instrumental social support,
and turning to religion. Indeed, of the five predictors, only opti-
mism related significantly to every coping tendency that was
assessed. Undoubtedly, this was due in part to the greater num-
ber of subjects who completed the LOT, but it was also due in
part to the greater magnitude of the correlations.

Is Optimism a Unique Predictor of Outcomes?

The first four columns of Table 3 present partial correlations
between optimism and the outcome variables, controlling for
each of the other predictors in turn. As can be seen, many of the
zero-order correlations involving optimism remain significant
even after the other predictors are statistically controlled, albeit
the correlations are usually reduced in magnitude. This charac-
terization holds for associations involving depression and many
aspects of coping, including active coping, planning, suppres-
sion of competing activities, positive reinterpretation and
growth, seeking social support for instrumental reasons, and
turning to religion. Taken in isolation, none of the other predic-
tors was able to render the correlations between optimism and
these various outcomes nonsignificant.

Associations involving number of symptoms and symptom
intensity fared less well, particularly the associations involving
symptom number. That is, with respect to symptom number,
the correlation with optimism was reduced to nonsignificance
if any of the other predictors was statistically controlled. With
respect to symptom intensity, inclusion of either trait anxiety or
neuroticism reduced the correlation with optimism to near
zero. It is interesting that the correlation between optimism and
symptom intensity remained significant even when the data
were first adjusted individually for self-mastery and self-esteem.

The last column of Table 3 presents partial correlations be-
tween optimism and the outcomes, controlling for all four of
the predictors simultaneously. This obviously is a much more
stringent test of the unique predictiv

1

Church Support and Family Religious Practices as Predictors of Children’s Spirituality

Joe Student

Online and Professional Studies, California Baptist University

BEH385: Statistics and Research Methodology II

Prof. Elisabeth Knopp

April 3, 2022

2

Church Support and Family Religious Practices as Predictors of Children’s Spirituality

Over half of Americans attend church or synagogue at least once a month, most

commonly for the purpose of obtaining “spiritual growth and guidance” (Gallup, 2007, 2010).

However, some prominent believers have openly questioned the necessity of church

involvement, emphasizing instead the virtue of individual spiritual pursuits (Barna, 2005). Even

less is known about the importance of church to the spiritual and moral development of children.

In a busy world filled with homework, sports, and chores, it is important to know whether a

child’s connection to the church is truly crucial to their own spirituality, especially if the child is

already receiving consistent religious support in the home. Thus, the purpose of this study is to

investigate children’s church relationships as a predictor of children’s spirituality controlling for

family religious practices.

Spirituality, defined as the dynamic, personal, and experiential relationship between God

and child (Desrosiers et al., 2010; Simpson et al., 2009) is conceptually distinct from one’s

religious identity or religious development, which are more concerned with shared specific

practices and teachings (Richert & Granqvist, 2013). Although spirituality has been studied

infrequently compared to other aspects of child development, the relative dearth of research is

not because children do not have spiritual experiences; on the contrary, spiritual experiences

appear to be common, even among young children. Furthermore, school-age children from both

religious and non-religious homes explain that God was responsible for features in the natural

world (Evans, 2001), suggesting that children’s spiritual leanings are not solely a function of

parent instruction and leaving the door open to outside influences.

3

Literature Review

Family Religious Practices

Family religious practices are defined by the centrality of religion to the life, structures,

and processes of the family (Layton et al., 2011). Given the socializing role of the family, it

would not be surprising that salient religious routines in the family, such as praying at mealtimes

and before bed and parents teaching their children about God (Layton et al., 2011), should

influence children’s perception of their relationship with God. For example, research indicates

that children as young as three and four can start asking metaphysical questions (Harris, 2000).

Family conversation about spiritual matters and maternal spiritual support are strong predictors

of children’s experience of God (Desrosiers et al., 2010; King et al., 2002). Guided participation

(Rogoff, 1995) in family religious practices may not only provide a foundation for explicit

religious beliefs, but also provide opportunities for increased experiences of connection to God.

