.. ...... ....
. up v1.r.'54.~r
A v , lot. Tet r
.u. .N. ....mn.._:..édra.
What": at»; ..v...%
..3 , ,v .. m».

; nan...

 

h. a A... A” .
a... it

13.3.25? H .. v. n
. ..3 ‘nt . . O.
1” ....laoval‘i- b1 . ”Ann“

.
.v 1.
. !

 
 

..L 13%

.1'

._, m... ”m.-.

I": ‘
MW}

3..

.x r 3m“?

n.
2

.
M
A.I

-
‘v

' “ ‘3 :u'r'r‘
I figg'fl'n 1:11., -
J ,
. ‘, 3.4"“
.. 1

H. .l..v-| r. O.

' .
int

2 ‘1... id‘s-.... I. f

 

Ewifiéfigg

l'.

 

a: , 3?

1t :1.

 

igmwmfivufgufifiw ,

1
$

‘ll

 

THESIb

2 Illllllllllllllllflllllllll‘IHJIlllllllllllllllllllll

1293 01714 1429

This is to certify that the

dissertation entitled

Coping with the stigma of AIDS: An investigation
of the effects of shame, stress, control, and
coping on depression in HIV-positive and -negative
gay men

presented by

Frank Joseph DeMarco

has been accepted towards fulfillment
of the requirements for

Ph. D. degree in Psychology

 

 

W/W

Major professor

r/ 7/ f?

MSU is an Affirmatiw Action /Equal Opportunity Institution 0- 12771

 

 

 

LIBRARY

Mlchlgan State
University

 

 

 

PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

 

DATE DUE MTE DUE DATE DUE

 

      

 

 

‘ Jr“. -’ duo
UEIEiBz 6 2002
lf‘P-Rl 109 029506

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

use WWpGS-p.“

 

.56
.N) C

 

 

 

. COPING WITH THE STIGMA OF AIDS:
AN INVESTIGATION OF THE EFFECTS OF SHAME, STRESS, CONTROL
AND COPING ON DEPRESSION IN HIV-POSITIVE AND NEGATIVE GAY MEN
By

Frank Joseph DeMarco

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Psychology

1998

 

 

AW

 

537:1er

ru- 1‘
”one:

ABSTRACT
COPING WITH THE STIGMA OF AIDS:
AN INVESTIGATION OF THE EFFECTS OF SHAME, STRESS, CONTROL
AND COPING ON DEPRESSION IN HIV-POSITIVE AND NEGATIVE GAY MEN
By

Frank Joseph DeMarco

The current study examines how HIV serostatus, internalized shame, physical
symptoms, stress, appraisals of control, and type of coping used (detachment versus
involvement) affect depression in a sample of HIV-positive (N = 50; 31 with AIDS
diagnosis) and HIV-negative (N = 57) gay men, as well as in a comparison sample of
heterosexual‘male college students (N = 112). Multivariate path analyses indicated that,
among gay men, being HIV -positive was associated with increased shame (b = .09),
which in turn was associated with increased use of detachment (avoidance) coping (b =
.51) and decreased use of involvement (active) coping (b = -.30). Increased stress was
associated with greater use of both involvement (b = .53) and detachment (b = .49)
coping. Involvement coping was associated with decreased depression (b = -.49) while
detachment coping was associated with increased depression (b = 1.04). Coping
explained 76% of the variance in depression, and the model fit the data well (chi square
11 df = 15.85, p = .147). Implications for clinical practice and suggestions for future

research are discussed.

This work is dedicated to my wife, Christine Edgar,
and in loving memory of Michael LeRoux.

iii

3:11 I.
053 r.
rese.

‘0 '12»:
I0 30‘?
IIIL'ne

alu-a‘,

 

ACKNOWLEDGMENTS

This dissertation represents the culmination of many years of study and toil, and is
an undertaking that would not have been possible without the kindness,.help, and support
of a great many people.

First and foremost I would like to thank the men who agreed to participate in my
research study. I am especially gratefiil to the persons living with HIV/AIDS who were
. generous enough to share their experiences with me. I

This study would have been far more difficult (if not impossible) to conduct
without the help of the Lansing 'Area AIDS Network (LAAN). In particular I would like
to thank former LAAN director Bill Bathie for believing in my project and allowing me
to isolicit research participants from among LAAN’s clientele, and Carrie Tarry for her
immeasurable help with the data collection and for taking my project on as her own. I will
always be grateful for your help and support.

My dissertation committee was carefully chosen not only for their scholarly
abilities, but for their human qualities as well. Each of these men has been an enormous
help to me throughout my graduate career at Michigan State. I would like to thank Bob
Caldwell, my chair, for his mentoring, patience, and kindness throughout this arduous
process; Gersh Kaufman, for inspiring this research and for his kindness and

encouragement throughout; Bert Karon, for keeping me inspired through his passion

iv

about psychology in general and psychoanalysis in particular; and Ralph Levine, for
agreeing at the last minute to join my committee, and for teaching me the “right” way to
do statistical analysis.

The Psychology Department is fortunate to have excellent administrative and
support staff who are critically important to helping students negotiate the sometimes
treacherous waters of the graduate school process. Many people have helped me
throughout my years at Michigan State, but I would eSpecially like to mention Suzy
Pavick (“Goddess of Knowledge and Keeper of the Chocolate”), Cheryl Forcia, and
Roger Halley. Thank you for your help and for making this a more humane process for
graduate Students.

There have been many other people whose friendship, mentoring, or support has
meant a great deal to me. throughout this process. I would just like to mention a few of
them: Brent Bilodeau, Jennifer Chabot, Georgia Davidson, Mary Edens, Jim F uendeling,
Lisa Gauthier, Lauren Harris, Debbie Kolk, Jon Lacey, Lisa Laughman, David
Lounsbury, David Ostrow, Denise Rajala, Peter Reilly, and Michael Teixeira.

I must thank my family for always believing in me and for their love, patience and
. support throughout a VERY long process. I love you all very much.

Last, but by no means least, I must acknowledge my spouse, Christine Edgar, for
her incredible patience and unconditional love and support throughout my entire graduate
school career. You’re a dissertation widow no more, sweetie!

This work was supported in part by a $3,000 Blue Cross/Blue Shield of Michigan

Foundation Student Award Program Grant (grant # 191-SAP/96).

 

 

L181 OF TABL
HST. 0i FIG!
NRODL'CTM
“A .\lE.\'T.1
LUEPATL'RE
THE PSY
DOES H1
AIDS A?
FACTO
VENT.
Ris
HI
PAC]
MEX

S

TABLE OF CONTENTS

LIST OF TABLES ..................................................................................................... xi

' LIST OF FIGURES .......................................................................................................... xiii
INTRODUCTION ................................................................................................................ 1
“A MENTAL HEALTH CATASTROPHE” ................................................................. 2
LITERATURE REVIEW, STUDY RATIONALE, AND HYPOTHESES ........................ 4
THE PSYCHOLOGICAL IMPACT OF AIDS ............................................................. 4

DOES HIV SEROSTATUS AFFECT MENTAL HEALTH? FINDINGS IN THE
AIDS AND MENTAL HEALTH LITERATURE ......................................................... 5

FACTORS MEDIATING THE RELATIONSHIP BETWEEN HIV/AIDS AND

MENTAL HEALTH: 1. RISK GROUP MEMBERSHIP & HIV SYMPTOMS ......... 11
Risk Group Membership ........................................................................................ 11
I HIV-Related Symptoms ......................................................................................... 12

FACTORS MEDIATING THE RELATIONSHIP BETWEEN HIV /AIDS AND

MENTAL HEALTH: II. STRESS, CONTROL, AND COPING ................................ 14
Stress And Cognitive Appraisal ................................................................... ' .......... 14
Appraisals Of Control ............................................................................................ 16
Coping .................................................................................................................... 16
Perceived Social Support ....................................................................................... 19

Vi

 

EXTEX

THE RC

Sta:
Phcr
Sha:
"AX EP
AN OI
VARIA
HYPOI

BIV

 

 

 

Cognitive Theory Of Emotions .............................................................................. 20

An Application: Folkman and Colleagues’ Study of Stress, Control, and Coping

among Gay Men in San Francisco ......................................................................... 20
EXTENDING THE STRESS AND COPING MODEL .............................................. 28
THE ROLE OF SHAME .............................................................. ....................... ...30

Affect Theory ......................................................................................................... 31

The Affect Of Shame ............................................................................................. 33

Shame Vs. Guilt ..................................................................................................... 34

Phenomenology Of Shame ..................................................................................... 34

Shame And Mental Health ..................................................................................... 35
“AN EPIDEMIC OF STIGMA” .................................................................................. 36
AIM OF THE PRESENT STUDY .............................................................................. 38
VARIABLES .............................................................................................. - ................. 4 O
HYPOTHESES ............................................................................................................ 40

Bivariate Hypotheses: ............................................................................................ 40

MultiVariate Hypotheses (Path Models): ............................................................... 42

METHOD .......................................................................................................................... 45
RECRUITMENT OF PARTICIPANTS ...................................................................... 45

Statistical Power Analysis ...................................................................................... 45

Gay Male Community Sample ............................................................................... 46

HIV-Positive Gay Male Sample ............................................................................. 47

Compensation for Gay Male Participants .............................................................. 48

Heterosexual Male College Student Sample .......................................................... 48

Vii

 

PROI

NIEA

RESL'L'

PRE

SC;
CH.

RE]

TEE

PROCEDURE .............................................................................................................. 49

MEASURES ................................................................................................................ 50
Demographic Information........................................: ............................................. 50
HIV Serostatus ....................................................................................................... 50
HIV Symptoms ............................................................ 51
Stress .......................................................... V ............ 1 ................................................ 51
Appraisals of Control ............................................................................................. 51
Stress and Control Quotients ................................................................................. 52
Coping Strategies ................................................................................................... 53
Intemalized Shame ................................................................................................. 57
Depression .............................................................................................................. 58

RESULTS .......................................................................................................................... 59

PRESENTATION OF RESULTS ............................................................................... 5 9
Confidence Intervals .............................................................................................. 59
Inference Probabilities and Odds Ratios ................................................................ 60
The d-Statistic for the Difference in Means ........................................................... 61
Correcting for Attenuation ..................................................................................... 61

SCALE PSYCHOMETRICS ....................................................................................... 62

CHARACTERISTICS OF THE STUDY SAMPLES ................................................. 66

RELATIONSHIPS BETWEEN DEMOGRAPHIC FACTORS AND SUBSTANT IV E
VARIABLES ............................................................................................................... 76

TESTS OF BIVARIATE HYPOTHESES .................................................. ' ................. 84

. viii

 

H)

h"
C

H)

5p

 

H)

“1

im

“3

SI?

H:

in‘

H:

do

“I

Sh

H:

WI

H:

Wi

Hypothesis 1: The gay sample will evince greater internalized shame than the
heterosexual sample. ........... . ................................................................................... 84
Hypothesis 2: The gay sample is expected to manifest more depressive

, symptomatology than the heterosexual sample ...................................................... 86
Hypothesis 3: Within the gay sample, HIV-positive serostatus -will be associated
with greater internalized shame. ............................................................................ 87
Hypothesis 4: More physical symptoms will be associated with greater
internalized shame. ................................................................................................ 88
Hypothesis 5: More physical symptoms will be associated with greater overall
stress level. .............................................................................................. .............. 90
Hypothesis 6: Greater perceived control will be associated with increased use of
involvement coping ........................ 92
Hypothesis 7: Lower perceived control will be associated with increased use of
detachment coping. ....... ' ......................................................................................... 9 4
Hypothesis 8: Lower perceived control will be associated with greater internalized
shame. ................................................................................. 96
Hypothesis 9: Greater use of involvement coping will be negatively associated
with depression. ..................................................................................................... 97 I
Hypothesis 10: Greater use of detachment coping will be positively associated
with depression. ..................................................................................................... 98
Hypothesis 11: Greater internalized shame will be positively associated with
depression. ............................................................................................................. 99

MULTIVARIATE ANALYSES ................................................................................ 101

ix

 

 

 

 

SLPPI
DISCL'SSI
METH

Sir

 

Lad
Liri

Sc;

SUPPLEMENTAL ANALYSIS ................................................................................ 110

DISCUSSION ....................................................... 116
METHODOLOGICAL CONSIDERATIONS ........................................................... 116
Sampling Issues .................................................................................................... 116
Lack of HIV Information for the Heterosexual Sample.............. ......................... 1 17
Limitations Of Cross-Sectional Data .................................. ....................... 117
Scale Psychometrics ............................................................................................. 117
Possible Confounding of Shame and Depression ................................................ 118
DISCUSSION OF RESULTS .................................................................................... 119
Sexual Orientation, HIV, and Intemalized Shame ............................................... 1.19
Sexual Orientation, HIV, and Depression ............................................................ 120
Physical Symptoms, Stress, and Intemalized Shame ........................................... 120
Appraisals of Control and Intemalized Shame .................................................... 121
Stress and Coping ................................................................................................ 122
Appraisals of Control and Coping ....................................................................... 123
Coping and Depression ........................................................................................ 123
Intemalized Shame and the Stress and Coping Model ........................................ 124
IMPLICATIONS FOR CLINICAL INTERVENTION ............................................. 125
SUGGESTIONS FOR FUTURE RESEARCH... ...................................................... 126
APPENDICES ................................................................................................................. 129
REFERENCES ................................................................................................................ 172 i

 

 

Ca.

cc
I2
Ta
Ta

Ia
to;

I31
inc
Tat

inc:

Ifib
Sam

Tab

Usec

LIST OF TABLES

Table L—l - Bivariate correlations among variables in the path analysis (from Folkrnan et

a1., 1993 ............................................................................................... 24
Table 1 - Items comprising the Ways of Coping subscales .................................... 55
Table 2 - Results of confirmatory factor analysis of the scales used in the study 63
Table 3 - Descriptive statistics for the study samples: demographic variables .............. 67
Table 4 - DeScriptive statistics for the study samples: HIV/AIDS testing and

diagnoses ............................................................................................. 69
Table 5 - Descriptive statistics for the study samples: physical symptoms, stress, control,
coping, shame, and depression ..................................................................... 71
Table 6 - Comparison of the study samples by ethnicity (collapsed categories). . . . . . . . ....72
Table 7 - Comparison of the study samples by education ...................................... 73
Table 8 - Comparison of the study samples by income ........................................ 73

Table 9 - Comparison of the study samples: age, physical symptoms, stress, control,
coping, shame, and depression ..................................................................... 74

Table 10 - Correlational analysis between demographic variables (age, education,
income) and substantive variables (gay sample) ................................................ 78

Table 11 - Correlational analysis between demographic variables (age, education,
income) and substantive variables (heterosexual sample) .................................... 80

Table 12- Crosstabulation of ethnicity (collapsed categories) by HIV serostatus (gay
sample) ................................................................................................ 81

Table 13 - Differences between ethnic groups (collapsed categories) on variables to be
used in the path analyses (gay sample) ........................................................... 82

xi

 

 

 

 

 

 

 

Table 14
used in tl

Table 15
shame :11:

Table 16
the heterc

able l"

9%
shame ma

Table 18 -
With inten

Table 19 -

With stress

Table 20 -

lm'OiVeme

Table 21 -
detachmer

Table 23 -
Shame .....

Table 23 -

Iiihle 24 -
COTTeIaIed

Table 25 .
depression

Table 20 _
121131.27 .
Tab1628 .

Table 29 -

 

 

Table 14 - Differences between ethnic groups (collapsed categories) on variables to be
used in the path analyses (heterosexual sample) ................................................ 83

Table 15 - Test of bivariate hypothesis 1: The gay men will evince greater internalized
shame than the heterosexual men ................................................................. 84

Table 16 - Test of bivariate hypothesis 2: The gay men will evince greater depression than
the heterosexual men ................................................................................ 86

Table 17- Test ofbivariate hypothesis 3: HIV + gay men will evince greater internalized
shame than HIV- gay men .......................................................................... 87

- Table 18 - Test of bivariate hypothesis 4: Physical symptoms will be positively correlated
with internalized shame ............................................................................ 88

Table 19 - Test of bivariate hypothesis 5: Physical symptoms will be positively correlated
with stress ............................................................................................. 90

Table 20 - Test of bivariate hypothesis 6: Control will be positively correlated with
involvement coping ................................................................................. 92

Table 21 - Test of bivariate hypothesis 7: Control will be negatively correlated with
detachment coping .................................................................................. 94

Table 22 - Test of bivariate hypothesis 8: Control will be negatively correlated with
shame ................................................................................................. 96

Table 23 - Test of bivariate hypothesis 9: Involvement coping will be negatively
correlated with depression ........................................................................ 97

Table 24 - Test of bivariate hypothesis 10: Detachment coping will be positively
correlated with depression ........................................................................ 98

Table 25 - Test of bivariate hypothesis 11: Shame will be positively correlated with

depression ............................................................................................ 99
Table 26 - Summary of results of bivariate tests ............................................... 100
Table 27 - Individual link analyses for Path Model 1 (gay sample) ........................ 103
Table 28 - Individual, link analyses for Path Model 2 (heterosexual sample) ........ i ..... 108

Table 29 - Correlations between coping preference and other variables in the study. . ...l 13

xii

LIST OF FIGURES

Figure L-l - Path analysis from Folkman et a1. (1993) ......................................... 25
Figure 1 - Path model for gay sample (HIV status known) .................................... 44
Figure 2 - Path model for heterosexual sample (HIV status unknown) ...................... 44
Figure 3 - Results of Path Model 1 (gay sample) .............................................. 102
Figure 4 - Results of Path Model 2 (heterosexual sample) ................................... 107

xiii

 

 

were re}:
World Ii
immune:
sobering
PTOPQIIIC

Ir
Of these c
rePorted 1'
African A
demOnstr
CDC. W}
OIAIDS \

199T),

 

T}
CHOImOUS
SInCe the

than the r.

 

INTRODUCTION _

In 1981, the frrst deaths related to acquired immune deficiency syndrome (AIDS)
were reported by the Centers for Disease Control and Prevention (CDC). Today, the
World Health Organization (WHO) estimates the worldwide prevalence of human
immunodeficiency virus (HIV) infection to be around 30.6 million people, a truly
sobering figure (WHO, 1997). This is clearly a worldwide epidemic of staggering
proportions.

In the US alone, 612,078 AIDS cases have been reported to date; 379,258 (62%)
of these eases have ended in death (CDC, 1997). Of these US AIDS cases, 7,902 were
reported in children under thirteen, 91% of whom incurred the disease perinatally.
African Americans, Latinos, and intravenous (IV) drug users and their partners
demonstrate the fastest-growing rates of new HIV infection in the US, according to the
CDC. While men who have sex with men continue to account for the largest proportion
of AIDS cases, the incidence of new cases has slowed among this population (CDC,
1997)

Thus, like the situation worldwide, the scope of the AIDS epidemic in the US is
enormous. To help put the loss of life in perspective, consider that in the fifleen years
since the epidemic began nearly 380,000 Americans have died—about six times more

than the number of Americans who died during the Vietnam War.

 

 

also a

lhusv '

health
edit:
profess
fit to di
battle c
With rh
the firs
"The In
amode
n0tes r1
Lmder c
AIDS, 1
“War
l 1934, l

l

 

It} I

Iii-L;
‘01;
I \\
CIT.

Sp;

e_\- }
i\' a
1%

5' I}

 

“A MENTAL HEALTH CATASTROPHE”

“Though widely perceived as an important medical issue, the AIDS epidemic is
also a mental health catastrophe perhaps unmatched in 20th century American history.”
Thus writes Walt Odets in his introduction to The second decade of AIDS: A mental
health practice handbook (1995, p. 1). Indeed, the AIDS pandemic nOt only confronts
medical science with one of its greatest conundrums, but also presents mental health
professionals with a new set of challenges. The American Psychological Association saw
fit to devote two special issues of American Psychologist to AIDS in 1984 and 1988. The
battle cry went out for psychologists to be involved in both AIDS prevention and assisting
with the emotional adjustment of those individuals diagnosed with HIV-related illness. In
the first special issue, Batchelor (1984) highlights the stigmatizing nature of the disease.
“The mention of AIDS causes people to draw back in fear; it has the emotional impact of
a modem-day black plague. AIDS has become a psychological emergency” (p. 1279). He
notes that AIDS presents psychological stressors that may not be adequately addressed
under current mental health treatment conditions. “In our interviews with people with
AIDS, they have repeatedly stressed-that, although their medical needs were being
adequately addressed, their psychological needs were not being given sufficient priority” ~
(1984, p. 1288).

By the time of the second special issue of American Psychologist in 1988, more
was known about the spread of AIDS and about its psychological impact. Three general
themes had emerged regarding the role of psychologists in the AIDS epidemic: Effecting
behavioral change to prevent transmission, changing social attitudes and beliefs regarding

AIDS and persons with AIDS (PWAs), and helping PWAs deal with the psychological

 

 

 

  
  

impar t 0:"
ese thre-
If

epide nic
order to u.
epidemics
disea es c3
u'hicl has
reacti ms :.
Pm c I the

Mr

 

 

impact of their condition. The concern of the present study lies mainly with the last of
these three themes.

Morin, in the 1988 special issue of American Psychologist, delineates the
epidemic of reaction and stigma that accompanies the medical epidemic of AIDS. “In
order to understand what we are facing, it may be useful to think of three separate
epidemiCs..i.First, an epidemic of HIV infection....The second epidemic: involved the
diseases characterized by the case surveillance definition of AIDS....The third epidemic,
which has received far less attention, involves the social, cultural, economic, and political
reactions to the HIV and AIDS epidemics. This third epidemic of reaction....is as much a
part of the pathology of AIDS as the Virus itself” (p. 838).

Morin elaborates on the stigma of AIDS: “Stigma associated with AIDS has lead
to people being shunned socially and suffering significant psychological damage even
though they maintain employment and housing....AIDS has disproportionately affected
groups, such as gay men and IV drug users, who are already subjected to substantial
stigma” '(1988, p. 840). In this same issue, Jones writes: “The AIDS epidemic is a
profound attack on the physical, mental, and emotional welfare of millions of human
beings. Psychology is a discipline whose expertise is critical to blunting the devastating
effects of this diSease....Another critical area in which we have much to offer is in
' understanding the attitudinal and behavioral dimensions of prejudice and discrimination”
(1988, p. 899). The present research is concerned with the stigrnatization that results from

these social attitudes, and its effect on the mental health of PWAs.

 

 

LII

HIV AIDS
and or am
Pollack, 3;
substance
at, 19931”.
WidiOChe:
Catalan. &

It is

effects on t

 

Olhers “EEC

to adVanCe

 

Chapter 1

LITERATURE REVIEW, STUDY RATIONALE, AND HYPOTHESES

THE PSYCHOLOGICAL IMPACT OF AIDS

There has been considerable research on the psychological impact of coping with
HIV/AIDS. The types of mental health problems investigated have included depression
and/or anxiety (e.g., Atkinson et a1., 1988; Cazzullo et a1., 1990; Folkman, Chesney,
Pollack, & Coates, 1993; Ostrow, Joseph, et a1., 1989; Ostrow, Monjan, et a1., 1989);
substance abuse (e.g., James, Rubin, & Willis, 1991; Perry et a1., 1990; Rosenberger et
a1., 1993); “reactive” or adjustment disorder (e. g., Bungener, Kosmadakis, Jouvent, &
Widlocher, 1993; Lipsitz et a1., 1994); and suicidal behavior (e. g., Gala, Pergami,
Catalan, & Riccio, 1992; McKegney & O’Dowd, 1992; Rajs & Fugelstad, 1992).1

It is commonsensical to assume that'HIV/AIDS has devastating psychological
effects on those infected. “Fear of contagion leads to social ostracism....The damaging
psychological effects to people with AIDS when they are faced with such abject fear in
others need not be quantified to be comprehended” (Batchelor, 1984, p. 1283). However,

to advance scientific knowledge of the mental health implications of HIV/AIDS, these

 

' Organic brain syndromes, psychotic disorders, and personality disorders have also been investigated but
do not fall within the purview of the current study. This study is concerned with “functional” (i.e.,
presumably non-organic) psychological problems (Sacks et a1., 1995), so AIDS-related dementia and other
organic mental disorders will not be considered. Psychotic disorders and personality disorders will not be
considered here because of their relative rarity as well as a scarcity of available research.

4

 

 

 

 

groups c
positite)
summm
S:
they hat. e
Other (non
Psychiatric
Kinskm’ 19

will! ed m

 

effects do need to be quantified so that we may better understand the nature and scope of '
the damaging effects of social ostracism and stigmatization. Fortunately, there have been
a number of research studies that have attempted to quantify the impact of HIV /A1DS on

mental health.

, DOES HIV SEROSTATUS AFFECT MENTAL HEALTH? FINDINGS IN THE
_ AIDS AND MENTAL HEALTH LITERATURE

A number of studies reported differences in mental health outcomes between
groups differentiated by serostatus (i.e., whether subjects are HIV-negative or HIV -
positive). The characteristics of the empirical studies discussed in this section are
summarized in Appendix F.

Several investigators reviewed hospital charts of AIDS inpatients to determine if
they had a higher rate of psychiatric consultations and psychological disturbance than
other (non-AIDS) inpatients. It was found that AIDS inpatients had a higher rate of
psychiatric consultations than did any other group of hospitalized patients (Johannet &
Muskin, 1990), that they had a significantly higher rate of repeat consultations and
required more attention from psychiatric staff than did non-AIDS inpatients (O’Dowd &
McKegney, 1990), and that a small but significant population of new psychiatric
inpatients (i.e., individuals who presumably would not otherwise have been hospitalized)
appears to have been brought on by the AIDS epidemic (Sacks et a1., 1995). The,
evidence provided by these chart reviews must be interpreted cautiously, as it is not clear
whether the higher rates of psychiatric consultations are due to AIDS per se or to

psychiatric problems related to risk-group status (an issue to be discussed in more detail

 

 

 

 

below).
Nex‘erti
by PW.-

AIDS ir

persons
users (I\
C
l l 990) ex
positiye 5
than HIV-
scores thg;
Sex
[\DL'S‘ wt

Ilealmem‘ l

 

mlllCCd cg;
Pakegch Cl .'

 

0‘frall diff;

report a gen

    
 

Whole (reea:

below). Further, the evidence is retrospective and focuses only on inpatients.
Nevertheless, these findings are suggestive of a growing demand for psychiatric services
by PWAs, and also imply that AIDS inpatients may be psychologically needier than non-
AIDS inpatients.

Other studies have examined inpatient, outpatient, and community samples of
persons considered at high risk for HIV /AIDS, including hemophiliacs, intravenous drug
users (IVDUs), and gay/bisexual men.

Only one study focusing on hemophiliacs was found. Dew, Ragni, and Nimorwicz
(1990) examined mental health outcomes among male hemophiliacs, and found that HIV-
positive subjects were more likely to have clinically elevated anxiety and anger-hostility
than HIV-negative controls. HIV -positive subjects also had higher mean depression
scores than HIV-negative controls, but the differences were not clinically meaningful.

Several investigations looked at outcomes for samples composed primarily of
IVDUs, with mixed results. In their Norwegian sample of drug addicts in residential
treatment, Hestad, Aukrust, Ellersten, and Klove (1994) found that HIV-positive subjects
evinced greater psychological disturbance than did HIV -negative subjects, and that these
differences were even more pronounced at follow-up nine months later. Similarly,
Pakesch et al. (1992) reported that HIV-positive drug users tended to have more elevated
depression scores than HIV-negative drug users. Finally, Lipsitz et al. (1994) found no
overall differences according to HIV serostatus in a sample of IVDU outpatients, but did
report a gender—serostatus interaction for men with regard to depression, with HIV-
positive men being more depressed than HIV-negative men. Significantly, the sample as a

whole (regardless of HIV serostatus) showed serious impairment in DSM Axis V global

 

 

 

 

functioni
interest i
positive :
HIV illre
experienc

T1
infection I
and biser
Elevated l:
lESpeciallj
("Bungener
& Liang, ;
general p5
negative 3‘

Int;
psi-ChOIog;
Population
a“1994) r

 

 

 

diSmeanc:

 

Chilang Cl .;

Rosellbere '.

 

functioning (mean = 49), approaching levels found in psychiatric inpatients. Also of
interest is the finding that, although there were no clear-cut differences between HIV-
positive and HIV-negative subjects, HIV-positive subjects in a more advanced stage of
HIV illness were more depressed (even when controlling for vegetative symptoms) and
experienced greater stress than subjects who Were less ill.

The largest number of studies supporting a positive relationship between HIV
infection and psychological distress have examined samples consisting primarily of gay
_ and bisexual men. These investigations reported that HIV -positive subjects exhibited
elevated levels of sexual dysfunction (Catalan et a1., 1992), any Axis I disorder
(especially adjustment disorder) (Chuang, Jason, Pajurkova, & Gill, 1992), depression
(Bungener et a1., 1993; Ostrow, Joseph, et a1., 1989), somatic complaints (Krikorian, Kay,
& Liang, .1995; Rosenberger et a1., 1993), anxiety, obsessive-compulsive behavior, and
general psychological distress (Ostrow, Joseph, et a1., 1989) when compared to HIV-
negative subjects.

Interestingly, the most consistent finding among these studies is that
psychological distress among gay and bisexual men appears to be elevated relative to the
population in general, regardless of HIV serostatus—a finding similar to that of Lipsitz et
a1. (1994) regarding IVDUs. While differences in psychological distress between HIV-
positive and HIV-negative subjects do exist, they tend to be slight; but the overall level of
disturbance in these samples relative to population norms is high (Bungener et a1., 1993;
Chuang et a1., 1992; Krikorian et a1., 1995; Rosenberger et a1., 1993). Further,
Rosenberger and her colleagues found that, among depressed HIV-positive gay men,

onset of the depression typically preceded that of HIV symptoms (or confirmation of

 

 

 

HIV-r
premo

to elex

distress
The hos
populat
hospital
Hestad e
Sample 0
positive ;
H
exception
Studies (1;
Scores am

Illness (as

 

 

HIV-positive serostatus) by a considerable margin. Taken together, these results suggest a
premorbid vulnerability to psychological distress among gay men in general, in addition
to elevations in distress related to HIV -positive serostatus.

. There is therefore some empirical support for the hypothesis that psychological
distress will be greater for HIV -positive individuals than for HIV -negative individuals.
The hospital chart reviews suggest that there is a small but not insignificant new
population of HIV-related psychiatric inpatients, and that AIDS inpatients in general
hospitals may constitute a psychologically needier group compared to other inpatients.
Hestad et al.’s Norwegian sample of drug addicts, as well as Bungener et al.’s French
sample of gay men, indicate some fairly clear-cut differences in distress between HIV-
positive and HIV-negative subjects.

However, findings of clear-cut differences related to serostatus are more the
exception than the rule. Most investigations reported equivocal results; for example, two
studies (Krikorian et a1., 1995; Rosenberger et a1., 1993) reported elevated somatization
scores among HIV -positive subjects, especially those at a more advanced stage of the
illness (as would be expected), but failed to show any differences on other measures of
psychological distress such as anxiety or depression. Likewise, Lipsitz et a1. (1994) found
no clear-cut differences related to serostatus in their sample of IV DU outpatients. Quite
commonly, then, results are mixed, with some meaningful differences between groups
emerging while others prove small and clinically insignificant (e. g., Catalan et a1., 1992;
Dew et a1., 1990). Perhaps most interesting, however, is the consistent finding that group

differences in psychological distress related to serostatus are ofien overshadowed by the

 

 

 

dmawd

bDL’sl I

associate.
borne out
find any I
negative 5
Tmm&
loseph,ét
Rabkin, R
.finther er;
Simptom.
dOCunienI:
SIHUIarl);
pSICllOIOg.
1993; \l'il;
(1993). At.-
pOSltil‘e ga;
bmpIOmS‘

\

I115 “litres:
$61,0ng gm

19939.

 

m. ' ...
FISHER

1 “agent:

 

elevated levels of distress found in so-called “high risk” samples (such as gay men and
IVDUs) regardless of serostatusz.

The intuitive, commonsense expectation that a diagnosis of HIV /AIDS would be
associated with clear-cut evidence of increased psychological morbidity is therefore not
borne out by the literature. In fact, there is a large group of investigations that failed to
find any iclear-cut, clinically significant differences between HIV -positive and HIV-
negative subjects (Atkinson et a1., 1988; F olkrnan et a1., 1993; Gala et a1., 1993; Hays,
Turner, & Coates, 1992; O’Dowd, Natali, Orr, & McKegney, 1991; Ostrow, Monjan,
Joseph, & VanRaden, 1989; Perry, J acobsberg, Card, & Ashman, 1993; Williams,
Rabkin, Remien, Gonnan, & Ehrhardt, 1991). However, these investigations do provide
0 further evidence of the importance of risk group membership and level of HIV-related
symptomatology. Several (Gala et a1., 1993; O’Dowd et a1., 1991; Perry et a1., 1993)
documented the high rate of psychiatric morbidity among IVDUs regardless of serostatus.
Similarly, gay and bisexual samples were found to have significantly elevated levels of
psycholdgical distress relative to population norms (Atkinson et a1., 1988; Folkman et a1.,
1993; Williams et a1., 1991). Further, as was the case for Rosenberger and her colleagues
(1993), Atkinson et a1. (1988) found that the psychological problems reported by HIV-
positive gay men in their sample tended to predate HIV diagnosis. Finally, HIV—related

‘ symptoms, rather than HIV serostatus per se, emerged as a significant predictor of mental

 

2 It is interesting to note that three of the studies showing the clearest differences in distress between
serostatus groups were examining European samples (Pakesch et al., 1992—Austn'a; Bungener et al.,
1993—France; and Hestad et a1., l994—Norway). It is uncertain to what degree the experience of PWAs
in these three cultures mirrors that of PWAs in the US. Regardless, two of the three (Pakesch et a1., 1992,
and Bungener et a1., 1993) are similar to most of the American studies in showing as many (if not more)

 

 

 

health 5

Monjan

risk for.
psycholc
serostatu
to serost;

explain tl

 

WM“ 51::

Hf"

 

 

10

health status in several investigations (Folkman et a1., 1993; Hays et a1., 1992; Ostrow,
Monjan, et a1., 1989).

Taken together, these findings suggest that members of certain groups at increased
risk for HIV infection (e.g., gay men and IVDUs) experience elevated levels of
psychological distress, but that there are no clear differences in distress related to HIV
serostatus per se. Most of the studies that failed to find any significant differences related
to serostatus utilized large samples3, indicating that a lack of statistical power does not
explain their. negative results.

In summary, three interesting findings emerged from the literature. First, members
of high-riSk groups such as gay men and IVDUs (regardless of HIV serostatus) evinced
high levels of psychological disturbance relative to population norms, while
comparatively few (and weak) differences related to serostatus were found. Second,
among HIV-positive gay men, it was shown that the onset of psychological problems
often predated the HIV diagnosis (Atkinson et a1., 1988; Rosenberger et a1., 1993),
suggesting a premorbid vulnerability to psychological distress. And third, while HIV-
positive serostatus per se was ofien not associated with greater psychological distress,
HIV-related symptomatology was associated with greater distress (F olkrnan et a1., 1993;
Hays et a1., 1992; Ostrow, Monjan, et a1., 1989).

HIV /AIDS and mental health therefore bear a complex relationship to one

another. As Rosenberger and her colleagues note, there are inconsistent results across

 

important similarities in psychological morbidity within the high risk samples (regardless of serostatus) as
they did differences related to serostatus.

 

 

 

studies
sympto

1993-,p

BEE

MINT!

psycholo
straightfc
an unders
Risk Gro

F i
members}
that target
501116 C35:-
Ksk gm;
m." be 1h.
that exist ;
lillettionj

general) is

 

(KESSier, It

I.l

beian‘ an.

Tm ”Edie:

 

11

studies regarding the effects of HIV /AIDS on mental health. “Whether psychiatric
symptomatology is higher in HIV-infected populations is unclear.” (Rosenberger et a1.,

1993,p.150)

FACTORS MEDIATING THE RELATIONSHIP BETWEEN HIV/AIDS AND

MENTAL HEALTH: 1. RISK GROUP MEMBERSHIP & HIV SYMPTOMS

The answer to the question, “Does having HIV/AIDS result in greater
psychological distress than not having HIV /AIDS?” does not appear to be a
straightforward “yes” or “no.” Rather, there appear to be other factors that are critical to
an understanding of this relationship.

Risk Group Membership

First, the AIDS—mental health relationship clearly seems to depend on risk group
membership (Folkmanet a1., 1993; Rosenberger et a1., 1993). The literature demonstrates
that variability in psychological distress related to serostatus is greatly reduced, and in
some cases disappears entirely, when comparisons are made within particular risk groups.
Risk group status seems to be the most salient predictor of mental health among PWAs. It
may be that being a member of certain groups involves living with many daily stressors
that exist irrespective of HIV serostatus (though they may be exacerbated by HIV
infection). For instance, it is well-documented that IV drug use (and substance abuse in
general) is associated with considerable comorbidity with other psychiatric disorders
(Kessler, 1991). Likewise, the shame, stigrnatization, and discrimination faced by gay,

lesbian, and bisexual individuals is pervasive in our society and is associated with

 

3 The median sample size was 304 in the eight studies that failed to find any significant differences between

 

 

 

 

 

 

 

 

pS}'ChOlC
Neisen. 1
members
members

distress r

.—
b-O'.

predictor
unimpon.
positive c'
also sugg;
high base
ll'lCl’t’meI-f~
discern th:
COuI’SC Of:
health at C
aCross diff

IS IIOI [he I”

 

In Predicm
l
1993; Hat-i

HIV‘REIaI

Tin;

dists
styrent ‘a

\

Stigma“ 8T6

12

psychological distress within these groups (Kaufman & Raphael, 1996; Morin, 1988;
Neisen, 1990). A thorough discussion of the mental health implications of risk group
membership is beyond the scope of the present investigation; but in any event,
membership in these groups has been associated with increased levels of psychological
distress relative to the general population.

If it is the case that the risk or transmission group a person belongs to is a better
predictor of psychological distress than HIV serostatus, does this imply that serostatus is
unimportant? Not necessarily. It does suggest the need to assess the impact of an HIV-
positive diagnosis within the context of the person’s daily life and sociocultural milieu. It
also suggests that when measuring mental health outcomes within a population with a
high base rate of psychological problems, it may be more difficult to discern the
incremental effects of additional stressors. Finally, it implies that it may be important to
discern the meaning of being diagnosed with HIV, how that meaning changes over the
course of the disease, and how the changing nature of HIV-related stressors affects mental
health at different stages of the illness. One potent set of HIV -related stressors that vary
across different stages of the illness are physical symptoms. The literature suggests that it
is not the HIV diagnosis per se, but rather the onset of symptoms, that is more important
in predicting psychological distress in PWAs (Bungener et a1., 1993; F olkman et a1.,
1993; Hays et a1., 1992; Lipsitz et a1., 1994; Ostrow, Monjan, et a1., 1989).

HIV-Related Symptoms

It must be emphasized that degree of HIV/AIDS-related symptomatology is a very

different variable than HIV serostatus. Many HIV-positive individuals are asymptomatic,

 

serostatus groups.

 

 

 

and ma}
adjustm
svmpton
to receix
transient

Hirsch. I

from HIT

appearant
may seem
fi’Om the i
emerge, d
inCllejng
Confrontir:

F0

 

Indirect]y .
mediated I?

 

 

13

and may remain so for months or even years (Volberding & Cohen, 1990). Psychological
adjustment to the HIV diagnosis therefore does not necessarily involve coping with actual
symptoms of HIV-related illness. It has been suggested that psychological distress specific
to receiving a diagnosis of HIV seropositivity while still asymptomatic consists mainly of
transient adjustment reactions (Bungener et-al., 1993; Krikorian et a1., 1995; Tross &
Hirsch, 1988). The onset of actual symptoms, however, and the progression of the disease
from HIV -positive but asymptomatic to AIDS-related complex (ARC) to AIDS,
introduces stressors that may place greater demands on psychological resources.