Therefore, to properly understand the role of church support in predicting children’s spirituality,

it is necessary to control for the likely influence of family religious practices.

Church Support

Church support is defined by the provision of love, empathy, caring, and trust by

coreligionists (Krause & Ellison, 2009), a sense of community, and the feeling of family when

one is around members of their church (Layton et al., 2011). Church support is theorized to

impact spiritual formation because one’s connection to God is developed and maintained through

reciprocal spiritual support within the church community, and one’s perception of God is

influenced by those relationships with others (Krause & Ellison, 2009). Prior research has found

significant correlations between adults’ spirituality and the quality of their relationships with

other people in the church (Krause & Ellison, 2009; Simpson et al., 2009; Winseman, 2005), and

4

limited research with adolescents has also linked positive spiritual development to positive,

impactful relationships with a church youth pastor (Strommen & Hardel, 2000).

As existing research on the relationship between church support and spirituality has been

conducted only with adults and adolescents, the current study extends the literature by

investigating whether church support predicts children’s spirituality. The hypothesis guiding this

study is that church support will positively predict children’s spirituality, controlling for the

influence of family religious practices.

Method

Participants

Participants were 127 sixth- and seventh-grade students (43 boys and 84 girls) enrolled in

coeducational Brisbane Catholic Educational primary schools. Australian schools typically do

not have middle schools. The sample was evenly split according to grade level with 62 sixth-

grade students and 65 seventh-grade students. The age of students ranged from 10 to 13 years

(M = 11.42, SD = .60). Of the students sampled, 89.8% reported that they had been born in

Australia, and 87.4% reported speaking solely English at home. In Australia, Catholic Schools

typically have a low fee structure and draw students from a similar range of socioeconomic

backgrounds as the local Government schools in the same geographic areas.

Measures

Religious Preference. Respondents were asked to indicate their religious preference.

Response options included Protestant Christian (38%), Catholic Christian (14%), Other (36%),

and None (12%).

Depression. Children’s depression symptoms were measured using the Children’s

Depression Inventory (CDI), a six-item self-report measure of depressive symptomology in

children and adolescents (Kovacs, 1992). Participants respond to items indicating a depressed

state (e.g., “I am often sad”) on a 3-point Likert-type scale ranging from 1 (not at all true) to 3

(very true). Evidence of strong internal reliability of the CDI has been found in previous studies,

with Chronbach’s alphas ranging from .71 to .89 (Reynolds, 1994). Children’s self-report CDI

scores are positively correlated with parent-report measures, as well as self-report measures of

loneliness, thus supporting the validity of the measure (Carey, Faulstich, Gresham, Ruggiero, &

Enyart, 1987).

Factor analysis. An exploratory factor analysis (EFA) was conducted in order to

examine the factor structure of the six-item scale.

Only one factor had an eigenvalue greater than 1, thus suggesting a one-factor solution

according to Kaiser’s K-1 rule. The first factor explained 51.66% of the variation in scores. A

scree test was also conducted in order to determine whether the scale represented a

unidimensional construct. The scree plot was somewhat ambiguous, with the point of inflexion

being at either the second or third component, thus suggesting a one- or two-factor solution.

Next, the component matrix was examined to see whether all of the items loaded heavily onto

the first factor. All six items had loadings with absolute values over .40 on the first factor.

Reliability analysis. Next, a reliability analysis was conducted in order to determine the

internal reliability of the depression scale. This analysis produced a Chronbach’s α of .80, which

was good for psychometric purposes.

Distribution of composite scores. Finally, a composite score was created from the six-

item scale, generating a standardized (M = 0.00, SD = 1.00) depression score for each child in the

sample, ranging from -1.01 to 3.57. The resulting distribution was leptokurtic with a moderate

positive skew. Most of the respondents scored on the low end of the depression inventory with

only a few indicating high levels of depression.

Variable 3. Repeat the above process for your second composite variable.