Both Hays et a1. (1992) and Folkman et a1. (1993) addressed the importance of the
appearance of HIV symptoms. Hays and his colleagues speculated that the threat of AIDS
may seem remote when the subject is still asymptomatic—in essence, one is protected
from the implications of the diagnosis by defensive denial. When symptoms finally
emerge, deleterious effect on mental health may result from a number of factors,
including feeling out of control, negative changes in self-image, worry over illness, and
confronting the possibility of death.

Folkman et a1. (1993) concluded that HIV symptoms affect mental health
indirectly through their effects on stress and coping processes. These processes are
mediated by individuals’ appraisals of control over their situation and by the availability

and perceived adequacy of social support.

 

 

FACTO ‘

SIESTA

I:
have bee
These in.I
control I}
includin 5
Stress Ai

L.

assess the

 

 

and utiliz
Stress 3m
PeTSOn as
signifies
51111311011
in the de
“TESS tal
the r elat

accordir

14

FACTORS MEDIATING THE RELATIONSHIP BETWEEN HIV/AIDS AND
MENTAL HEALTH: II. STRESS1 CONTROL, AND COPING

In addition to risk-group membership and HIV-related symptoms, other factors
have been implicated in the literature as affecting psychological morbidity among PWAs.
These include stressors PWAs may be experiencing, their appraisals 'of the level of
control they have over these stressors, and specific coping strategies they may employ,
including the utilization of social support.

Stress And Cognitive Appraisal

Lazarus and F olkman (1984) provide an integrative framework within which to
assess the effects of stress, appraisals of control, and coping (the latter including seeking
and utilizing social support) on psychological well-being. In their cognitive theory of
stress and coping, stressors are defined as situations or events that are appraised by the
person as significant to his well-being and as taxing or exceeding his resources. It is
significant that, in this definition, one’s appraisal of the situation determines whether the
situation is or is not a stressor. This approach recognizes that individual differences exists
in the degree to which events and situations are evaluated as stressful. This definition of
stress takes into account characteristics of both the person and the situation, emphasizing

the relationship between the two. Cognitive appraisal is crucial to the stress experience

according to this framework."

 

’ Lazarus and Folkman emphasize conscious, reality-oriented cognitive appraisals rather than unconscious
processes. “It is our premise that although personality factors such as needs, commitments, and preferred
styles of attention influence perception, appraisals are generally correlated with reality” (1984, p. 53). Yet
Lazarus and Folkman also acknowledge the possible influence of unconscious processes. Appraisal is not
necessarily an entirely conscious, rational, or intentional process, according to the authors. This idea will

 

. . .. on. I . -
.V ' l.“- .t A. " ' ._ ' I. ‘ ' ' ' —
1.. . ‘3 I. ._3' .' ‘Uo .5; ’ ' , ‘ ‘

    
  

' insane. '* '

a“.

. r O'CII. Iqu.”m;h.'im, - - . ' '. . .. .: ; ‘l
I _ I

 

Manspss

 

include the grade
problems; the fa;
caregiver burden.
Edition of sex:
alteration from f
ofsharne and gt
(Cazzullo et al..
Some of
simple. gay rr.
'I‘hether the} ar;
Stirred multip'
53333013 $6611] 3
”1310315. trea
itessors will v
offiSESSing lt‘
“hat a

P

“AS? ACCOI’

on mental bea’

C0 in
plug.

15

Many psychosocial stressors are cited in the literature as “AIDS-related.” These
include the gradual, insidious, and unpredictable cottrse of the disease; HIV-related health
problems; the fact of being stricken during the prime of life; multiple losses due to AIDS;
caregiver burden; financial difficulties; the fear of infecting others, with consequent
inhibition of sexual relationships in order to avoid transmission; social isolation and
alienation from family; societal disapproval and stigrnatization, with consequent feelings
of shame and guilt; and stressors surrounding HIV testing and treatment decisions
(Cazzullo et a1., 1990; Folkman et a1., 1993; Rabkin & Rabkin, 1995).

Some of the above-mentioned stressors may be endemic to risk; group status—for
example, gay men and IVDUs may experience societal disapproval and discrimination -

. whether they are HIV -infected or not, and members of these communities have often
suffered multiple losses or experienced caregiver burden as a result of the disease. Other
stressors seem reasonably specific to living with HIV /AIDS, particularly those involving
symptoms, treatment, and fear of infecting others. The nature and severity of these
stressors will vary by illness stage (Hays et a1., 1992), again highlighting the importance
of assessing level of symptomatology as a key HIV—related stressor.

What are the mechanisms by which stressors effect the mental well-being of
PWAs? According to Lazarus and Folkman (1984), Stressful situations exert their effects
' on mental health via two mediating processes: situational appraisals of control, and

coping.

 

emerge as important in the application of Lazarus and F olkrnan’s theoretical framework in the present
investigation.

 

Appraisals 01 C

Accordin

 

to one's evaluat: ..
notified. Cont: ;
HIT-related stre:
sness'ful situatie:
1993; Cazzullo L
1995). In additic
associated with t
depression (Cat.
Coping
Iazams
behavioral dip:
is taxing or ere
hereiore made
tonnollability c
This definition
°lPOsed to a m
heme“ routin
9nd filo”, and
also caution a5

Situation ine-v
'l

 

16

Appraisals Of Control

According to Lazarus and F olkman (1984), situational appraisal of control refers
to one’s evaluation of the degree to which a stressful situation can be controlled or
modified. Control is often cited in the literature as an important mediator‘of the effect of
HIV-related stressors on psychologicaldistress; specifically, feeling out of .control of
stressful situations is associated with increased psychological distress (Catalan et a1.,
1992; Cazzullo et a1., 1990; Folkman et a1., 1993; Hays et a1., 1992; Krikorian et a1.,
1995). In addition, some investigators have found that negative appraisals of control are
associated with maladaptive coping strategies, which are in turn associated with increased
depression (Catalan et a1., 1992; Folkman et a1., 1993; Krikorian et a1., 1995).

Coping i

Lazarus and Folkman define coping as “constantly changing cognitive and
behavioral efforts to manage specific external and/or internal demands that are appraised
as taxing or exceeding the resources of the person” (1984, p. 141). A distinction is
therefore made by the authors between appraisals of control (i.e., beliefs about the
controllability of a given situation) and actrial eflorts to control the situation (i.e., coping).
This definition also emphasizes a process-oriented conceptualization of coping, as
opposed to a more static trait-oriented model. The definition also makes a distinction
between routine, automatic adaptive behaviors and coping; the latter implies mobilization
and efi’ort, and is reserved for times of unusual psychological stress. Lazarus and Folkman
also caution against confusing coping with outcome—coping is an eflort toqmanage a

situation, irrespective of that effort’s effectiveness. A distinction is therefore made

 

Ween manJgi
mastery. but rat?
Lazarus

to In older trait
the latter often 1
stable traits lsnc
argon ‘hat past I
although the an:
charge in copir.
proach that e:
According to 1?.
overall “copin
Coping
manage or mod
bonded to mi;
3031 bpes of c

tailitate or irrr

 

 

17

between managing a situation, and mastery over the situation. Coping does not imply
mastery, but rather refers to an attempt to manage a given stressful situation.

Lazarus and Folkman assert that this process-oriented model of coping is superior
to the older trait-oriented models that grew out of psychoanalytic ego psychology because
the latter often failed to predict actual coping behavior. They do not deny the existence of

stable traits (such as characteristic defenses) that affect coping processes, but they do i

' argue that past research focusing primarily on stable traits did not prove fruitful. Thus,
although the authors state that “we should recognize that there is both stability and
change in coping” (1984, p. 130, emphasis added), they advance a process-oriented
approach that emphasizes the changing nature of coping over time and across situations.
According to the authors, therefore, it does not make sense to speak of a person having an
overall “coping style.”

Coping is broken down by Lazarus and Folkman into those strategies used to
manage or modify the stressful situation itself (problem-focused coping), and those
intended to modulate the emotional response to the situation (emotion-focused coping).
Both types of coping may be invoked during a stressful encounter and may interact to
facilitate or impede one another. It has been shown that in situations where people feel
they can control the outcome of events, problem-focused coping strategies are more likely
to be employed, whereas in situations where people feel matters are out of their control,
emotion-focused coping strategies (e. g., escape-avoidance, distancing, keeping feelings to
oneself, seeking emotional support, putting a positive spin on the situation) are more
common (Folkman et a1., 1993). Emotion-focused coping that involves avoidance is

associated with greater psychological distress (N amir, Wolcott, Fawzy, & Alumbaugh,

 

 

hm therapy it;

1937), wherefi
positive spin or
associated will
was). Ptoblsn
where there is :
thidtt'in. l9Sl ;
Several
1995) have linl
strategies. with
problem-focuse
were in tum ass
Wolf et al.. 199
pathological c
Miller and chc
scores among p
{hit found that
Obsessive and ,
[hf Other handi’

Seeking, reliin

 

 

 

18

1987), whereas emotion-focused coping that involves selective attention (e. g., putting a
positive spin on the situation or using the situation to promote personal growth) is
associated with less psychological distress (Billings & Moos, 1981; Folkman & Lazarus,
1985). Problem-focused coping is associated with less psychological distress in situations
where there is some possibility of controlling or affecting the outcome of the situation
(Aldwin, 1981; Forsythe & Compas, 1987).

Several investigators (Catalan et a1., 1992; Folkman et a1., 1993; Krikorian et a1.,
1995) have linked PWAs’ appraisals of control to their coping and problem-solving
strategies, with positive appraisals of control over situations being associated with active,
problem-focused coping (e. g., planful problem solving, active conflict resolution), which
were in turn associated with better mental health outcomes. Others (Namir et a1., 1987;
' Wolf et a1., 1991) did not explicitly measure appraisals of control, but did find favorable
psychological outcomes to be related to an active-behavioral coping style. Similarly,
Miller and Riccio (1990), in their review of the literature, found that higher depression
scores among PWAs were associated with avoidance reactions to HIV illness. Further,
they found that an active-cognitive coping style was problematic because it lead to
obsessive and ruminative thoughts and attitudes, while an active-behavioral strategy, on
the other hand, was productive. The latter included expressing feelings, infonnation-
seeking, relying on others and seeking social support, taking care of oneself, involvement
with AIDS programs, political involvement, membership in AIDS support groups, and
active pursuit of self-growth. Likewise, Tross and Hirsch (1988) recommended that the

best therapy for PWAs is to teach them coping skills in a supportive environment,

 

 

especially throe
may be among
Perceived Soci

The net‘
constitutes a d}
can serve as im
source of stress
network is syn
They argue tha‘
important. Soc
@Pmpriatel}: 1
atom of COpir

Social 1
ofPll'As (Chu

1993; Hays et

19% T1055 &

0115 oflhe 5C3:

scale loaded c-

problem-soles

 

dsSOciated Wi.‘

 

 

19

especially through participation in peer support groups. Indeed, seeking social support
may be among the most important types of active-behavioral coping.

Perceived Social Support

The network of social relationships within which an individual functions
constitutes a dynamic, ever-changing contextifor stress and coping. Social relationships
can serve as important resources for the person, but may also present demands and be a
source of stress as well. This is why it should not be assumed that a person’s social
. network is synonymous with social support, according to Lazarus and Folkman (1984).
They argue that it is the level and quality of social support perceived by the person that is
important. Social support is an important resource that must be cultivated and used
appropriately; utilization of social support is therefore viewed by Lazarus and Folkman as
a form of ~c0ping.

Social support has been widely cited as beneficial to the psychological adjustment
of PWAs (Chung & Magraw, 1992; Conwell, 1994; Dew et a1., 1990; Folkman et a1.,
1993; Hays et a1., 1992; Lackner et a1., 1993; Michels & Marzuk, 1993; Miller & Riccio,
1990; Tross & Hirsch, 1988; Wolf et al., 1992). “Seeking social support” is included as
one of the scales in the Ways of Coping questionnaire (Folkman & Lazarus, 1988). A
second-order factor analysis of the instrument showed that the “seeking social support”
scale loaded on an involvement coping factor (along with items measuring planfiil
problem-solving and positive reappraisal). Involvement coping was in turn positively

associated with less depressed mood in gay men (F olkman et a1., 1993).

 

Cognitive Tl

Acco:
health by a Cl
commitments
determine at.
appraisal lead
social suppon
lhese etlects
cognitive appi

All Applicatit
among Gay .\

This st
mental health '
and Folkmztn'
steady.

Panici

SUblietts were

 

dining the pas

 

filelids, relati.

 

 

hereon“
Stressful don:

QUeSil-OI'mair;

20

Cognitive Theory Of Emotions

According to Lazarus and Folkman (1984), all of the foregoing is linked to mental
health by a cognitive theory of emotions which posits that the values, beliefs,
commitments, and goals that influence a person’s cognitive appraisal of a situation
determine whether that situation is experienced as stressful and controllable; this
appraisal leads to the mobilization of certain coping strategies (including the utilization of
social support) which in turn help to shape the emotions that the person experiences.
These effects operate reciprocally as well—emotional responses can shape subsequent
cognitive appraisals.

An Application: Folkman and Colleagues’ Study of Stress, Control, and Coping
among Gay Men in San Francisco

This study (Folkman et a1., 1993) of the effects of stress, control, and coping on
mental health merits a more detailed examination because it draws directly upon Lazarus
and Folkman’s (1984) theory of stress and coping and serves as a model for the current
study. i

Participants were 425 gay and bisexual men recruited in the San Francisco area.
Subjects were asked to rate on a 4-point scale the degree of stress they had experienced
during the past month in each of nine domains (primary relationships, relationships with
friends, relationships with family, work, finances, illnesses of close others, deaths of close
others, own health, and political issues). They then were asked which of these nine
domains had been the most stressful for them, and to indicate on a 4-point scale the
degree of control they feel they have over that domain. Coping relative to this most
stressful domain was assessed using a short (19-item) version of The Ways of Coping

Questionnaire (WOC; Folkman & Lazarus, 1988). The WOC has participants indicate on

alpoint scale
attempting to c
seren scale so
avoidance. dis
reappraisal. D
Bnei Symptor
measured trait
in l989.

HIV st
Pl’litlpanls u
recent test. th
They were the
ellpounced c
0“ lhfise reSp
Positive 35m
“moms“

The a
“ten W0C
emerged: dB:
and dlSIanci;
Social SUPPO:
sex‘ drugs, 0

bEhAViorai fit

21

a 4-point scale the extent to which they used each item (i.e., coping strategy) in
attempting to deal with the most stressful domain of their lives. The short WOC yields
seven scale scores: self-controlling, cognitive escape-avoidance, behavioral escape-
avoidance, distancing, planful problem-solving, seeking social support, and positive
reappraisal. Depressive mood was measured using a 6-item depression subscale from the
Brief Syrhptom Inventory (Derogatis & Melisaratos, 1983). DepreSsive mood was
measured twice, once in 1988 and again in 1989. All other variables were measured once,
in 1989.

HIV serostatus and symptomatology were assessed by subject self-report.
Participants were asked whether they had been tested for HIV, the date of their most
.recent test, their current serostatus, and whether they had been diagnosed with AIDS.

' They were then asked to indicate which of 15 physical symptoms, if any, they had
experienced during the past year (please see Appendix D for symptom checklist). Based
on these responses, “HIV status” was categorized as status unknown, HIV-negative, HIV-
positive asymptomatic (zero to one symptom), HIV-positive symptomatic (two or more
symptoms), or AIDS (diagnosed).

The authors performed asecond order factor analysis (oblique rotation) of the
seven WOC scale scores to reduce the number of coping variables. Three factors
emerged: detachment coping (consisting of self-controlling, cognitive escape-avoidance,
and distancing), involvement coping (consisting of planful problem-solving, seeking
social support, and positive reappraisal), and behavioral escape-avoidance (e.g., using
sex, drugs, or alcohol to escape problems), which loaded by itself. Item analysis of the

behavioral escape-avoidance factor indicated that those strategies were used relatively

 

infrequently; it.
was excluded f

It was f
relative to prom
standard deviat
previous findir
Rosenberger et
population. A l
depression sco

C ross-s
3333.!) =01) ;
ltporting Sig“;
alTflptomatic

diagnosed and

 

Long:

1-6i,p=.17'

HIV 5
COHiIOHlng {C

to.- -
aria-“Ce \l. _

 

 

22

infrequently; in addition, the scale had poor reliability (alpha = .29). For these reasons it
was excluded from subsequent analyses.

It was found that the sample of gay men as a whole was significantly depressed
relative to population norms both at time one (T1) and time two (T2), scoring 2.8
standard deviations above the mean forlnon-patient males. This result is in line with
previous findings (e.g., Bungener et a1., 1993; Chuang et a1., 1992; Krikorian et a1., 1995;
Rosenberger et a1., 1993 ) that gay men as a group are more depressed than the general
population. A logarithmic transformation was used to normalize these positively skewed
depression scores.

Cross-sectionally, HIV status was related to depression both at T1 (F [4, 411] =
3.28, p = .01) and T2 (F [4, 416] = 4.44, p = .001), with AIDS-diagnosed subjects
reporting significantly more depressive mood than HIV-negative, HIV-positive
asymptomatic, or status unknown. There were no significant differences between AIDS-
diagnosed and HIV-positive symptomatic subjects, or between HIV-positive symptomatic
subjects and those in the other three categories.

Longitudinal results were somewhat different; it was found that HIV status had no
significant effect on depression at T2, after controlling for depression at T1 (F [4, 415] =
l.61,p = .17).

HIV symptoms, however, were related to depressive mood at T2, even when
controlling for depressive mood at T1. Subjects were grouped according to whether they
reported zero or one symptom versus two or more symptoms, and an analysis of

covariance was performed using T1 depressive symptoms as the covariate (F [1, 422] =

5.66. p = .02?!

subsequent p;

C orre

l. The bivariz
mood at T2 \\

moodatll u

 

 

23

5.66, p = .02). Based on this finding, the authors covaried HIV symptoms in the
subsequent path analysis.

Correlations among the variables used in the path analysis are presented in Table
1. The bivariate analyses revealed, not surprisingly, that the best predictor of depressive
mood at T2 was depressive mood at T1 (r = .70, p < .001). For this reason, depressive

mood at T1 was also covaried in the subsequent path analysis.

 

 

Table L-l

Bivariate cor.
__________
’-
l Depressive

2 an 55mph

4 Control
5 Detachmcn
6 Involvemer

7 Depressive

N
p< .03; top

Folkn
S’Ll’Eszu] dOUJ

While those \

 

 

 

With depTESs
depreSSiVe pil
mediailng 6‘

“glue 1,

 

24

Table L-l

Bivariate correlaiions among variables in the path analysis (from Folkman et a1., 1993)

 

 

1 2 3 4 5 6 7
1' Depressive mood Ti 1.00 .18" .41** -.15** .31“? -.21** .70**
2 HIV symptoms 1.00 .23" -.01 .O9* -.02 .21**
3 Stress 1.00 -.20** .29M -.09* .54“
4 Control 1.00 -.03 .21“ -.18"'*
5 Detachment 1.00 —.23** .38"
6 Involvement 1.00 -.31**
7 Depressive mood T2 1.00

 

*p < .05; **p < .001.

Folkman and her colleagues hypothesized that subjects who appraised the most

stressful domain of their lives as. controllable would rely more on involvement coping,

while those who felt out of control of this domain would rely on detachment or avoidance

coping. It was further expected that involvement coping would be negatively associated

with depressive mood, while detachment coping would be positively associated with

depressive mood. Thus, stress would exert its effect on depressive mood via the

mediating effects of control and coping. The results of the path analysis are shown in

Figure 1.

l

l

V

‘ Con

‘p < .05; ‘l

IDEPICSSlVe

AS p l
related 10 {Ce
ml'Olvemem
Al

Emma”. [O {

 

SIT

SSS Ellen;

 

25

Figure L-l

Path analjgis from F olkman et a]. (1993L

 

 

Stress

 

 

 

.070

   

 

 

 

 

 

Detachment &.

-. 181 * * -.150** Depressive Mood T2a

 

 

 

   

 

 

 

Involvement

 

.077

 

V

Control

 

 

 

 

*p < .05; ”p < .01

a Depressive mood at T1 and HIV symptoms are covaried in all analyses.

As predicted, the path analysis indicated that increased stress was negatively
related to feeling in control; feeling in control was in turn positively related to
involvement coping; and involvement coping was negatively related to depressed mood.
Also as expected, detachment coping was positively related to depressed mood. However,
contrary to expectations, this effect was not mediated by appraisal of control; instead,
stress exerted its effect on detachment coping directly. Also contrary to expectations was
a'significant direct effect of stress on depressive mood. As Figure 1 indicates, this direct
path from stress to depressed mood is the strongest effect in the model (beta =' .27, p <

.01).

 

Overa
(adjusted R‘ =
for 50% ofth
statistically si

These
that stress aff:
focused copira
individuals a;
is'i‘v'OiVEd. pro]

QEPTESSCd met

 

 

 

26

Overall, the model accounted for 60% of the variance in T2 depressed mood
(adjusted R2 = .60). The covariates, depressed mood at T1 and HIV symptoms, accounted
for 50% of the total variance; stress, control, and coping accounted for the other 10%, a
statistically significant increase (F = 21.89, p < .001).

, These results provide partial support for Folkman et al.’s model. The data suggest
that stress-affects depression in part through the negative effects of passive, avoidance-
focused coping strategies that are positively associated with depressed mood. However, if
individuals appraise the stressful situation as controllable, they are more likely to bring
involved, problem-focused coping strategies to bear, which are negatively associated with
depressed mood.

The data fall short of F olkman et al.’s expectations in two important ways,
however. First, there is the unexpected asymmetry in the relationship among stress,
appraisals of control, and coping. That is, appraisals of control mediated the effects of
stress on involvement coping, but not on detachment coping. Instead, stress exerted its
effect on detachment coping directly. F olkman et al. suggest that this problem may stem
from assessing control and coping in one domain only, thus lowering the reliability of
these measures. The reliability of the coping measure is further compromised by the fact
that it is a short version of the original WOC, and omits two of the scales found in the
firll-length version of the instrument. The current investigation will attempt to address
some of these methodological concerns.

Second, the significant direct pathway between stress and depressed mood

suggests that there may be other important variables mediating between stress and

 

 

depression th.
factor—sham

Anoth
confounding l
by Folkman e
has received a
varLable used
ot‘collapsing ‘
l'vith the addi
collapsing HP
information is
and one becon
positive serosr
{claim Simple
(fishnet adaptj
PSIv'ChologiCaI
For this ”3350‘
Sarcasm“ Wii:

|'

negalll'e. NU?

Emilie.

rial

Se"Wants an

 

 

27

depression that are not accounted for by Folkman et al.’s model. One candidate for such a
factor—shame—will be introduced and discussed in more detail below.

Another potential problems with Folkman and colleagues’ study is the
confounding of HIV serostatus with HIV symptoms. The variable “HIV status” as defined
by Folkman et a1. combines HIV serostatus, number of HIV symptoms, and whether one
has received'an AIDS diagnosis. This confounds “HIV status” with the “HIV symptoms”
variable used in the path analysis. In effect, these two variables are simply different ways
of collapsing the continuous variable “number of symptoms” into a categorical variable
(with the addition of HIV serostatus in the case of the “HIV status” variable). By
collapsing HIV symptoms into a categorical variable and adding in HIV serostatus,
information is lost (by decreasing the measurement level of the HIV symptoms variable)
and one becomes unable to differentiate the psychological effects of knowledge of HIV-
positive serostatus (irrespective of symptoms) from those of the experience of HIV-
related symptoms. It is an assumption of the current study that these constitute two
distinct adaptive contexts within which the individual may experience different kinds of
psychological distress presenting different demands on coping and adjustment resources.
For this reason these two variables will be kept conceptually distinct in this study. HIV
serostatus will be' measured as a binary categorical variable: HIV-positive or HIV-
-negative. Number of HIV symptoms, on the other hand, will be treated as a continuous
variable.

The present study will also depart from Folkman et al. (1993) in that HIV

serostatus and HIV symptoms will be included in the full path model. It is believed that

 

 

these are theort
rather than exc
These c

example of res

 

perspective of
nature of the s
investigation ).
permits the ex
investigation i.

some of the n:

 

described belt:

I
Lazar.

investigate st:
(1993)app1y p
and Coping 1T.
(rather than S‘
efforts in that
instead With 1
hem» in Panic

thuman res-

 

 

28

these are theoretically important variables that should be included as terms in the model,
rather than excluded or covaried.

These concerns notwithstanding, Folkman et al. (1993) provide a valuable
example of research on psychological functioning in PWAs from the stress and coping
perspective of Lazarus and Folkman (1984). Especially important are the longitudinal
nature of the study (which unfortunately cannot be duplicated in'the present
investigation), and the use of the powerful statistical technique of path analysis, which
permits the examination of structural relationships among variables. The current
investigation will use F olkman et al.’s (1993) work as a model (attempting to address
some of the methodological concerns noted above), and will seek to extend that model as

described below.

EXTENDING THE STRESS AND COPING MODEL

Lazarus and F olkman (1984) provide an elegant theoretical model with which to
investigate stress and coping processes and their emotional outcomes, and Folkman et a1.
(1993) apply this theoretical framework to empirical research involving PWAs. The stress
and coping model is an advancement over other approaches in that it is process-oriented
(rather than static) and allows for reciprocal causality. It is also an improvement over past
efforts in that it does not look solely at static person factors (i.e., traits), but is concerned
instead with the dynamic interplay of both person and environmental factors that come to
bear in particular stressful situations. The model therefore does justice to the complexity

of human responses to stressful situations.

In the
existence oft
Appraisal is 1
authors. “[A]
appraisal" ( l‘.
theoretical mt

“Our 1

psyche

persor

Specif
Their claim 0

heed 0n this

 

0069‘

y

Folkman are

Ollellled am?
|
St .

an.
mm may a

 

 

29

In their model of stress and c0ping, Lazarus and Folkman acknowledge the
existence of unconscious and non-rational influences on the appraisal and coping process.
Appraisal is not necessarily a conscious, rational, or intentional process, according to the
authors. “[A]n individual may be unaware of any or all of the basic elements of an
appraisal” (1984, p. 52). Lazarus and Folkman appear to welcome the integration of their
theoretical model with psychodynamically-oriented perspectives. i

“Our position allows the concept of appraisal to be integrated with depth or

psychoanalytic-type theories....Appraisal theory thus need not be restricted to

personal agendas that are accessible and easily operationalized; less accessible
agendas and processes, about which psychoanalytic theorists have been most
vocal, are also fair game. Appraisal theory is in a sense neutral with respect to the

specific personal agendas that are conceived to shape it” (1984, p. 52).

Their claim of neutrality notwithstanding, Lazarus and F olkman’s model—including the
cognitive theory of emotion linking appraisals and coping to emotional responses to
stressors—emphasizes conscious cognition and overt behavior. In practice, the research
based on this model has not adequately addressed the role that unconscious or
intrapsychic factors may play in shaping appraisals of stress and control, coping
behaviors, and mental health.

One such intrapsychic factor is internalized aflect (Kaufman, 1989). Lazarus and
F olkrnan are concerned with overt emotional responses to stressors, and their process-
oriented approach acknoWledges that these emotional responses will in turn influence
subsequent cognitive appraisals. However, internalized, unconscious, or unexpressed

affects may also play a major role in psychological symptom formation and distress

 

(Kaufman.
Lazarus ant
assessed alc
Reg;
shame assoc
Kaufman. l'
PWAs vvho
HIV AIDS c
orientation c
PT0poses to I
booed that n

relationship ,

THE ROL

    

 

 

30

(Kaufman, 1989; Lewis, 1971; Lewis, 1987). This observation is not inconsistent with
Lazarus and Folkman’s model, but these internalized affects have not been adequately
assessed alongside overt cognitive appraisals and coping behaviors.

Regarding PWAs, it is hypothesized that the most salient internalized affect is
shame associated with the stigmatizing nature of the disease (Herek,& Glunt, 1988; '
Kaufman, 1989; Levenson, 1988; Morin, Charles, & Malyon, 1984; Nichols, 1983). For
PWAs who are also members of particular risk or transmission groups, the shame of the
HIV/AIDS diagnosis is compounded by already-existing stigma concerning sexual
orientation or IV drug use (Morin, 1988; Nichols, 1983). The current project therefore
proposes to examine shame in the context of stress, appraisal, and coping in PWAs. It is
hoped that the inclusion of shame will result in a more fully specified model of the

relationship among HIV /AIDS, stress, coping, and depression.

THE ROLE OF SHAME

This study draws primarily from an affect theory perspective on shame. This
perspective was developed by Silvan Tomkins, and fiirther elaborated by Gershen
Kaufman with respect to shame. Affect theory has been selected as the theoretical
vieWpoint which can best guide the present investigation for several reasons. Affect
theory, unlike more traditional psychoanalytic approaches (e. g., Lewis, 1971; 1987)5,
differentiates the affect system from the drive system, thereby providing a more precise

language with which to describe and predict emotional experience (Kaufman, 1989).

 

5 It is acknowledged that many other theoretical treatments of shame exist as well, notably those of Lynd
(1958), Nathanson (1994), Tangney (1995), and Wurmser (1981). However, a full discussion of the many
varied perspectives on shame is beyond the scope of the current study.

 

Secon
over ll
theory
innate
povver
practic
develo-
attest;
the int

ot‘the :

distres-
COmpa
366 App

Affect

mil-31m]

31

Secondly, Tomkins and Kaufman argue persuasively for the primacy of the affect system
over the drive system. Affect theory subsumes and integrates classical psychoanalytic
theory and object relations theory within a general theory of the emotions, one in which
innate drives and relationships with both “real” and internalized objects derive their
power over behavior and personality through their. fusion with affect. Thirdly, from a
practical standpoint, affect theory, unlike other approaches, has led directly to the
development of standardized instruments for measuring internalized shame. Rather than
. attempting to infer shame from clinical material as Lewis (1971) has, instruments such as
the Intemalized Shame Scale (ISS; Cook, 1994) allow for more objective investigations
of the impact of internalized shame on psychological health.

What follows is a brief discussion of shame and its impact on psychological
distress from the perspective of affect theory. For a more expanded discussion of affect
theory in general, the place of shame within that framework in particular, and a
comparison of the affect theory and psychoanalytic approaches to understanding shame,
see Appendix C.

Affect Theory

Tomkins (1995) considers affect, both positive and negative, to be the primary
innate motivating mechanism of human behavior. This view differs from the traditional
psychoanalytic perspective, in which the drives are primary. Tomkins considers drives
and affects to constitute separate systems, both of which are important motivators of
behavior (Kaufman & Raphael, 1996). The affect system, however, is considered primary
over the drive system because the drives require amplification by affects in order to

function. For example, the sex drive must be fused with excitement affect in order to

 

quicklj
tunctic
can be

noxiou

that m,
intense
pomert
magtiif
infiuen
35 rules
10 Spec;
appears

Foikma

reailll'a
{WllVa
£0115ch

Scene re

I11

32

function properly; any impediment to excitement, such as shame, disgust, or fear, will
quickly disrupt the sex drive and in many instances make it impossible for individuals to
function sexually until the impediment is removed. Likewise, the hunger and thirst drives
can be quickly overridden by disgust, which functions to prevent the ingestion of toxic or
noxious substances (Kaufman & Raphael, 1996).

Combinations of affects (positive or negative), along with the stimuli or situations
that trigger them, are stored in memory as scenes according to Tomkins (1995). The more
intense the affect attached to the original object or event, the more memorable and
powerful the scene. Scenes involving similar affects can become interconnected and
magnified, thereby increasing their psychological “footprint” and exerting considerable
influence over one’s psychic life. Scenes in turn generate scripts, which Tomkins defines
as rules for action and cognition that allow one to predict, interpret, control, and respond
to specific scenes (Kaufman & Raphael, 1996). The affect theory concept of script
appears similar to the notion of coping strategies found in the work of Lazarus and
Folkman.

Scenes eventually disappear from full consciousness until they are later
reactivated by a similar scene, or by language that evokes the old scene. When scenes are
reactivated, a substantial portion of the original scene may remain inaccessible to
conscious experience. As long as important aspects of the scene remain inaccessible, the
scene resists control or modification, and one is prone to reexperience the scene again and

again (Kaufman, 1989).

 

The A:

inhibitc
incorn;
Thus. a
TECUCllt
Words.

Ofposn
lfiCOmp
for rein
breach

longing
Silamg v

relatior.

33

The Affect Of Shame

Tomkins defines shame as “an innate affect auxiliary response and a specific
inhibitor of continuing interest and enjoyment....The innate activator of shame is the
incomplete reduction of interest or joy....” (Tomkins, 1995, p. 84, emphasis in original).
Thus, according to Tomkins, shame is evoked whenever one experiences a partial
reduction in" the positive affects of interest—excitement or enjoyment—joy. In other
words, shame is triggered when there is a breach or interruption in the selfs experience
of positive affect vis-a-vis self or other, but only when the breach or interruption is
incomplete, leaving vestiges of the positive affect intact. This creates a “sense of longing
for reinstatement or reparation of the relationship or situation that existed prior to the

.breach (Kaufman, 1989). If the positive affect is completely reduced—if there is no
longing for a continuation or reinstatement of the previous positive affective state—then
shame will not be experienced. There needs to be an ongoing investment in the
relationship in order for shame to be evoked.

This idea is readily applicable to interpersonal relationships and affective bonds.
Kaufman (1989;1992) refers to the affective bond between two people as the
“interpersonal bridge.” Whenever this bridge is severed or damaged, whether due to a
fight, betrayal, or other failure, strong negative emotions ensue, including shame. The

~ deeply disturbing feeling of shame strongly motivates one to restore or repair the breach
to the interpersonal bridge. Thus shame functions to motivate the reparation and
maintenance of interpersonal bonds. Shame does not, however, require the presence of

another person in order to be felt. The self can and does shame the self (Kaufman, 1989).

 

 

Sham

and at
U

U]

C0051

34

Shame Vs. Guilt

Unlike other authors (e. g., Lewis, 1971), Tomkins does not differentiate shame
and guilt as different affects per se. Rather, they represent different “affect complexes”
consisting of shame as the core affect, but with differing “causes and consequences.” The
same may be said of shyness, inferiority, embarrassment, etc. All are considered by
Tomkins (1995) to be variants of the shame experience. '

Phenomenology Of Shame

“Shame....is an affect auxiliary to the affect of interest—excitement. Any
perceived barrier to positive affect with the other will evoke lowering the eyelids and loss
of tonus in the face and neck muscles, producing a head hung in shame” (Tomkins, 1995,
p. 85). This image of shame as “loss of face” is a familiar one. Shame is a deeply
disturbing experience in which one feels profoundly defective, diminished, inferior.
These feelings are accompanied by hanging the head and lowering the eyes, along with
increased autonomic activity and heightened self-awareness. Kaufinan describes shame as
feeling “seen in a painfully diminished sense” (1989, p. 17). It is the experience of being
exposed as inferior, unworthy, lacking. Shame is experienced as an interruption of the
smooth functioning of the self, and as a breach in the interpersonal bond between self and
other. Shame is so disturbing precisely because it is so central to the self and to the
formation of identity and self-concept, as well as to interpersonal relationships (Kaufman,
1989)

Shame is also disturbing because of the ambivalence it engenders toward the
perceived source of that shame. There exists a longing for a reestablishment of the

interpersonal bridge between self and other, a yearning to heal of the rift within the self

 

 

(Kaufman.
because it n
contribute ti
experienced
Shame Ant
Star
in fact plays

connection a

 

35

(Kaufman, 1989). In an evolutionary sense, therefore, shame may be seen as adaptive
because it motivates individuals to maintain and repair interpersonal relationships, which
contribute to survival. Simultaneously, however, there may be rage and hatred
experienced toward the shaming other as well as the shameful self.

Shame And Mental Health

Shame, though always disturbing, is not necessarily harmful in and of itself, and

l in fact plays the important role of motivating one to reestablish a positive emotional
connection with the source of one’s shame. However, like all affects, shame can be
“magnified in frequency, duration, and intensity” according to Tomkins (1995) and
thereby become psychologically harmful or “malignant.” The shame experience may also
be made malignant by combining it with other affects such as distress; when prolonged,
that combination of affects results in depression. This process of psychological
magnification or interconnection of shame scenes, in combination with the
internalization of shame through shame binds, results in a self that is profoundly and
malignantly bound by shame (Kaufman, 1989).

Kaufman (1989) posits that shame in combination with other negative affects will
lead to a variety of shame-based syndromes. For the purposes of the present study, the
relationship of shame to depressive syndromes is most relevant. Both Kaufman and
Tomkins view depression as a prolonged state of shame combined with distress. Unlike
traditional views that consider depression to be a direct result of inwardly-directed anger,
Kaufman suggests that while this dynamic does exist, it is secondary to the more
important shame-distress dynamic. SetV-blame and self-contempt identity scripts are also

common according to Kaufman. In the self-blame script, the self is repeatedly accused by

 

 

 

the self for p
denouncerr.
involves Ill;
reactivate a:

depression.

“AV EPID

 

liierel; 8; C,
stigma of A
as from the
as gabhise)

AID

 

 

Iransmittedl
tmsmlSsio
engagi‘d sh,
thus PTOne 1
tenor 0\‘er'
Willi reSDEC
clear in Soc

36

the self for real or imagined transgressions, with the blaming accompanied by angry
denouncement and humiliation. The self-contempt identity script, on the other hand,
involves the self rejecting the self entirely. Self-blame and self-contempt scripts
reactivate and intensify powerful shame-distress scenes and thereby exacerbate

depression.

“AN EPIDEMIC OF STIGMA”

9

AIDS, more than any epidemic in modern times, is truly “an epidemic of stigma’
(Herek & Glunt, 1988). To be stigmatized is, in effect, to be shamed by society. The
stigma of AIDS stems in part from the fact that it is a deadly, infectious disease, as well
as fi'om the association of AIDS with traditionally stigmatized subgroups in society such
as gay/bisexual men, IV drug users, people of color, and the poor.

AIDS is also shameful and stigmatized because of its identification as a sexually
transmitted disease, despite that fact that sexual contact is only one of several
transmission routes. As Tomkins reminds us: “Sexuality has from the beginning of time
engaged shame” (1995, p. 406). In our Western society, sexuality is a moral issue, and is
thus prone to evoke guilt (immorality shame); further, this guilt has been merged with
terror over the consequences of disobedience to a wrathful God. This is especially true
with respect to gay/bisexual/lesbian sexuality. This “immorality and punishment” motif is
clear in social attitudes that consider AIDS to be a punishment from God visited upon
homosexuals and other “wicked persons.” Neisen (1990) cites such attitudes as an
example of the intensification of shame due to heterosexism that is inflicted upon all

PWAs regardless of sexual orientation.

 

addn
parti:

then

dean
sens:

pO‘W :

37

Finally, the powerlessness and dependency that accompany AIDS engender
additional shame. Tomkins writes that “shame may be felt at the vicissitudes of fate,
particularly in the face of death, when the self reluctantly acknowledges the fragility of
cherished relationships” (1995, p. 404). Kaufman (1989), and Kaufman and Raphael
(1996), expand upon this point. Powerlessness, when encountered in adulthood,
reactivates primitive scenes of the primary helplessness experienced by infants. This state
of helplessness and impotence is incompatible with adult existence, which is built upon
the sense of competence and mastery gained through maturation. The adult, once again
faced with the helplessness and powerlessness of childhood, feels shamed.

Kaufman (1989) discusses the affect dynamics of powerlessness in PWAs in more
detail. An HIV/AIDS diagnosis, or just the fear or threat of contracting HIV, evokes a
sense of acute powerlessness and uncertainty. PWAs must contend not only with their
powerlessness, but also with the fear, anger, shame, distress, dissmell,6 and disgust that
powerlessness activates. According to Kaufman (1989), and Kaufman and Raphael
(1996), the most toxic combination is fear and shame, which become magnified into
humiliation and terror for PWAs. These magnified negative affects may fiirther suppress
the person’s already compromised immune system, and may also reactivate old shame
scenes surrounding the sense of self-identity (e. g., one’s identity as a gay man) that had
ostensibly been resolved long ago. The stigmatizing societal reaction to AIDS further

magnifies the terror and humiliation experienced by PWAs.