Procedure

The students were administered the questionnaires in combined class groups, such that all

participating students from a particular school were presenting a single location in the school at

the time of the administration. The questionnaires were administered by graduate psychology

students, with classroom teachers also in attendance. The administration took place during

regular class time, and students were encouraged to work privately in answering questionnaires,

as if undertaking an exam.

Running head: GENDER AND STRESS AS PREDICTORS OF DEPRESSION 1

Gender and Stress as Predictors of Depression

Ana Marie Agee Booth

California Baptist University

Gender and Stress as Predictors of Depression 2

Gender and Stress as Predictors of Depression

Over 17 million adults in the United States suffer from some sort of depression, making it

one of the most common mental illnesses in America. “Depression affects an estimated one in 15

adults (6.7%) in any given year. And one in six people (16.6%) will experience depression at

some time in their life.” (What is Depression?). There are an overwhelming number of factors

that can lead to depressive symptoms in both men and women, one of which is said to be an

elevation in stress hormone disturbances. Both genders rely on different coping methods to

process stressors effectively and successfully in a manageable method. Thus, the purpose of this

study is to investigate gender as a predictor of depression controlling for stress.

Depression is defined as a common and serious medical illness that negatively affects how you

feel, the way you think and how you act and causes feelings of sadness and/or a loss of interest in

activities you once enjoyed. It can lead to a variety of emotional and physical problems and can

decrease your ability to function at work and at home (What is Depression?). Doctors and

scientists have studied depression for many years. They have conducted research on many

individual symptoms as well as clustered symptoms, in addition to the plethora of factors that

may be the cause of developing depressive symptoms. Many of the research studies identify

participants according to their gender and then their ages. For example, a recent study indicated

that depressive symptoms based on life stressors of medical interns ages twenty-six to thirty,

increased by 71% over a 3-, 6-, 9-, and 12-month time period (Fried, 2015). This particular study

consisted of a total of three thousand and twenty-one individuals, with 48.4% being male and

52.6% being female. Furthermore, both males and females explain their depressive symptoms

almost equally in terms of life stress as a causal factor of major depressive disorder, suggesting

that gender may not be a significant factor when correlating stress and depression.

Gender and Stress as Predictors of Depression 3

Literature Review

Stress. Stress is defined as the degree to which you feel overwhelmed or unable to cope

as a result of pressures that are unmanageable (Stress MHF, 2021). At the very least, stress is our

body’s response to specific life events and/or situations (Sha, 2006). Social circumstances,

economic circumstances, environment, and genetics are all contributors to a person’s level of

stress. As the many contributions of stress can vary immensely from person to person, male and

female, it also varies in how responses are processed. Given the fact that positive and negative

life changes both have the ability to cause one to experience some form of stress, it would not be

surprising if stress factors such as divorce, job loss, death, or financial difficulties led to feelings

of depression. Stress factors are managed very differently among men and women. Both genders

typically report similar stress levels, yet report varying physical and emotional symptoms (Shih,

2004). Many of the symptoms of depression have a rather similar look as having life changing

stressors. Therefore, in order to properly understand the role of gender in predicting depression,

it is necessary to control for the likely influence of stressors.

Gender. The World Health Organization defines gender as the characteristics of women,

men, girls and boys that are socially constructed including norms, behaviors and roles associated

with being a woman, man, girl or boy, as well as relationships with each other. In a socially

constructive manner, gender will vary from environment to environment and can change over

time. Gender is theorized to impact depression because one’s stress hormones and one’s genetics

offer largely influential factors that contribute to a mental illness such as depressive disorder.

Recent research has found that women are more likely than men to experience depression

(Zwicker & DeLongis, 2010). “Some studies show that one-third of women will experience a

major depressive episode in their lifetime.” (Hyde, 2020). There is a high degree of heritability

Gender and Stress as Predictors of Depression 4

(approximately 40%) when first-degree relatives (parents/children/siblings) have depression

(Sowa & Lustman, 1984).