 

6 Dissmell is defined in affect theory as an affect auxiliary that serves to protect the oxygen drive; it is the
olfactory equivalent of disgust. See Appendix C for a more complete discussion.

 

 

Pot
according :
family. ho."

experience

 

additional

llhen sec;
Pll'As. the
rapidly ma
tapenence
Which in tt
Other end-o.

1h" 3315 t0 t1

 

pSE'CllO 30m
|

H989). Th

reliabilitv

38

Powerlessness may impinge upon any one of a number of security areas
according to Kaufinan (1989). Security areas include one’s intimate relationships and
family, home, health, career and work, identity, etc. For the PWA, powerlessness is
experienced most immediately in the domains of health and future prospects, with serious
additional threats to relationships, home, employment, and other sources of support.
When security areas are threatened, rapid magnification of negative affect results. For
PWAs, these affects include primarily shame, terror, rage, and distress. Much of this
rapidly magnifying negative affect may be systematically suppressed, to prevent its being
experienced as overwhelming. Such suppression, however, leads to backed—up affect,
which in turn may produce physical stress reactions such as elevated blood pressure and
other endocrine changes. This chain of events, beginning with powerlessness or other
threats to the person’s security areas and culminating in physical stress and
psychosomatic illness, is termed the powerlessnessaffect-stress cycle by Kaufman

(1989). This cycle may further degrade the already fragile health of PWAs.

AIM OF THE PRESENT STUDY

The powerlessness and uncertainty that accompany an AIDS diagnosis, combined
with the stigmatizing reactions of society, family, and fiiends, make AIDS the most
powerfully shame-infused disease since leprosy. Given this, it is surprising that no
systematic empirical study of the role of shame in mental health outcomes for PWAs has
yet been published. Not a single research article was found that systematically measured
shame in PWAs using standardized instruments with demonstrable validity and

reliability. This is not to say that none of the articles discusses the shame experienced by

 

 

 

Pll'As. Or.

 

context of

James. St E
l990). sup
observatior
(Carr. 19%
general th:

these articl

 

increasing
them; hovv
psjv'ChO lo g

Th
mOGel\\.
aPPTai Salg
OI imema

39

PWAs. On the contrary, such discussions are common, with shame issues reviewed in the
context of psychotherapeutic work with PWAs (Barret, 1989; Bradley, 1991; Lippmann,
James, & Frierson, 1993; Martin & Henry-Feeney, 1989; Schaffner, 1994; Shemoff, '
1990), support groups for PWAs (Chung & Magraw, 1992; Millan & Ivory, 1994), and
observation and psychiatric consultations with PWAs in hospice and hospital settings
(Carr, 1989; Dilley, Ochitill, Perl, & Volberding, 1985). There are also a number of
general think-pieces and review articles that raise the issue of shame and stigrnatization in
PWAs (Herek & Glunt, 1988; Levenson, 1988; Morin et al., 1984; Nichols, 1983). All of
these articles make important contributions to the literature, particularly with respect to
increasing our understanding of the phenomenology of living with HIV/AIDS. None of
them, however, explicitly measures shame, nor documents the relationship of shame to
psychological distress among PWAs.

The present study therefore extends Lazarus and F olkrnan’s stress and coping
model—which emphasizes the importance of cognitive appraisals of stress, cognitive
appraisals of control, and subsequent coping behaviors—by adding to it the measurement
of internalized shame. The effects of stress, coping, and shame on depressive
symptomatology were investigated within a sample of gay men with and without
HIV/AIDS, as Well as within a comparison sample of heterosexual male college students.
' It was hoped that this more elaborated model would increase our understanding of the
factors associated with psychological distress among gay men with (and without)
HIV/AIDS, as well as increase our understanding of the relationships between health

status, stress, coping, shame, and depression among men in general.

 

 

Tn

um'ortunat

most of th

uch less

not permit

However.

mptoms

 

40

The effects of HIV serostatus were assessed within the gay sample only, due to the

unfortunate fact that the serostatus of the college student sample is largely unknown since

most of these men have not been tested for HIV. Heterosexual male college students are

much less likely to get tested for HIV compared to gay men, and time and resources did

not permit recruitment of a sample of heterosexual men aware of their HIV status.

However, all of the men in both samples were asked about the presence of physical

symptoms related to HIV/AIDS.

99°F9‘5099NE“

VARIABLES

Sexual orientation

HIV serostatus (gay sub-sample only)

Degree of internalized shame

Number of physical symptoms

Degree of perceived stress

Degree of perceived control over stressful situations
Degree to which involvement (active) coping is used
Degree to which detachment (avoidance) coping is used
Degree of depressive symptomatology

HYPOTHESES

Bivariate Hypotheses:

-1.

The gay sample will evince greater internalized shame than the heterosexual
sample, due to the stigma associated with homosexuality (Kaufman & Raphael, 1996;
Nichols, 1983).

The gay sample is expected to manifest more depressive symptomatology than
the heterosexual sample, based on previous findings (Bungener et al., 1993; Chuang

et al., 1992; Folkman et al., 1993; Krikorian et al., 1995; Rosenberger et al., 1993).

 

 

 

3. \l'ithir
intern
Glunt.

4. hlorep
part dc.
asvvel'.
(Kaufn

5. More]
S}mptc

1993;]

 

6. Create
COping
7- LOWer
c0pin:

8' LOWe]

41

. Within the gay sample, HIV-positive serostatus will be associated with greater
internalized shame because of the stigma associated with HIV/AIDS (Herek &
Glunt, 1988; Kaufman, 1989).

. More physical symptoms will be associated with greater internalized shame, in
part due to the stigrnatizing and disfiguring nature of manyiAIDS-related symptoms,
as well as the increasing sense of powerlessness one experiences with declining health
(Kaufman, 1989).

. More physical symptoms will be associated with greater overall stress level, since
symptoms are perhaps the most significant AIDS-related stressor (Folkman et al.,
1993; Hays et al., 1992; Ostrow, Monjan, et al., 1989).

. Greater perceived control will be associated with increased use of involvement
coping (Folkman et al., 1993; Lazarus & Folkman, 1984).

. Lower perceived control will be associated with increased use of detachment
coping (Lazarus & Folkman, 1984).

. Lower perceived control will be associated with greater internalized shame, due
to an increased feeling of powerlessness (Kaufman, 1989).

. Greater use of involvement coping will be negatively associated with depression

(Folkman et al., 1993; Lazarus & Folkman, 1984).

10. Greater use of detachment coping will be positively associated with depression

(Folkman et al., 1993; Lazarus & Folkman, 1984).

11. Greater internalized shame will be positively associated with depression

(Kaufman, 1989).

 

42

Multivariate Hypotheses (Path Models):

A pair of path models, based in part on Folkman et al. (1993), will‘be tested (see
Figures 1 and 2, below). Figure 1 represents the central multivariate research hypothesis.
In this model, the structure of the relationships among stress, control, coping, and
depressed mood is based on the path model tested by Folkman and her' colleagues (1993;
see Figure L-l, page 25, above). The model depicted in Figure 1 differs fiom that tested
by Folkman et al. in that physical symptoms and HIV serostatus are explicitly included as
variables in the model, rather than controlled for. It also differs from Folkman’s model in
the inclusion of shame. These three variables—HIV status, physical symptoms, and
internalized shame—are hypothesized to be important precursors to the
cognitive/behavioral stress and coping processes modeled by F olkman and her
colleagues. Finally, the models to be tested differ from Folkman’s in that the direct links
between stress and depression, as well as between control and depression, have been
removed. In F olkman et a1. (1983), the direct link between stress and depression was
empirically observed but was not predicted by the theory. F olkman speculated that the
existence of this direct link indicated that the other variables in her model did not
adequately account for the relationship between stress and depression. It is hoped that,
with the addition of an affect component to the model (i.e., shame), the direct link
between stress and depression will become less important, if not trivial. Similarly, the
direct link between control and depression is not predicted by the theory and was shown
by Folkman and colleagues to be non-significant, so is excluded here as well. Excluding
these direct links provides a stricter test of the theory that the effects of stress and

appraisals of control on depression are not direct, but are mediated by coping behaviors.

43

The model shown in Figure 1 predicts that positive HIV serostatus will be
associated with increased shame, increased stress, and increased physical symptoms.
Increased symptoms will in turn be associated with increased stress, increased shame, and
decreased feelings of being in control. Increased shame will be associated with increased
stress, decreased control, greater use of detachment coping, and lesser use of involvement
coping. Increased stress will be associated with decreased control, greater use of
detachment coping, and lesser use of involvement coping. Decreased control will be
. associated with greater use of detachment coping and lesser use of involvement coping.
Finally, greater use of detachment coping strategies will be positively associated with
increased depressed mood, while greater use of involvement coping strategies will be
negatively associated with depressed mood. This model will be tested for gay subjects
only, since it includes HIV serostatus as a variable.

Figure 2 depicts a model very similar to Figure 1, except that HIV serostatus is not

included. This model will be tested for the heterosexual sample only.

 

Figure 1

filth model for gay sub-sample (HIV status known)

      
  

 

     
  

 

mv INVOLVEMENT
SEROSTATUS COPING
* DEPRESSION
DETACHMENT
5“" "ms COPING
Figure 2

Bath model for heterosexuarl sub-sample (HIV status unknown).

INVOLVEMENT
COPING

 
   
 

DETACHMENT
COPING

 
   
 

 

CONTROL

 
 

 

 

 

REC Rl'l
Statistica
II
size requi
1116 sampl
'each \‘an';
H€€d6d to
7S3. The
COHdIIiO:
t0 recruit

Pfimit dr;

 

II
aPmeim
Sample 0:
Positive
L.
recruited

511mm 3;

 

 

Chapter 2

METHOD

RECRUITMENT OF PARTICIPANTS
Statistical Power Analysis

It has been suggested (Richard DeShon, personal communication) that the sample
size required for sufficient power in structural modeling can be estimated by considering
the sample size needed for suffiCient power to detect the individual correlations between
. each variable in the model. Using guidelines provided by Cohen (1992), the sample size
needed to achieve power of .80 to detect a small (e.g., r = .10) effect with alpha = .05 is
783. The N needed to detect a medium-sized effect (e.g., r = .30)‘under these same
conditions is only 85. While it was not feasible (given limitations of time and resources)
to recruit a sample size of nearly 800, it was possible to recruit a sample large enough to
permit detection of medium-sized effects.

It was proposed that a minimum of 100 gay men be recruited for the study,
approximately 50% HIV-positive and 50% HIV -negative. This was achieved; the gay
- sample consists of 107 men with known HIV serostatus, 50 of whom (47%) are HIV-
positive.

Likewise, it was proposed that a minimum of 100 male college students be
recruited as a comparison sample. This, too, was achieved; the heterosexual college

student sample size is 112.

45

 

 

 

Gay Ma!

 

L'nix'ersi:
approach
Lesbian-i
about 1h:
Associat.
lesbian-b
Coording
DeVeiOp:

The Slud:

 

 

 

 

the list, vi
\1
These 1114

both cm.

C'Onsem

re‘Umin

reqUEsu‘

46

Gay Male Community Sample

HIV -negative gay men were recruited in a variety of ways. Michigan State
University (MSU) campus organizations serving gay students, faculty, and staff were
approached and requests made for their assistance in recruiting subjects. The Alliance of
Lesbian-Bi-Gay and Transgendered Students (ALBGTS) placed a link to a.wn'tten notice
about the study on their World Wide Web page. The Gay-Lesbian Faculty-Staff
Association (GLF SA) maintains an electronic mail (email) list that serves the MSU
lesbian-bi-gay (LBG) community. A notice about the study was posted to this list by LBG
Coordinator Brent Bilodeau, an MSU staff member working within the Multicultural
Development Unit of the Office of the Vice President for Student Affairs and Services.
The study announcement was thereby circulated via email to all individuals subscribing to
the list, with good response.

Written notices were also placed in print publications with a high LBG readership.
These included the Capital Times and Lansing Association for Human Rights Newsletter,
both covering the Lansing area, and the Outpost, covering Ann Arbor, Detroit, and the
rest of southeastern Michigan. In addition, ilyers were posted (with permission) in several
Lansing-area businesses that cater to a LBG clientele. Finally, participants were recruited
through word-of-mouth notification about the study in the Lansing LBG community.

Each survey packet contained a cover letter, the questionnaire (including informed
consent notification), and a pre-addressed business reply envelope to be used for
returning the completed survey. In all, 81 survey packets were mailed to persons

requesting them, and 69 were returned, for a response rate of 85%.

 

names a
survey I;
informs:
themsel'
comple:
name ar.’
survey r;
will be J

HIV-P0

AIDS .\
the Lang,
commUr
Panlcip '~
I0 all of
some). “
"0111mm
One West
Client 0.
“Sing t::
given th.

his COW-L,

I“

 

47

The above recruitment procedures required potential participants to reveal their
names and mailing addresses (and often their email addresses as well) in order to have
survey packets and compensation (discussed below) mailed to them. All such personal
information was securely stored and kept strictly confidential. The survey questionnaires
themselves were identified by unique ID number only, and also securely stored. Once a

I completed survey was returned and the participant’s compensation mailed to him, his
I name and address information was destroyed, unless he requested that a summary of the
survey results be sent to him at a future date, after which time the identifying information
will be destroyed.

HIV-Positive Gay Male Sample

Gay men with HIV/AIDS were recruited primarily through the Lansing Area
AIDS Network (LAAN), a local agency that coordinates services for PWAs throughout
the Lansing area, although several HIV positive men were recruited from the wider
community using the techniques outlined above. The identities of LAAN clients
participating in the study were unknown to the researcher. LAAN staff mailed out letters
to all of their gay male clientele informing them about the study, notifying them that a
survey would be sent to them, and explaining that their participation was completely
voluntary and that they would remain anonymous. Survey packets were then mailed out
one week later. Each survey packet was assigned a unique ID number, matched to each
client by LAAN staff. As each completed packet was returned by mail to the researcher
(using the enclosed pre-addressed business reply envelope), LAAN staff members were
given the survey ID number, which allowed them to identify the participant and mail him

his compensation. One wave of reminder letters was sent out to individuals who had not

 

 

yet reu
LUA‘

Comp

their ti
receive
of goo
purcha
LIAN

surrey

men (.1
men, 21
5610512
COmm'

below

rt‘vcruit

Heter.

This“

Student .

48

yet returned their surveys after three weeks. In all, 70 survey packets were mailed out to
LAAN clients, and 44 were returned, for a response rate of 63%.

Compensation for Gay Male Participants

To provide an incentive for study participation, as well as to compensate them for
their time and effort, all gay male participants who returned a completed survey packet
received a $10 gift certificate7 from Meijer, a local discount chain selling a wide variety
of goods including groceries, clothing, toiletries, and medicine. The gift certificates were
purchased directly from Meij er and mailed to each participant by the researcher (or by a
LAAN staff member if respondent was a LAAN client) upon receipt of his completed
survey.

The final N for the gay male sample was 116, which includes 50 HIV-positive
men (44 LAAN clients and 6 men recruited from the wider community), 57 HIV Qnegative
men, and 9 men who had not been tested for HIV and who therefore did not know their
serostatus. Of the 66 gay men not recruited through LAAN, 63 were recruited from the
community and 3 were identified from within the college student sample, discussed
below.

Appendix A contains an example of written notices used in participant
recruitment.

Heterosexual Male College Student Sample

Male college students were recruited from the MSU Department of Psychology

subject pool. Students received academic credit (psychology experiment credits, part of

 

7 This compensation was made possible in part through a Blue Cross/Blue Shield of Michigan Foundation
Student Award Program grant in the amount of $3,000.

 

the co
studer
respor
homo.

hetero

M

b? 31 it
“‘fli‘ p
CORSET

questic

study I
lhfi tag

above)
Classrt

Student

Iener u-

49

the course requirements for certain psychology courses) for their participation. All 1 15
students who attended the group data collection sessions completed the survey, for a
response rate of 100%. Three of the students indicated in the survey that they are
homosexual, and therefore were added to the gay male sample. The final N for the

heterosexual sample was 112.

PROCEDURE

Because of the sensitive nature of this research topic, participants were required to
be at least 18 years of age to be eligible for the study. Participants were told that they
were participating in a study of health, stress, and emotional well—being. Informed
consent was obtained from all participants using a form included in each survey
questionnaire ( see Appendix B).

For the gay sample, self-administered questionnaire packets were distributed to
study participants by mail (or, less commonly, in person) by the principal investigator. In
the case of LAAN clients, the surveys were distributed anonymously by LAAN staff (see
above).

MSU psychology students completed their questionnaires in large groups.
Classroom space was reserved for these data collection sessions, and students signed up
to attend at their convenience. These sessions typically ran 30-40 minutes, and each
student received his experiment credit immediately upon completion of the survey.

All respondents received a debriefing letter after returning their surveys. This

letter was mailed to gay respondents along with their gift certificates. For the. student

 

...—w-

samplr

sessior

packet
stored
enterec

files or

.’
—-
PF!
J,
CI)

descrih

Demog

Orientat
also ask
Web, ar
11 had b
fimihtar

USCd as

 

50

sample, the debriefing letter was given to each participant at the end of the data collection
session when he turned in his survey and received his credit.

No identifying information was included in the research packets; individual
packets were identified by a unique case number only. Completed research protocols were
stored in a locked filing cabinet located in the home of the principal investigator. Data
entered into electronic form for the purpose of analysis were stored in password-protected

files on a personal computer located in the home of the principal investigator.

MEASURES

An example of the final survey, containing all of the research instruments
described below, can be found in Appendix B.

Demographic Information

Participants were asked for demographic information such as date of birth, sexual
orientation, race/ethnicity, level of education, and household income. Participants were
also asked about frequency of computer usage, access to the Internet and World Wide
Web, and whether they would have been more or less likely to fill out the questionnaire if
it had been available online. These computer-oriented questions were included to
facilitate comparison of these data with future data collected over the Internet, and are not _
used as variables in the current study.

HIV Serostatus

HIV serostatus was determined by subject self-report. Participants were asked
whether they have had the HIV antibody test, the date that the test was most recently
administered, and their current serostatus. HIV -positive participants were also asked

whether they have been diagnosed with AIDS. As expected, a majority (94%) of the

 

heterosexuz
effects and
H1\' Sim;
lo'
physical 5
past year.
experienc
coded, tc
Stress
5
. asks par
extreme
relation
lilneSse
Particir
create ;
for this
{Of One

APillar

felt the)

30mm] u

51

heterosexual college student sample had never been tested for HIV, and therefore the
effects and correlates of HIV serostatus could not be assessed for this group.

HIV Symptoms

Following Folkman et al. (1993), participants were presented with a list of 16
physical symptoms and asked to indicate whether they had experienced each during the
past year. Space was also provided for participants to mention any symptoms they had
experienced that were not included in the list; up to 5 additional symptom mentions were
coded, for a possible range of 0-21 symptoms.

Stress

Similarly to Folkman et al. (1993), stress was assessed using an ll-item scale that

' asks participants to rate on a 4-point Likert scale how stressful (0 = not at all to 3 =
extremely) the following 10 life-domains had felt during the past month: primary
relationships, relationships with. friends, relationships with family, work, school, finances,
illnesses of close others, death of close others, own health, and political issues.
Participants could also nominate and rate an “other” category. Ratings were summed to
create a total stress score (range = 0-33). Test-retest reliability information is not available
for this measure or for the measure of control (see below), as these measures were created
for one specific study (F olkman et al., 1993) and haVe not been widely used.

. Appraisals of Control

For each of the domains rated for stress, participants rated the extent to which they
felt they had control over that domain. Ratings were given on a 4-point scale (0 = almost
never to 3 = almost always). Unlike Folkman and her colleagues (1993), who assessed

control using a single item concerning the most stressful life domain only, control over all

u /_ "I

10 (
an r
reli.
coll
indi

Sll‘l

that
cun
him

she

C0ll

erp
(rat

0111;

nor
C01]

“UH

52

10 (11 if “other” is endorsed) domains was measured, and these ratings summed to create
an overall control score ranging from 0-33. This method was intended to provide a more
reliable and comprehensive measure of control, the lack of which F olkman and her
colleagues noted as a shortcoming of their 1993 study. Participants were still asked to
indicate their most stressfirl life area in order to assess coping within that domain.