As existing research study on the relationship between gender and depression has been

conducted primarily focused on age groupings, the current study will extend the literature by

investigating whether gender predicts depression. The hypothesis is that gender will significantly

predict depression controlling for the influence of stressors.

Method

Participants

Participants were 101 adult college students enrolled in the Online and Professional

Studies division at California Baptist University (CBU). CBU is a private Christian university

located in Riverside, California with an enrollment of approximately 11,317 students on campus

and online. The age of the adult college student participants ranged from 18 to 54 years

(M=28.29, SD=7.41). Of the 101 participants, 88 were female (87%) and 12 were male (12%).

The Online and Professional Studies division at California Baptist University offers more than

40 fully online and fully accredited degree programs while also providing spiritual and social

development opportunities.

Measures

Sex/Gender. Respondents were asked to indicate their gender. Response options

included female (87%) and male (12%).

Stress. Adult college student’s levels of stress were measures using the DASS-21 Stress

Subscale, a four item self-report measure of stress levels in adult college students. Participants

Gender and Stress as Predictors of Depression 5

responded to items indicating a stressed state (e.g., “I feel that I use a lot of nervous energy”) on

a 7-point Likert-type scale ranging from 1 (does not apply to me at all) to 7 (most of the time).

Evidence of moderate to strong internal reliability has been found in previous studies, with

Chronbach’s alphas ranging from .59 to .81 (Osman et al. 2012). DASS-21 Stress subscale scores

are positively correlated with scores on a measure of mixed depression and anxiety/stress, thus

supporting the validity of the measure.

Factor analysis. An exploratory factor analysis (EFA) was conducted in order to

examine the factor structure of the four-item scale.

Only one factor had an eigenvalue greater than 1, thus suggesting a one-factor solution

according to Kaiser’s K-1 rule. The first factor explained 57.57% of the variation in scores. A

scree test was also conducted in order to determine whether the scale represented a

unidimensional construct. The scree plot shows the point of inflection clearly at the second

component, thus suggesting a one-factor solution. Next, the component matrix was examined to

see whether all of the items loaded heavily onto the first factor. All four items had loadings with

absolute values over .40 on the first factor.

Gender and Stress as Predictors of Depression 6

Reliability analysis. Next, a reliability analysis was conducted in order to determine the

internal reliability of the stress level scale. This analysis produced a Chronbach’s alpha score of .

75, which was acceptable for psychometric purposes.

Distribution of composite scores. Finally, a composite score was created from the four-

item scale, generating a standardized (M=0.00, SD=1.00) stress level score for each adult college

student in the sample ranging from -1.52 to 2.71. The resulting distribution was slightly

leptokurtic with a moderate positive skew. Most of the respondents scored on the low to middle

end of the stress level scale with only a few indicating high levels of stress.

Gender and Stress as Predictors of Depression 7

Depression. Adult college student’s levels of depression were measured using the DASS-

21 Depression Subscale, a four item self-report measure of depression levels in adult college

students. Participants responded to items indicating a depressed state (e.g., “I feel that I have

nothing to look forward to”) on a 7-point Likert-type scale ranging from 1 (does not apply to me

at all) to 7 (most of the time). Evidence of moderate to strong internal reliability has been found

in previous studies, with Chronbach’s alphas ranging from .63-.70 (Osman et al. 2012). DASS-

21 depression subscale scores are positively correlated with scores on a measure of mixed

depression and anxiety/stress, thus supporting the validity of the measure.

Factor analysis. An exploratory factor analysis (EFA) was conducted in order to

examine the factor structure of the four-item scale.

Gender and Stress as Predictors of Depression 8

Only one factor had an eigenvalue greater than 1, thus suggesting a one-factor solution

according to Kaiser’s K-1 rule. The first factor explained 71.74% of the variation in scores. A

scree test was also conducted in order to determine whether the scale represented a

unidimensional construct. The scree plot shows the point of inflection clearly at the second

component, thus suggesting a one-factor solution. Next, the component matrix was examined to

see whether all of the items loaded heavily onto the first factor. All four items had loadings with

absolute values over .40 on the first factor.

Gender and Stress as Predictors of Depression 9

Reliability analysis. Next, a reliability analysis was conducted in order to determine the

internal reliability of the depression level scale. This analysis produced a Chronbach’s alpha

score of .86, which was good for psychometric purposes.

Distribution of composite scores. Finally, a composite score was created from the four-

item scale, generating a standardized (M=0.00, SD=1.00) depression level score for each adult

college student in the sample ranging from -.81 to 3.48. The resulting distribution was

leptokurtic with a positive skew. Most of the respondents scored on the low end of the depression

level scale with only a few indicating high levels of depression.

.

Gender and Stress as Predictors of Depression 10

Procedure

The adult college students were administered the questionaires anonymously. All of the

students, majoring in psychology, were currently enrolled in the Online and Professional Studies

division of California Baptist University. The surveys were completed online, outside of

classroom hours and each student was given extra credit for completion of the online

questionaire.

Results

Descriptive Statistics

First, figures were constructed in order to visually examine the relationship between each

independent variable and the dependent variable. A boxplot was used to examine the

relationship between sex and depression because sex is a categorical predictor. A scatterplot was

used to examine the relationship between stress and depression because stress is a continuous

predictor.

A visual inspection of the boxplot revealed that the median depression level for females

appears to be similar to that of the males. There is more variability among the male scores than

Gender and Stress as Predictors of Depression 11

the female scores. Some males scored relatively high, while the females scored somewhat low

with several potential outliers and otherwise unusual responses. A visual inspection of the

scatterplot revealed that there appeared to be a slight positive relationship between stress levels

and depression. In most cases, as stress levels increased, depression also increased. However,

some college students scored low on stress levels and also low on depression levels.

Next, a correlation test was conducted to examine the relationships between each

independent variable and the dependent variable. There was a significant moderate positive

correlation between stress and depression, r(96) = .49, p<.01. There was also a very small

positive correlation between gender and depression r(96) = .06, p=.59. The two independent

variables were not significantly correlated; thus, collinearity was not an issue in this model.

Gender and Stress as Predictors of Depression 12

Checking Assumptions

Homogeneity of variance. In order to test the assumptions, the analysis was run with

both independent variables included in the model. Predicted values and unstandardized residuals

were saved as new variables in the dataset. To test for homogeneity of variance, a scatterplot

was created with the predicted values on the x-axis and the residuals on the y-axis.

A visual inspection of the scatterplot revealed that the residuals are not evenly distributed

and there appears to be clusters and some slight fanning. This means that the model does not

have equally predictive power for individuals at both the high end and the low end of the

depression level scale. Thus, the assumption of homogeneity of variance is not supported.

Gender and Stress as Predictors of Depression 13

Normally distributed residuals. In order to determine whether residuals were normally

distributed, a histogram of the distribution was created with a normal curve superimposed on top.

A visual inspection of the histogram revealed that the residuals are not normally distributed.

Thus, the assumption of normally distributed residuals is not supported.

Multiple Regression Analysis

Finally, each of the output tables in the multiple regression analysis was interpreted in

order to determine whether my hypothesis was supported.

Gender and Stress as Predictors of Depression 14

The R2 coefficient was .26. This means that 26% of the variation in adult college

student’s depression levels are predicted by the model, while 74% of the variation was not

predicted by the model. The amount of predictive ability was determined to be substantively

significant.

The F-statistic was statistically significant, F(2, 93) = 16.28, p < .001, indicating that the

model results in significantly better predictions than those based solely on the mean.

This table reveals that stress level is a statistically significant predictor of adult college

student’s depression levels. Gender is not a statistically significant predictor of adult college

student’s depression levels, controlling for stress. Thus, my hypotheses is not supported.

Gender and Stress as Predictors of Depression 15

The regression equation for this model is:

ŷ = -.03 – .51×1 + .41×2

Where .03 is the expected depression level score for a student who has an average level of stress.