Stress and Control Quotients

Participants varied in the number of life areas that were applicable to them and
that could therefore be rated for stress and control. For example, not every participant was
currently attending school, or had recently experienced the death of someone close to
him. These life areas were therefore marked “N/A” and did not contribute to the overall
stress or control summary score. Thus, the total raw scores for stress and control varied
partly as a function of the number of applicable life areas rated. This made it difficult to
compare the raw scores of participants who endorsed different numbers of life areas.

For example, a person for whom all 11 life areas were applicable, and who
experienced “a little” stress (rating = 1) in 10 of those areas and “quite a bit” of stress
(rating = 2) in the 11‘h area would have a total raw score of 12. Another person for whom
i only 4 life areas were applicable, but who rated all four as “extremely” stressful (rating =
3), would also produce a raw score of 12. Are these two individuals really experiencing
an equivalent amount of stress? It is very difficult to say.

To deal with this problem, stress and control quotients were computed. The
numerator of the quotient is the raw score (i.e., the sum of the ratings for stress or
control). The denominator is the total amount of stress or control possible, given the

number of life areas being rated (i.e., the number of life areas multiplied by 3, the highest

 

 

 

&l

and

reer
’hes
Pal“.
ima

COp

itsr

53

possible rating). The result is expressed as a proportion of the total stress or control
possible, given the number of life areas rated (range 0.00 — 1.00). This allowed
comparison of the stress and control scores of participants whose ratings were based on
differing numbers of life areas.
Coping Strategies

C0ping was measured using the Ways of Coping (WOC) questionnaire (F olkrnan
I & Lazarus, 1988), a self-report scale that presents participants with a range of cognitive
and behavioral strategies that may be used in attempting to cope with stressful situations.
The WOC scale used in this study is a modified version developed by F olkrnan and her
colleagues (Folkman, Chesney, Cooke, Boccellari, & Collette, 1994) for use with gay
men. It is slightly longer than the standard WOC (73 vs. 66 questions). This is due in part
to the inclusion of four separate items that assess coping by using food, alcohol,
recreational drugs, and prescription drugs (the standard WOC uses just one item for all of
these). In addition, the modified scale includes items about having sex with one’s primary
partner, having anonymous sex, having non-anonymous sex, and using meditation or
imagery. These additional items presumably make the scale more inclusive of the types of
coping utilized by gay men, but the items are also non—specific enough to be used with a
general population sample. Therefore, participants in both the gay and heterosexual
samples were given the modified WOC questionnaire for the sake of comparability of
responses. I

Participants were asked to rate each item according to the frequency with which
they used that strategy to attempt to cope with their most stressful life area. Ratings were

made on a 4—point scale (0 = does not apply/not used to 3 = used a great deal). Folkman

 

and .
by St
sell-
conf
SUPP
detac
\l'Ol

andr

54

and colleagues’ (1994) scoring rules for the instrument yield eight scale scores, obtained
by summing the ratings for the items that load on each scale: accepting responsibility (or
self—blaming), behavioral escape-avoidance, cognitive escape-avoidance, distancing,
confrontive coping, planful problem solving, positive reappraisal, and seeking social
support. These scales can subsequently be summed into two second-order scales:
detachment coping and involvement coping. See Table 1 for the item composition of the
WOC subscales. See “Results” section (below) for data on intemal/external consistency

and reliability of the WOC scales as well as the other scales used in this study.

55

Table 1
Items comprising the Wayg of Coping subscales.

 

Involvement Coping

Confrontive Coping
1. Tried to get the person responsible to change his or her mind.

2. I expressed anger to the person(s) who caused the problem.
3. I apologized or did something to make up.

4. I let my feelings out somehow.

5. I stood my ground and fought for what I wanted.

Planful Problem-Solving
1. Just concentrated on what I had to do next—the next step.

2. I tried to analyze the problem in order to understand it better.

3. Bargained or compromised to get something positive from the situation.
4. I made a plan of action and followed it.

5. I changed something so things would turn out all right.

6. I knew what had to be done, so I doubled my efforts to make things work.
7. I came up with a couple of different solutions to the problem.

Positive Reappraisal
. Looked for the silver lining, so to speak; tried to look on the bright side of things.

. I told myself things that helped me to feel better.

I was inspired to do something creative.

. Changed or grew as a person in a good way.

. I came out of the experience better than when I went in.
. I found new faith.

. l rediscovered what is important in life.

I changed something about myself.

I prayed.

10. I meditated or used imagery.

ppqomuyna

Seeking Social Support
1. Talked to someone to find out more about the situation.

2. Accepted sympathy and understanding from someone.

3. I talked to someone who could do something concrete about the problem.
4. I asked a relative or friend I respected for advice.

5. I talked to someone about how I was feeling.

 

(Table continues)

 

la‘4llue7

D

IN\12.AJ\U.
s

56

Table 1 (concluded)
Items comprising the Ways of Coping subscales.

 

Detachment CopinfigL

Behavioral Escape-Avoignce
1. I tried to make myself feel better by drinking.
2. I tried to make myself feel better by using recreational drugs, e. 9....
3. I tried to make myself feel better by using prescribed mood-altering drugs, e. g..
4. I took a big chance or did something very risky.
5. I had anonymous sex to feel better.

Cognitive Escape-Avoidance
1. Hoped a miracle would happen.

2. I wished that I could change what had happened or how I felt.

3. I daydreamed or imagined a better time or place than the one I was in.
4. I wished that the situation would go away or somehow be over with.
5. I had fantasies or wishes about how things might turn out.

6. I prepared myself for the worst.

7. I went over in my mind what I would say or do.

' Distanclng
1. Went on as If nothing had happened.

2. I tried to keep my feelings to myself.

3. Tried to forget the whole thing.

4. I waited to see what would happen before doing anything.

5. I didn’t let It get to me; I refused to think too much about it.

6. I made light of the situation; I refused to get too serious about It.

7. I accepted it, since nothing could be done.

8. I tried to keep my feellngs from interfering with other things too much.

Self-Blaming‘I
1. Criticized or lectured myself.

2. I realized I brought the problem on myself.
3. I made a promise to myself that things would be different next time.

 

'This scale called "Accepting Responsibility” by Folkman.

in

Pa
sui
als

re;

We:

57

According to F olkrnan and Lazarus (1985), and Folkman (personal communica-
tion), it is inappropriate to speak of test-retest reliability for the WOC because of the
process-oriented nature of the construct it purports to measure. The instrument is not
meant to measure stable coping traits or styles, and it is expected that a person’s coping
scores on the WOC will change from situation to situation and across time as the degree
of threat of particular stressors waxes and wanes. There are some, however, who take
issue with this assertion (Parker, Endler, & Bagby, 1993).

Intemalized Shame

Shame was measured using the Intemalized Shame Scale (188; Cook, 1994), a 30-
item self-report questionnaire that focuses on the phenomenological experience of
.intemalized shame. Participants were asked to rate how often they feel or experience what
is described in each item. Ratings were made on a 5-point Likert scale (0 = Never to 4 =
Almost Always). The instrument consists of 24 negatively-worded shame items and 6
positively-worded “self-esteem” items included primarily to reduce negative respOnse set.
Factor analyses have consistently yielded one unitary shame construct derived from
summing the response categories of the 24 shame items (Cook, 1994). Investigators may
also use the 6 self-esteem items as a separate self-esteem scale if they wish. Cook (1994)
reported good internal consistency for the shame scale with respect to both clinical (alpha
.= .96) and non-clinical (alpha = .95) samples. He reported a median item-total correlation
of .70 in the clinical sample and .63 in the non-clinical sample. Cook also reported a 7-

week test-retest reliability of .84 in a subset of the non-clinical sample.

 

58

Depression

Depressive symptomatology was measured using the Center for Epidemiologic
Studies Depression Scale (CBS-D; Radloff, 1977), a 20-item self-report scale designed to
measure current depressive symptomatology in the general population. The CES-D items,
culled from previously validated depression scales, were selected to emphasize the
affective (as opposed to vegetative) aspects of depression, which include depressed mood
and feelings of guilt, worthlessness, helplessness, and hopelessness, though psychomotor
retardation and disturbances in sleep and appetite are also assessed.

Participants were instructed to indicate how often during the past week they felt or
behaved in the way described by each item. Ratings were made on a 4-point scale,
ranging from. 0 = “rarely or none of the time” (< 1 day/past week) to 3 = “most or all of
the time” (5-7 days/past week). Ratings were summed to obtain an overall depression
score ranging from 0-60. The scale contains 4 positively-worded items that are reverse-
coded when calculating the total depression score.

Radloff (1977) reported good internal consistency for the CBS-D. Across four
standardization samples (three general population samples and one clinical sample), item-
.whole correlations ranged from .30 to .79; 80-95% of the correlations were .40 or greater.
Alphas across the four samples ranged from .84-.90, and split-half reliabilities ranged
from .76 to .85. The highest internal consistency was found within the clinical sample.
Test-retest reliabilities were more moderate, ranging from a high of .67 at four weeks to a
low of .32 at twelve months. Radloff attributes these moderate test-retest correlations to
the fact that the CES-D was designed to measure current symptomatology, which varies

over time.

Chapter 3

RESULTS

PRESENTATION OF RESULTS

A number of conventions that may be less familiar to some readers are used in this
study for the presentation of results. These conventions are described below.

Confidence Intervals

Following; Hunter and Schmidt (1990), confidence intervals are used in place of
most traditional tests of significance in this study. The confidence interval provides the
same information as the traditional significance test, in that one may accept or reject the
null hypothesis depending upon whether or not the hypothesized value of the null
hypothesis is contained within the observed interval. In addition, the confidence interval
is correctly centered about the observed value rather than a hypothetical null value, and
provides information about the likelihood of committing a Type H error (see “Inference
Probabilities and Odds Ratios,” below).

In keeping with the tradition of the .05 alpha level for statistical significance, a
95% two-sided confidence interval will be used to test non-directional (exploratory or
two-tailed) hypotheses. For testing directional (confirmatory or one-tailed) hypotheses, a

90% two-sided (i.e., equivalent to a 95% one-sided) confidence interval will be used.

59

 

 

:2

ha

Stir".

Pres

lead

60

Inference Probabilities and Odds Ratios

When a sample correlation is small in magnitude and/or the sample size is small,
the sample correlation may be determined to be “non-significant” when in fact the true
population correlation is non-zero (i.e., a Type II error). When this situation occurs, zero
is contained within the confidence interval, but may be located very near to one of the
boundaries of the interval. This suggests that zero is an improbable choice for the true
population value, since values at the extremes of the confidence interval are less probable
estimates of the true population parameter. In such cases, calculation of the inference
probability and odds ratio can provide additional information regarding the sign of the
true population value. The inference probability (IP) can be defined as the probability that
the true population parameter lies in the predicted direction (negative or positive).
Assuming a normal distribution centered about the sample correlation, the IP is the area
under the normal curve above (or below, if the correlation is predicted to be negative) a z-
score of 0 (i.e., the null hypothesis).

The odds ratio is simply the ratio of the inference probability to the reverse
probability l-IP; it is the odds that the directional hypothesis is true. Note that the IP and
odds ratio make sense only with respect to directional (i.e., confirmatory or one-tailed)
hypotheses.

Levine (class notes, p. 7-21) suggests a rule of thumb that an odds ratio greater
than or equal to 2:1 (corresponding to an IP of .68 or higher) should be considered
sufficient support for a directional hypothesis. This rule of thumb is observed in the
present study, though the traditional 95% significance level is used as well to allow the

reader to decide whether a given directional hypothesis is supported by the data.

 

61

The d-Statistic for the Difference in Means

Whenever a difference in means is tested, a d-statistic will be calculated in place
of the traditional t or F to facilitate interpretation. The d—statistic is a standard score (like
a z-score) and thus a straightforward measure of effect size because it expresses the

difference in means in standard deviation units. The formula for d is as follows:

 

where D is the absolute value of the difference in means, and v, and v; are the variances
of the two groups being compared.

Correcting for Attenuation

Hunter and Schmidt (1990) have demonstrated that error of measurement in either
variable systematically attenuates the magnitude of sample bivariate correlations. Because
this effect is systematic, it is correctable if the amount of measurement error can be
estimated. Therefore, all sample correlations in this study have been corrected for

attenuation using the following formula:

for

r corrected = ——
Jar-J05:

where rer is the observed correlation, and 5 J5; is the product of the square roots of
the reliabilities (alphas) for variables x and y. Where alpha is unknown, it is assumed to
be equal to 1.00 (i.e., perfect measurement), a conservative assumption in the, sense that it

reduces the amount of correction provided by the formula.

 

62

If correction for attenuation is applied to a sample correlation, it must also be
applied to the upper and lower boundaries of the confidence interval around that
correlation. This insures a correct interpretation of whether or not the correlation is
statistically significant. The formula used to correct the confidence interval is analogous
to that used to correct the correlation coefficient, replacing the uncorrected correlation
coefficient r,ry with the uncorrected upper or lower boundary of the confidence interval

(both boundaries must be corrected).

SCALE PSYCHOMETRICS

 

The composition of (and rationale for) the scales used in this study are described
above in the “Measures” section of the “Methods” chapter. Table 2 presents the results of
a confirmatory factor analysis of these scales using Hunter and Hamilton’s (1992)
statistical program CFA. Hunter and Hamilton’s program yields the traditional reliability
coefficient (alpha) as well as the average inter-item correlation for each scale. In addition,
the program provides tests of “unidimensionality” (i.e., internal consistency, or the degree
to which items in the scale measure a single construct) and “parallelism” (i.e., external
consistency, or the degree to which the items in the scale relate in a consistent way to
items in other scales). The tests for unidimensionality and parallelism are each performed .
with and without the assumption of uniform or “flat” item quality (i.e., the correlation
between an item and the construct it purports to measure). The results of these tests are
expressed as chi-squares that indicate deviations from unidimensionality and parallelism.

Thus, non-significant values of chi-square are desirable.

63

Table 2
Results of confirmaftoryfjctor analysis of t_he study scales (N = 228).

 

 

Average

Inter-item
Scale Name No. of items Alpha correlation
Stress 1 1 .74 . .21
Control. _ 11 .75 .21
Physical Symptoms 20 .86 i ' .23
WOC”: Confrontlve Coping 5 .66 .28
WOC: Planful Problem-Solving 7 .72 .27
WOC: Positive Reappraisal 10 .76 .24

WOC: Seeking Social Support . 5 .77 .40
WOC: Behavioral Escape-Avoidance 5 .63 .26
WOC: Cognitive Escape-Avoidance 7 .81 ' .37
WOC: Distanclng 8 .73 .26
. WOC: Self-Blaming 3 .65 .38

4

4

WOC: Involvement Copingc .73 .40

WOC: Detachment Copingd .64 .31
Intemalized Shame Scale _ 24 .96 .51
Depression (CES-D°) 20 .93 .39

 

'p < .05; “p < .01; “‘p< .001

'Values are Chi-squares; significant values indicate deviations from unidimensionality/parallelism.
Ways of Coping.

cSecond-order scale composed of Confrontlve Coping, Planful Problem-Solving, Positive

Reappraisal. and Seeking Social Support.
dSecond-order scale composed of Behavioral Escape-Avoidance, Cognitive Escape-Avoidance.

Distanclng, and Self-Blaming.

°Center for Epidemiologic Studies Depression Scale.

(Table continues)

Ta

Isle.”

a.”

 

64

Table 2 (concluded)

Results of confirmatogy factor analysis of the study scales (N = 228).

 

Deviation from Deviation from

 

 

 

 

 

unidimensionality‘ parallelism'
Item quality Item quality
Flat Gradient . Flat Gradient
362.46” 243.74” 263.43*** 311 16*”
508.63""""r 486.01“ 121.04* 129.30“
1430.84” 771 .67” 22.34 ns 15.68 ns
8.75 ns 0.96 ns 100.04“ 126.09“
57.08“” 40.56""r 98.50“ 98.28”
121.57“ 70.77” 150.22“ 175.43”
63.49”" 23.07” 32.43 ns 42.19 ns
42.58”" 14.60 ns 77.32*** 52.86"
174.16*** 77.91“" 7824* 105.03***
95.23“" 59.42“ 1 17.56“" 135.45”
2.17 ns 0.00 ns 30.36* 4520*"
5.79 ns 4.53 ns 27.28 ns 30.36"
22.09“" 5.19 ns 63.09“” 19.05 ns
3063.86” 1561.52” 85.31“" 63.67“
1856.90“ $53.84*” 136.07*** 81 .44***

 

JO < .05: ”p < .01; "‘p < .001

 

'Values are Chi-squares; significant values indicate deviations from unidimensionality/parallelism.
I’Ways of Coping. . '
cSecond-order scale composed of Confrontlve Coping, Planful Problem-Solving, Positive

Reappraisal, and Seeking Social Support.
“Second-order scale composed of Behavioral Escape-Avoidance, Cognitive Escape-Avoidance,

Distancing. and Self-Blaming.
°Center for Epidemiologic Studies Depression Scale.

65

As Table 2 indicates, alphas for the scales range from .63 for WOC Behavioral Escape-
Avoidance to .96 for the ISS. Alphas for the WOC subscales tended to be lower than for
the other scales, probably due to the relatively small number of items in these subscales.
Average inter-item correlations ranged from .21 for the stress and control scales (perhaps
reflecting the relative independence of the various life areas rated for stress and control)
to .51 for the ISS.

Given these respectable alphas, it is surprising to note how poorly most of these
scales performed on the “stricter” chi-square tests of internal and external consistency
(see Table 2). Even the seemingly robust ISS, which had the highest coefficient alpha and
inter—item correlation of all the scales, failed the tests for unidimensionality and
parallelism, regardless of whether uniform or gradient item quality was assumed. The
only scales that passed both tests were the second-order Detachment and Involvement
Coping scales. These results may reflect the practice of relying on coefficient alpha as the
primary measure of internal consistency/reliability during scale construction and
validation. Since the value of coefficient alpha is based in part on the number of items in
the scale, scales with large numbers of items may yield high alphas, yet still demonstrate
poor internal and external consistency when stricter tests are used.

Attempts were made to improve the scales’ internal and external consistency by
deleting weaker items, but in no case did this improve any scale sufficiently to pass the
unidimensionality and parallelism tests. The scales were therefore left as—is in the interest
of preserving comparability to other studies using these scales. Hence, the following
results should be interpreted with the caveat that these scales may not measure unitary

constructs.

 

for t
edut‘
the l
samr

The

are 5
Was

Hl

66

CHARACTERISTICS OF THE STUDY SAMPLES

Table 3 presents basic demographic data (age, ethnicity, education, and income)
for the two study samples. As the table shows, the samples differed markedly in age, 5
education, and annual income.8 This was to be expected, given the student composition of
the heterosexual sample. This sample was younger and less variable in age than the gay
sample, averaging 19.3 years (SD = 2.7 years) compared to 36.6 years (SD = 9.6 years).
The college student sample was (by definition) uniform in education level, while the gay

sample showed considerable variability and was skewed toward the upper range (not

 

surprising in a university community). The income distribution of the heterosexual
sample was skewed toward the upper range, probably because most of these respondents
are still being supported by their parents, while the income distribution of the gay sample
was skewed toward the lower range, reflecting the fact that 78% (39 out of 50) of the
HIV-positive men fell into the two lowest income categories. The two samples did not
differ in ethnicity, with 84% of both samples self-identifying as Caucasian. Due to the
small number of respondents falling into the other ethnicity categories, ethnicity was

collapsed into White and Non-White for the purpose of statistical analysis.

 

8 Formal statistical comparisons of the two samples are found in Tables 6-9.

67

 

 

 

 

 

 

 

 

 

 

 

 

Table 3
Descriptive statistics for the study samples: demographic variables.
Gay Male Heterosexual Male
Sample ~ Sample
Variable (N = 116) (N = 112)
Mean Std. Dev. Range Mean Std. Dev. Range
Age 36.6 9.6 . 18' - 68 19.3 2.7 17 - 45
N % Cum. % N °/o Cum. %

Ethnicity
African Am. 3 2.6 2.6 3 2.7 2.7
Asian 2 1.7 4.3 8 7.1 9.8
Caucasian 97 83.6 87.9 94 83.9 93.7
Latino 8 6.9 94.8 2 1.8 11.6
Mid. Eastern 1 0.9 95.7 2 1.8 13.4
Native Am. 2 1.7 97.4 1 0.9 14.3
Other 3 2.6 100.0 2 1.8 16.1
Total 116 100.0 112 100.0

Education
< HS Grad. 1 0.9 0.9 0 0.0 0.0
HS Grad. 11 11.2 12.1 0 0.0 0.0
Some College 48 39.6 51.7 107 95.5 95.5
College Grad. 24 20.7 72.4 5 4.5 100.0
Adv. Degree 32 27.6 100.0 0 0.0 100.0
Total 116 100.0 112 100.0

Annual Income
< $10K 30 25.9 25.9 7 6.3 6.3
$10-30K 42 36.2 62.1 7 6.3 12.5
$30-60K , 28 24.1 86.2 32 28.6 41.1
$60-100K 8 6.9 93.1 40 35.7 76.8
> $100K 8 6.9 100.0 26 23.2 100.0

 

 

Total ‘ 116 100.0 112 100.0

 

68

Table 4 presents descriptive statistics related to HIV/AIDS testing and diagnosis.
It can be seen that over 92% of the gay sample had been tested for HIV, compared to less
than 15% of the heterosexual sample. Of the 107 gay men with known HIV serostatus, 50
(46.7%) were HIV-positive; of these, 31 (62%) had been diagnosed with AIDS. None of

the 16 heterosexual men with known HIV serostatus were HIV -positive.

69

Table 4
Descriptive statistics for the study samples: HIV/AIDS testing and diagnoses.

 

 

 

 

 

 

 

 

 

 

 

Gay Male Heterosexual Male
Sample Sample
Variable (N = 116) (N = 112)
N % Cum. % N % Cum. %
Been tested '
for HIV? ~
Yes 107 92.2 92.2 16 14.3 14.3
No 9 7.8 100.0 94 83.9 98.2
Don't Know 0 0.0 100.0 2 1.8 100.0
Total 116 100.0 112 100.0
IF YES:
HIV Positive?

' Yes so 46.7 46.7 o 0.0 0.0
No 57 53.3 100.0 14 87.5 87.5
Don't Know 0 0.0 100.0 2 12.5 100.0
Total 107 ‘100.0 16 100.0

IF YES:
AIDS Diagnosis?
Yes 31 62.0 62.0
No 17 34.0 96.0 NIA
Don't Know 2 4.0 100.0
Total 50 100.0

 

70

Descriptive statistics for physical symptoms, stress, control, coping, shame, and
depression are presented in Table 5. At a glance, the table indicates that gay respondents’
scores on these variables are more elevated, more variable, and have a wider range than
those of the heterosexual respondents. Formal statistical comparisons between samples on

these variables are presented in Table 9.

Table 5

71

Descriptive statistics for the study samples: physical symptoms, stress,

controlI copingI shameI and depression.

 

 

 

 

Gay Male Heterosexual Male
Sample Sample

Variable (N = 116) (N = 112)

Mean so Min. Max. Mean so Min. Max.
Physical , , . -
Symptoms 3.11 3.87 .00 16.00 .99 1.53 .00 6.00
Stress
Quotient .44 .1 8 .04 1 .00 .34 .14 .04 .70
Control
Quotient .51 .19 .05 .97 .59 .17 .23 1.00
Involvement .
Coping 33.48 13.05 3.00 64.00 31.13 10.88 3.00 63.00
Detachment
Coping 23.87 10.90 .00 55.00 23.09 9.94 3.00 54.00
Shame 36.26 21.33 .00 87.00 29.79 19.01 3.00 86.00
Depression 18.84 13.27 .00 53.00 16.76 10.55 .00 48.00

 

72

Tables 6-9 show the results of statistical comparisons of the study samples. Table
6 presents a crosstabulation of the two samples by ethnicity (collapsed categories). Not

surprisingly, the samples did not differ in ethnic composition (kappa = -.003, p < .05).

Table 6
Comparison of the study samples by ethnicig (collapsed categories).

 

 

 

 

95% 2-slded
Gay Hetero. Confidence Interval 5‘9-
Ethniclty Sample Sample Total Kappa 55in». Lower Upper (p < .05)
White 97 94 191 -.003 .05 -.10 .09 ns
Non-White 19 18 37
Total 1 16 1 12 228

 

Tables 7 and 8 compare the samples on education and income, respectively. In
both cases the samples differed significantly. The gay men evinced more variability and
higher attainment in education than the heterosexual men, which as mentioned above is a
sampling artifact. The gay men also reported lower annual household income than did the
heterosexual men, which as previously noted probably reflects ongoing financial support

of the college students by their parents as well as low income among HIV-positive gay

men.

73

Table 7
Comparison of the study samples by education.

 

Gay Hetero.

 

 

 

 

Education Sample Sample Total

< HS Grad. 1 0 , 1

HS Grad. 11 0 ' ' 11

Some College 48 107 155

College Grad. 24 5 29

Adv. Degree 32 0 32

Total 116 112 228
value df sig. (2-tailed)

Chi-square 78.86 4 .000

Table 8

Comparison of the study samples 9y income.

 

Gay Hetero.

 

 

Income Sample Sample Total '

< $10K 30 7 37
$10-30K 42 7 49
$30-60K 28 32 60
$60-100K 8 40 48

> $100K 8 26 34
Total 1 16 1 12 228

value df sig. (2-tailed)

 

Chi-square 70.38 4 .000

 

74

Table 9 compares the study samples on age, physical symptoms, stress, control,
involvement coping, detachment coping, shame, and depression. As the table indicates,
the gay men were significantly older on average than the heterosexual men (d = 2.46, p <
.05), which again is attributable to sampling.

The gay men as a group reported significantly more physical symptoms than the
heterosexual men (3.11 symptoms versus .99 symptom, respectively; d = .72, p < .05).
This effect is almost certainly attributable to the effects of HIV/AIDS within the gay
sample; in fact, the HIV -negative gay men reported fewer symptoms on average than the
heterosexual men (.75 versus .99 symptom, respectively), while the HIV -positive men

(not diagnosed with AIDS) reported 3.77 symptoms and men with AIDS 7.77 symptoms.

Table9
Com ri n t stu sam les: a e slcals m oms stress control co in shame ndd ession.

 

 

 

 

Sample 95% 2-slded
Gay Hetero. . Confidence Interval . Sig.

Variable (N - 116) (N -112) OWN“ San Lower Upper d (P < .05)
M0 M 36.57 19.25 17.32 .93 15.49 19.15 2.46 sig

so 9.58 '2.71
Physical M 3.11 .99 2.12 .39 1.35 2.89 .72 sig
Symptoms SD _ 3.87 1.53
Stress M .44 .34 A .10 .02 .06 .14 .62 sig
Quotient so .18 .14
Control iii .51 .59 .08 .02 .03 .12 .42 sig
Quotlent so .19 .17
Involvement M 33.48 31.13 2.35 1.59 -.79 5.49 .20 ns
Coping so 13.05 10.88
Detachment iii 23.87 23.09 .78 1.38 -1.95 3.50 .07 ns
Coping SD 10.90 9.94
Sham M 36.26 29.79 6.47 2.68 1.19 11.75 .32 sig

so 21.33 19.01

M 18.84 16.76 2.08 1.59 -1.05 5.22 .17 ns
0"" ”m" so 13.27 10.55

 

aAbsolute value of the difference in means.
°Effect size; the magnitude of the difference in means expressed in standard deviation units.

75

Table 9 indicates that the gay men reported significantly greater stress and
significantly lower appraisals of control than did the heterosexual men. The gay men’s
mean stress quotient was .44, meaning that on average they experienced 44% of the
amount of stress possible given the number of life areas they endorsed, compared to 34%
for the heterosexual men (d = .62, p < .05). Further, the gay men felt in control of
stressful situations only about half of the time (mean control quotient = .51), compared to
59% of the time for the heterosexual mm (d = .42, p < .05).

Once again, these findings’appear to be related in part to the HIV/AIDS-status of
the gay men. When the gay sample is divided according to HIV serostatus, HIV -negative
.respondents reported significantly less stress than the HIV -positive men (.39 versus .48,

‘ respectively; d = .51, p < .05), though still not as little as the heterosexual men (.39 versus
.34, respectively; d = .38, p < .05). Likewise, HIV-positive men reported less control than
HIV -negative men (.46 versus .55, respectively; d = .48, p < .05); in this respect the HIV-
negative‘gay men did not differ significantly from the heterosexual men (.55 versus .59,
respectively; d = .18, ns). Interestingly, no significant differences in stress or control
emerged when the HIV -positive group was divided according to absence versus presence
of AIDS diagnosis.

Continuing with Table 9, the two samples did not differ significantly in the
amount of involvement and detachment coping strategies used to deal with stress. Both
groups tended to use involvement coping strategies more frequently than detachment

coping strategies.

76

Table 9 indicates that the gay men reported significantly greater internalized
shame than did the heterosexual men, with shame scores of 36.3 versus 29.8, respectively
(d = .32, p < .05). Once again, this difference disappears when HIV serostatus is taken
into account, with the HIV-negative gay men reporting a level of shame similar to that of
the heterosexual men (32.4 versus 29.8, respectively; d =. .14, ns).

Finally, as Table 9 indicates, the two samples did not differ significantly in mean
depression level, though the gay men were slightly more elevated than the heterosexual
men on this measure. This relationship will be examined in more detail below (see

analysis for Hypothesis 2).

RELATIONSHIPS BETWEEN DEMOGRAPHIC FACTORS AND

_S_U_BSTANTIVE VARIABLES

Tables 10 and 11 examine the correlations between the demographic variables
age, education, and income, and the substantive variables HIV serostatus (gay
respondents only), physical symptoms, stress, control, involvement coping, detachment
coping, shame and depression. Table 10 explores these relationships for the gay sample,
and Table 11 for the heterosexual sample.

Looking first at Table 10, it can be seen that age is unrelated to the substantive
variables for the gay sample. Education and income, however, are associated with many
of these variables. More educated men were less likely to be HIV-positive, and reported
fewer physical symptoms, than less educated men. More educated men also reported less
stress and greater feelings of control over stressfiil situations than less educated men.

Perhaps due to their decreased stress, more educated men reported using fewer coping

77

strategies (both detachment and involvement) than less educated men. Finally, more
educated men reported less depression than less educated men. The only variable not
significantly related to education among these men was shame, though there was a non-

significant trend for more educated men to report less shame than less educated men.

 

Table 10

Correlational analysis between demographic variables (age, education, income)

and substantive variables (gay sample; N = 107*).

78

 

 

 

 

95% 2-sided
Demographic Confidence Interval' Sig.
Variable 1” SE, Lower Upper N (p < .05)
Age
HIV Serostatusb .12 .10 -.07 .31 107 ns
Physical Symptoms .05 .10 -.15 .26 107 ns
Stress Quotient -.06 .10 -.28 .16 107 ns
Control Quotient -.13 .10 -.34 .09 107 ns
Involvement Coping .01 .10 -.22 .23 107 ns
Detachment Coping -.12 .10 -.36 .11 107 ns
Shame -.10 .10 -.29 .09 107 ns
Depression -.02 .10 -.22 .18 107 ns
Education
HIV Serostatus -.47 .08 -.61 -.32 107 sig
Physical Symptoms -.39 .08 -.56 -.21 107 . sig
Stress Quotient -.27 .09 -.48 -.06 107 sig
Control Quotient .31 .09 .1 1 .51 1 O7 sig
Involvement Coping -.23 .09 -.44 -.02 107 ' sig
Detachment Coping -.25 .09 -.47 -.02 107 sig
Shame -.17 .09 -.36 .02 107 ns
Depression -.26 .09 -.44 -.07 107 sig
Income

HIV Serostatus -.28 .09 -.46 -.11 107 sig
Physical Symptoms -.39 .08 -.57 -.22 107 sig
Stress Quotient -.30 .09 -.50 -.10 107 sig
Control Quotient .35 .08 .1 6 .55 1 07 sig
Involvement Coping -.19 .09 -.41 .02 107 ns
Detachment Coping -.22 .09 -.45 .01 107 ns
Shame -.27 .09 -.45 -.09 107 sig
Depression -.33 .09 -.51 -.16 107 sig

 

'These analyses exclude gay men who have not been tested for HIV.

'Corrected for attenuation.
”HIV-Negative = o; HIV-Positive = 1.

 

79

Similarly, Table 10 shows that higher income was associated with HIV-negative
serostatus, fewer physical symptoms, less stress, greater control, less shame, and less
depression among the gay respondents. Income was not significantly related to the coping
strategies, though there were non-significant trends toward higher income being
associated with decreased use of both kinds of coping.

These findings might suggest that education and income were “protective” factors
that buffered the effects of stress and facilitated adjustment. However, the effects of
. education and income may also be explained by the nature of the sample, particularly
with respect to the HIV-positive men. The majority of the HIV-positive men in the study
were recruited through an HIV/AIDS service organization, and it is likely that poorer and
less-well-educated persons are over-represented among the clientele of this organization.
If a true community sample of gay men with HIV/AIDS were recruited, the relationship
of income and education to physical and emotional well-being might differ from what is
observed in the present data. In any case, the effects of income and education were
controlled for in the multivariate path analyses for gay men.

Table 11 presents these data for the heterosexual men. Age and education were
unrelated to the substantive variables, not surprising given the lack of variability in age
and education in the sample. Income was related to physical symptoms and detachment
coping, with higher income associated with fewer physical symptoms and decreased use
of detachment coping strategies. This may be because higher income is associated with
better access to health care and other resources that make it less necessary to detach in

order to cope. Income was controlled for in the path analyses for heterosexual men.

80

Table 11

Correlational analysis between demographic variables (age, education, Income)
and substantive variables (heterosexual sample; N = 112).

95% 2-slded

 

 

 

Demographic Confidence Interval' Sig.
Variable r' SEr Lower Upper N (p < .05)
Age
Physical Symptoms -.09 .09 -.29 .1 1 1 12 ns
Stress Quotient .08 .09 -.13 .30 1 12 ns
Control Quotient -.13 .09 -.34 .08 1 12 ns
Involvement Coping -.14 .09 -.35 .08 112 ns
Detachment Coping .15 .09 -.07 .38 112 ns
Shame .06 .09 -.13 .25 112 ns
Depression .02 .09 -.17 .22 112 ns
Education
Physical Symptoms -.16 .09 -.36 .03 112 ns
Stress Quotient .06 .09 -.16 .27 112 ns
Control Quotient .01 .09 -.20 .23 1 12 ns
Involvement Coping -.03 .09 -.25 .19 1-12 ns
Detachment Coping .01 .09 -.22 .24 1 12 ns
Shame -.06 .09 -.25 .12 112 ns
Depression .01 . .09 -.19 .20 112 ns
Income

Physical Symptoms -.30 .09 -.48 -.1 1 112 sig
Stress Quotient .01 .09 -.21 .22 112 ns
Control Quotient .10 .09 -.12 .31 1 12 ns
Involvement Coping -.13 .09 -.34 .08 1 12 ns
Detachment Coping -.30 .09 -.51 -.09 1 12 sig
Shame -.12 .09 -.30 .07 112 ns
Depression -.06 .09 -.25 .13 1 12 ns

 

3Corrected for attenuation.

81

Tables 12-14 examine the relationship between ethnicity and the substantive variables.
Table 12 presents a crosstabulation of ethnicity and HIV serostatus within the gay sample.
As the table indicates, ethnicity was unrelated to HIV serostatus within this sample
(kappa = -.02, us). This is an interesting finding, given the fact that ethnic minorities
(particularly African Americans and Latinos) are over-represented among persons living
with HIV /AIDS in the US. If the sample had included more people of color or had been
drawn from a larger urban area, these results may have been quite different. In their large
multiracial sample of gay and bisexual men living in the Chicago area, DeMarco, Ostrow,
DiFranceisco, and Halman (1997) found that Afiican Americans were over-represented
among HIV -positive respondents. This is the case nationally as well; African Americans.
represent 36% of all US AIDS cases, while making up only about 15% of the population

(Centers for Disease Control, 1997).

Table 12
Crosstabulation of ethnlclt colla sed cate ories b HIV serostatus a sam Ie' N =107“ .

 

 

 

 

95% 2-slded
HIV Serostatus Confidence Interval Sig.
Ethnlclty Negative Positive Total Kappa 35mm Lower Upper (p < .05)
Non-White 9 9 . 18 -.02 .07 -.16 .11 ns
White 48 41 89
Total 57 50 107

 

 

. 'This analysis excludes gay men who have not been tested for HIV.

82

Table 13 examines the differences between white and non-white gay men with
respect to physical symptoms, stress, control, coping, shame, and depression. As the table
shows, the only significant differences between ethnic categories within the gay sample
was in the use of coping strategies. Non-white gay men reported more fretwent use of
both involvement (d = .72) and detachnient (d = .50) coping strategies. This may reflect
the greater amount of stress reported by the non-white gay men, though this difference in

stress level (d = .41) fell short of statistical significance.

Tablo 1 3

Diffegncpg betwgn ethnic groups (collapsed catpgories) on variables to be used In the path ana)ysg
(ggy gmple; N I 107').

 

 

 

 

Ethnic category 95% 2-slded
Non-White White Confidence Interval Sig.
Varlable (N - 18) (N - 89) ”WNW" SEom Lower Upper 6" (p < M)
Physical M 3.67 3.24 .43 1.03 -1.58 2.44 .10 ns
Symptoms SD 4.61 3.84
Stress M .50 ' .42 .08 .05 -.01 .17 .41 ns
Quotient SD .21 .18
Control M .50 .51 .01 .05 -.09 .11 y .05 ns
Quotlent SD .21 .19
Involvement M 41.39 32.32 9.07 3.21 2.77 15.37 .72 sig
Coplng SD 12.63 12.40 '
Detachment M 28.72 22.72 6.00 2.78 .55 11.45 .50 sig
Coplng SD 13.49 10.15
Shame M 32.33 35.94 3.61 5.39 -6.95 14.17 .16 ns
SD 24.41 20.08

D. Ion M 19.89 18.54 1.35 3.46 -5.44 8.14 .10 ns

9"“ so 13.49 13.39

 

'These analyses exclude gay men who have not been tested for HIV.
'Absolute value of the difference in means.

l’Effect size; the magnitude of the difference in means expressed in standard deviation units.

83

Turning next to Table 14, it can be seen that no differences on these variables

existed between ethnic categories within the heterosexual sample.

 

 

 

 

Table 14
Dlflflw between ethnic groups (collapsed categories) on variables to be used in the mth analyses
(hfimgxgl gmplp; N I 112). -
Ethnic category 95% 2-slded, , .
Non-White White Confidence Interval Sig.
Variable (N - 18) (N - 94) ”WWW SEorr Lower Upper 6" (P < -°51
Physical M 1.00 .99 .01 .39 -.76 .78 .01 ns
Symptoms SD 1.41 1.56
Stress M .29 .35 .06 .03 -.01 .13 .48 ns
Quotient SD .10 .14
Control M .61 .58 .02 .04 -.06 .11 .14 ns
Quotient SD .15 .18
Involvement M ' 32.53 30.86 1.66 2.81 -3.84 7.17 .16 ns
Coplng SD 9.96 1 1 .08
Detachment M 21.67 23.36 1.69 2.57 -3.33 6.72 .16 ns
Coping SD 10.82 9.81
Shame M 30.20 29.71 .49 4.91 -9.14 10.12 .03 ns
SD 15.84 19.63
De ‘0" M 17.67 16.59 1.08 2.72 -4.26 6.42 .10 ns
9"“ so 11.63 10.39

 

'Absolute value of the difference in means.
I’El‘fect size; the magnitude of the difference in means expressed in standard deviation units.