Controlling for stress, there is a .10 standard deviation decrease in gender which does not

correspond to a .03 standard deviation in depression levels.

Discussion

Based on the results, gender was not a significant predictor of depression, controlling for

stressors and my hypothesis was rejected thus being disconfirmed. This study consisted of 101

participants, 89 being female and 12 being male. This makes for an unbalanced participant pool,

especially when using gender as an independent variable. A more equally balanced selection of

participants would have led to significantly better and more accurate results. Although study

findings are certainly accurate, there are some limitations. First, this study used a wide range of

adults ages 18-54. I consider this a very strong limitation as young adults, whether male or

female, process stressors more immaturely than adults who have more life experience.

The results of this study correspond with the results in the 2015 study conducted by

Fried, in which males and female explained their depressive symptoms almost equally and

suggesting that gender may not be a significant factor when correlating stress and depression.

This study, in particular, can easily be generalized to the college and university

population by simply inferring the results from a sample population and applying it to the

population at large: Increasing the sample size to a population significantly larger than 101

college students and making the study more equal as far as gender.

Gender and Stress as Predictors of Depression 16

Many theorize that females have more life stressors and therefore suffer from depression

more than men, but these theories are just that and have not been shown to possess any statistical

significance.

Based on the results of this study, future researchers should follow a different

methodology to obtain more accurate and precise findings. Future researchers would benefit

from selecting an equal amount of male and female participants as well as separating the

participants into age groups. This will better serve additional future studies and set a more

specific path for future research to follow.

My data-based conclusion disputed my speculation that gender would play a large role in

stress and depression. In fact, this study somewhat drastically negated my initial hypothesis. I

strongly felt that women would definitely have more stressors and therefore higher levels of

depression. This was not in line with my referenced study or the present study. My overall

impression of this research study was actually quite surprising and I would be interested in

continuing it with a more specific range of participants to determine if there is any possibility of

truth to my initial investigation and research.

Gender and Stress as Predictors of Depression 17

Works Cited

Fried, E. I., Nesse, R. M., Guille, C., & Sen, S. (2015). The differential influence of life stress
on individual symptoms of depression. Acta Psychiatrica Scandinavica, 131(6), 465-
471. https://doi.org/10.1111/acps.12395

Hyde, J. S., & Mezulis, A. H. (2020). Gender differences in depression: Biological, affective,
cognitive, and sociocultural factors. Harvard Review of Psychiatry, 28(1), 4-
13. https://doi.org/10.1097/HRP.0000000000000230

Sha, T. (2006). Optimism, Pessimism and Depression; The Relations and Differences by Stress
Level and Gender. Acta Psychologica Sinica, 38(6), 886-901.

Shih, J. H.-F. (2004). Sociotropy/autonomy and depression: Gender differences and the
mediating role of stressful life events ProQuest Information & Learning]. APA
PsycInfo. http://libproxy.calbaptist.edu/login?
url=https://search.ebscohost.com/login.aspx?direct=true&db=psyh&AN=2004-99010-
099&site=ehost-live&scope=site

Sowa, C. J., & Lustman, P. J. (1984). Gender differences in rating stressful events, depression,
and depressive cognition. Journal of Clinical Psychology, 40(6), 1334-
1337. https://doi.org/10.1002/1097-4679(198411)40:6<1334::AID-
JCLP2270400609>3.0.CO;2-8

Stress. Mental Health Foundation. https://www.mentalhealth.org.uk/a-to-
z/s/stress#:~:text=Stress%20can%20be%20defined%20as,of%20pressures%20that
%20are%20unmanageable.

What is Depression? American Psychiatric Association. https://www.psychiatry.org/patients-
families/depression/what-is-depression

Zwicker, A., & DeLongis, A. (2010). Gender, stress, and coping. In J. C. Chrisler & D. R.
McCreary (Eds.), Handbook of gender research in psychology, Vol 2: Gender research
in social and applied psychology. (pp. 495-515). Springer Science + Business
Media. https://doi.org/10.1007/978-1-4419-1467-5_21