84

TESTS OF BIVARIATE HYPOTHESES

Tables 15-25 present the results of tests of the 11 bivariate hypotheses outlined
above. Note that 90% 2-sided confidence intervals are used to test these directional (i.e.,
l-tailed) hypotheses; this is equivalent to using a 95% l-sided confidence interval (i.e.,

alpha level = .05).

Hypothesis 1: The gay sample will evince greater internalized shame than the
heterosexual sample.

Table 15
Test of bivariate hyppthesis 1: The gay men will evince greater Intemalized shame than the heterosexual

 

 

 

 

 

 

 

 

 

 

 

 

Group 90% 2-slded
Gay Hetero. Confidence Interval Odds Sig.
Varlable (N I 107) (N I 112) Diff.‘ SEom Lower Upper 11" IP‘ Ratlo‘ (p < .05)
M 35.34 29.79 5.55 2.69 1.10 9.99 .28 .98 49.00 sig
Shame
SD 20.79 19.01
Group 90% 2-slded
HIV- Gay Hetero. Confidence Interval Odds Sig.
(N I 57) (N I 112) Diff. SE0... Lower Upper d IP Ratio (p < .05)
M 32.36 29.79 2.57 3.04 -2.46 7.59 .14 .80 4.00 ns
Shame
SD 17.99 19.01
Group 90% 2-slded
HIV+ Gay Hetero. Confidence Interval Odds Slg.
(N I 50) (N I112) Diff. SEorrr Lower Upper d IP Ratio (p < .05)
SI M 38.73 29.79 8.94 ’ 3.75 2.69 15.19 .42 .99 99.00 519

SD 23.30 19.01

 

'Value of the difference in means.

'Elfect size; the magnitude of the difference in means expressed in standard deviation units.

‘lnference Probability. the probability that the true population difference is positive.

‘Ratio of the inference probability (IP) to the reverse probability (MP); the odds that the directional hypothesis is true.

As the table indicates, gay men as a group reported over a fourth of a standard
deviation ((1 = .28) more shame than heterosexual men. This difference was statistically
significant and the [P was .98, corresponding to an odds ratio of about 49: 1.

However, as noted earlier, the magnitude of the difference in shame between gay

and heterosexual men varied according to the HIV serostatus of the gay men. As Table 15

85

shows, the mean‘shame score for HIV-negative men was 32.36, compared to 29.79 for
heterosexual men (d = .14, ns; IP = .80). This difference was not statistically significant,
though the odds are still 4:] that the true population difference is positive (IP = .80). In
contrast, when HIV -positive gay men were compared to heterosexual men, a more
pronounced difference emerged, with HIV-positive gay men on average reporting about 9
points more shame than heterosexual men ((1 = .42, p < .05.; IP = .99).

In summary, Hypothesis 1 is supported by the data, though the difference in
' shame between gay and heterosexual men is smaller when comparing HIV-negative gay
men to heterosexual men, and larger when comparing HIV-positive gay men to

heterosexual men.

86

Hypothesis 2: The gay sample is expected to manifest more depressive
symptomatology than the heterosexual sample.

   

 

 

 

 

 

 

 

 

 

 

 

 

Table 16
.,, in en.
Group 90% 2-slded
Gay Hetero. Confidence Interval Odds Sig.
Variable (u - 107) (N - 112) pm.“ SE0... Lower Upper d” IP‘ Ratio‘ (p < .05)
M 18.76 16.76 2.00 1.63 -.69 4.70 .17 -89 8.09 ns
Depression
SD 13.35 10.55
Group 90% 2-slded
HlV- Gay Hetero. Confidence Interval Odds Sig.
Variable (N I 57) (N I 112) Diff. SEW Lower Upper d IP Ratio (p < .05)
M 14.81 16.76 -1.95 1.75 -4.84 .93 -.18 .13 .15 ns
Depression
SD 1 1.07 10.55
Group 90% Z-slded
HIV+ Gay Hetero. Confidence Interval Odds Sig.
Variable (N I 50) (N I 112) Diff. 3501!! Lower Upper (1 IP Ratio (p < .05)
, M 23.27 16.76 6.51 2.26 2.74 10.28 .52 .99 99.00 sig
Depression

80 14.37 10.55

 

'Value of the difference in means.
'Effect size; the magnitude of the difference in means expressed in standard deviation units.

¢Inference Probability; the probability that the true population difference is positive.
‘Ratio of the inference probability (IP) to the reverse probability (MP); the odds that the directional hypothesis is true.

Gay men as a group reported slightly more depression ((1 = .17) than heterosexual
men, as shown in Table 16. This difference did not reach statistical significance, but the
odds are better than 8:1 (89% probability) that the true population difference is non-zero.

When HIV serostatus was taken into account, however, a different picture
emerged. As Table 16 shows, HIV-negative gay men actually reported slightly less
depression than heterosexual rnen (d = -.18, ns). HIV-positive men, on the other hand,
reported significantly more depression than heterosexual men (d = .52, p < .05; IP = .99).

Thus, Hypothesis 2 is supported, but for HIV-positive gay men only.

87

Hypothesis 3: Within the gay sample, HIV-positive serostatus will be associated with
greater internalized shame.

 

 

 

 

Table 17
T53 gf bivariatg hyMMis 3: HIV+ ggy men will evince greater internalized shame than HIV- ggy
HIV Serostatus 90% 2-sided
Positive Negative Confidence Interval Odds Sig.
Variable (N I 50) (N I 57) Diff.‘ 3501:: Lower Upper d” IP‘ Ratio“ (p < .05)
M 38.73 32.36 6.37 4.00 -.27 13.01 .31 .94 15.67 ns
Shame .

SD 23.30 17.99

 

'Absolute value of the difference in means.

'Efiect size; the magnitude of the difference in means expressed in standard deviation units.

‘lnference Probability. the probability that the true population difference is positive.

‘Ratio of the inference probability (IP) to the reverse probability (1-lP); the odds that the directional hypothesis is true.

Table 17 shows the difference between HIV-positive and HIV-negative gay men
with respect to internalized shame. The HIV-positive men reported more shame (d = .31)
.than the HIV -negative men. This difference did not reach statistical significance, but the
odds are nearly 16:1 (94% probability) that the true population difference is non-zero.
Thus, the data support Hypothesis 3.

Tables 18-25 present the results of bivariate correlational tests of hypotheses 4
through 11. In these tables, correlations are displayed separately for heterosexual men (N
= 112) and gay men (N = 107), for HIV-negative (N = 57) and HIV -positive (N = 50) gay
men, and for HIV-positive gay men with (N = 31) and without (N = 17) an AIDS
diagnosis. By presenting the results in this way, the applicability of a given hypothesis to

.each subgroup can be evaluated.

88

Hypothesis 4: More physical symptoms will be associated with greater internalized
shame.

 

 

 

Table 18
. 0 01:1: : t rm 11% ~ ’ 't r -1 m °1°ll~ . a: un-thl - 1:14:1sz l :u: "1' »t.::tt:.
90% 2-slded
Confidence Interval' Odds Sig.
Group E SE. Lower Upper lP’ Ratio" N (p < .05)
Heterosexual Men .18 .09 .02 .35 .98 49.00 _1 12 sig
Gay Men (all) .28 .09 .12 .45 .99 99.00 107 - sig
HIV- .18 .13 -.05 .41 .92 ‘ 11.50 57 ns
HIV+ .27 .13 .03 .51 .98 49.00 50 sig
AIDS Dx .47 .14 .22 .73 ' .99 99.00 31 sig
No AIDS OX .25 .23 -.18 .67 .85 5.67 17 ns

 

'Corrected for attenuation.
'lnferenoe Probability, the probability that the true population correlation is positive.
‘Ratio of the inference probability (IP) to the reverse probability (MP); the odds that the directional hypothesis is

Table 18 presents the correlation between physical symptoms and internalized
shame for each of the study subgroups. As the table shows, there was a modest but
significant correlation (r é .18, p < .05) between these variables for heterosexual men. For
gay men, the effect was stronger (r = .28, p < .05); however, when the gay sample was
broken down by HIV serostatus it can be seen that the relationship between symptoms
and shame for HIV-negative gay men was the same as that for heterosexual men (r = .18).
This correlation did not reach statistical significance for the HIV -negative men due to the
smaller size of this subgroup (N = 57), but the odds are nearly 12:1 (92% probability) that
the true population. correlation is positive.

For HIV -positive men, the correlation between physical symptoms and shame was
.27 (p < .05). This effect was much stronger, however, when AIDS diagnosis was taken
into account. For men with AIDS, the correlation between symptoms and shame
increased to .47 (p < .05), while for HIV-positive men without AIDS the effect was

similar to that for HIV-positive men in general (r = .25). Again, due to the small number

89

of HIV -positive men without AIDS (-N = 17), the latter correlation did not reach statistical
significance, but the odds are nearly 6:] (85% probability) that the true population
correlation is positive.

In general, the data support Hypothesis 4, particularly among HIV-positive gay

men, and especially among gay men with AIDS.

90

Hypothesis 5: More physical symptoms will be associated with greater overall stress
level.

Table 19

at 91 -

   

 

 

 

90% 2-sided

Confidence Interval’ Odds Sig.

Group R SE, Lower Upper If!" Ratioc N (p < .05)
Heterosexual Men .29 .09 .11 .47 .99 99.00 ‘ 112 sig
Gay Men (all) .55 .07 .41 .69 .99 99.00' 107 sig
HIV- .40 .1 1 .17 .63 .99 99.00 57 sig
HIV+ .53 .10 .32 .74 .99 99.00 50 sig
AIDS Dx .69 .09 .50 .89 .99 99.00 31 sig
No AIDS OX .25 .23 -.23 .74 .86 6.14 17 ns

 

'Corrected for attenuation.
bInference Probability, the probability that the true population correlation is positive.
‘Ratio of the inference probability (IP) to the reverse probability (1-IP); the odds that the directional hypothesis is

Table 19 presents the correlation between physical symptoms and stress for each
of the study subgroups. For heterosexual men, there was a moderate significant
correlation between these variables (r = .29, p < .05). For gay men, the effect was more
pronounced (r = .55, p < .05), especially for HIV-positive men (r = .53, p < .05) as
compared to HIV-negative men (r = .40, p < .05).

When the HIV -positive group was broken down by AIDS diagnosis, the observed
correlation between symptoms and stress was very strong for the men with AIDS (r = .69,
p < .05), not surprisingly. However, for HIV -positive men without AIDS, the correlation
between symptoms and stress was relatively modest (r = .25, ns), and in fact was more
similar in magnitude to the correlation observed for heterosexual men than to those
observed for either HIV-negative gay men or HIV-positive gay men with AIDS. It may be
the case that, for HIV-negative gay men, physical symptoms cause increased stress due to
fears about becoming HIV-infected. For men with AIDS, on the other hand, symptoms

may cause increased stress due to fears of impending death. But for HIV positive men

91

without an AIDS diagnosis, the weaker correlation between symptoms and stress may be
due to the fact that they are no longer living in fear of HIV infection, nor are they sick
enough to fear imminent death. In any case, the data provide strong support for

Hypothesis 5, especially among HIV-negative gay men and gay men with AIDS.

92

Hypothesis 6: Greater perceived control will be associated with increased use of
involvement coping.
Table 20

t

   

 

 

 

90% 241606
6 Confidence Interval‘ Odds Sig.
Group 1’ SE. Lower Upper IP" Ratio‘ N (p < .05)

Heterosexual Men .01 .09 -.20 .22 .53 1.13 . 112 ns
Gay Men (all) -.30 .09 -.49 -.10 .00 .00 107 sig“
HIV- . -.39 .11 -.64 -.14 .oo .oo 57 sig“
HIV+ -.05 .14 -.37 .26 .36 .56 50 ns
AIDS DX .35 .16 -.01 .70 .99 99.00 31 ns

NO AIDS OX -.55 .17 -.94 -.17 .00 .00 17 sig‘1

 

'Corrected for attenuation.

'lnrerence Probability. the probability that the true population correlation is positive.

‘Ratio of the inference probability (IP) to the reverse probability (1-IP); the odds that the directional hypothesis is
‘Relationship is in opposite direction of what was predicted.

Table 20 presents the correlation between appraisals of control and use of
involvement coping for each of the study subgroups. For heterosexual men, these
variables were unrelated (r = .01, ns). For gay men, however, the results were mixed. For
HIV-negative gay men, contrary to expectations, feeling in control of stressful situations
was negatively correlated with involvement coping (r = -.39, p < .05). For HIV -positive
gay men without AIDS, this negative relationship was even stronger (r = -.56, p < .05).
However, for HIV-positive gay men with AIDS, the effect was in the expected direction
(r = .35). This correlation fell short of statistical significance, but there is about a 99%
probability (99:1 odds) that the true population correlation is positive.

These mixed findings are difficult to explain. It may be the case that for HIV -
negative gay men, as well as HIV-positive gay men without AIDS, feeling more in
control lessens the need for coping behaviors of any kind (involvement or detachment),
especially since control is negatively correlated with stress. However, for the men with

AIDS, who reported the most stress (mean stress quotient = .50) and least control (mean

93

control quotient = .46) of any subgroup, coping in an involved and active way may be
crucial to maintaining their sense of control. Strictly'speaking, this interpretation deviates
from Lazarus and Folkman’s (1984) conceptualization, which holds that appraisals of
control precede coping responses. However, it seems reasonable to suggest that appraisals
of control and coping responses have a reciprocal effect on one another. That is, feeling in
control may Spur one to engage in positive, involvement-oriented Coping behaviors,
which in turn may reinforce one’s sense of control.

In summary, the data support Hypothesis 6 only among gay men with AIDS;
while the observed positive correlation for this subgroup was not statistically significant,

the odds are 99:1 that the true population correlation is positive (IP = .99).

94

Hypothesis 7: Lower perceived control will be associated with increased use of
detachment coping.

 

 

 

 

Table 21
-'~ - I1I-31Px' ¢-~ ' 011-1 "In: lit—- 1 1° ll-ix' hoax-At in ”I-
90% Z-Sldod
Confidence Interval' Odds Sig.
Group r’ SE. Lower Upper IP" Ratioc N (p < .05)
Heterosexual Men -.27 .09 -.48 -.07 .99 99.00 _112 sig
Gay Men (all) -.52 .07 -.69 -.36 .99 99.00 107 . Sig
HIV- -.39 .11 -.66 -.12 .99 99.00 57 sig
HIV+ -.52 .10 -.76 -.27 99 99.00 50 sig
AIDS DX -.17 .18 -.59 .25 ‘ .84 5.25 31 ns
NO AIDS OX -.82 .08 -.99 -.62 99 99.00 17 sig
'Corrected for attenuation.

”Inference Probability; the probability that the true population correlation is negative.
‘Ratio of the inference probability (IP) to the reverse probability (MP); the odds that the directional hypothesis is

Table 21 presents the correlation between appraisals of control and use of
detachment coping for each of the study subgroups. As the table shows, heterosexual men
with lower appraisals of control were more likely to use detachment coping strategies (r =
-.27, p < .05), as predicted. This effect is nearly twice as strong for gay men (r = -.52, p <
.05). Much of this effect is attributable to HIV -positive men (r = -.52, p < .05), and is
particularly strong for HIV-positive men without an AIDS diagnosis (r = -.82, p < .05).
Interestingly, the relationship between control and detachment is much weaker for men
diagnosed with AIDS (r = -.17, us) than for any other subgroup, particularly when
compared to HIV -positive men without AIDS. This difference cannot be explained by
lesser use of detachment coping among men with AIDS, since men with and without
AIDS reported similar levels of use of detachment coping (means = 27.29 and 26.42,
respectively). Perhaps for men diagnosed with AIDS, a relatively “stable” (i.e., less
reactive to changes in such factors as appraisals of control) level of detachment is

maintained in order to stave off intolerable levels of worry and anxiety.

95

In summary, Hypothesis 7 is supported by the data. The effect was strongest
among HIV-positive men without AIDS, and weakest among those with AIDS, for whom

detachment coping was less tied to changes in their appraisals of control.

96

Hypothesis 8: Lower perceived control will be associated with greater internalized

shame.

 

 

 

 

 

90% Zilded
Confidence Interval' Odds Sig.
Group r’ SE, Lower Upper IP" Ratio‘ N (p < .05)
Heterosexual Men -.37 .08 -.53 -.21 .99 99.00 .112 sig
Gay Men (all) -.57 .07 -.70 -.44 .99 99.00 107 . sig
HIV- -.36 .12 -.58 -.13 .99 ‘ 99.00 57 sig
HIV+ -.69 .07 -.83 -.55 99 99.00 50 Sig
AIDS Dx -.69 .10 -.87 -.51 .99 99.00 31 sig
No AIDS OX -.69 .'13 -.94 -.44 99 99.00 17 sig
'Corrected for attenuation.

'lnterenoe Probability. the probability that the true population correlation is negative.
‘Ratio of the inference probability (IP) to the reverse probability (t-lP); the odds that the directional hypothesis is

Table 22 presents the correlation between appraisals of control and internalized

shame for each of the study subgroups. As predicted, decreased control was associated

with increased shame for all of the subgroups. This effect was particularly pronounced

among HIV-positive gay men, regardless of whether or not they were diagnosed with

AIDS (r = -.69, p < .05). Hypothesis 8 is supported by the data.

97

Hypothesis 9: Greater use of involvement coping will be negatively associated with
depression.

 

 

 

 

90% 2-dlded
Confidence Interval' Odds Sig.

Group 1’ 3E. Lower Upper IP" , Ratlo‘ N (p < .05)
Heterosexual Men .10 .09 -.08 .29 .13 .15 112 ns
Gay Men (all) .15 .09 -.04 .34 .06 .06 p 107 . ns
HIV- .14 .13 -.12 .40 .14 .16 57 ns
HIV+ -.07 .14 -.35 .21 .68 2.13 50 ns
AIDS OX -.21 .17 -.56 .13 .89 8.09 31 ns
No AIDS Dx .18 .24 -.30 .66 .23 .30 17 ns

 

‘Corrected for attenuation.
'Inference Probability; the probability that the true population correlation is negative.
“Ratio of the inference probability (IP) to the reverse probability (1-IP); the odds that the directional hypothesis is true.

Table 23 presents the correlation between involvement coping and depression for
each of the study subgroups. In general, the data do not support Hypothesis 9, although a
modest negative correlation was observed for men with AIDS (r = -.21); this effect failed
to reach statistical significance, but the odds are better than 8:1 that the true population
correlation is negative (IP = .89). Similarly to what was suggested above with respect to
the relationship between involvement coping and control (Hypothesis 6), it may be the
case that men living with AIDS are in particular need of continual involvement and active

coping in order to stave off depression.

98

Hypothesis 10: Greater use of detachment coping will be positively associated with
depression.

 

 

 

 

90% 2-sided

Confidence Interval' Odds Sig.

Group 1’ SE. Lower Upper IP" Ratio‘ N (p < .05)
Heterosexual Men .55 .07 .41 .69 .99 99.00 ' 112 sig
Gay Men (all) .78 .04 .70 .86 .99 99.00 107 sig
HIV- .67 .07 .51 .82 .99 99.00 57 sig
HIV+ .79 .05 .67 .90 .99 99.00 50 sig
AIDS Dx .65 .11 .42 .87 .99 99.00 31 sig
No AIDS Dx .93 .03 .86 .99 .99 99.00 17 sig

 

'Corrected for attenuation.
'lnterence Probability. the probability that the true population correlation is positive.
‘Ratio of the inference probability (IP) to the reverse probability (1-IP); the odds that the directional hypothesis is

Table 24 presents the correlation between detachment coping and depression for
each of the study subgroups. As the table shows, Hypothesis 10 is strongly supported by
the data. The effect was stronger for gay men (r = .78) than for heterosexual men ‘(r =
.55), and among gay men was stronger for HIV-positive men (r = .79) than for HIV-
negative (r = .67) men. Within the HIV -positive group, the effect was much stronger for
men without AIDS (r = .93) than for men with AIDS (r = .65). As suggested above in the
discussion of Hypothesis 7, perhaps detachment plays a different, less malignant role for
men with AIDS than for HIV-positive men without an AIDS diagnosis. In any case,

detachment is strongly related to depression for all of the study subgroups.

99

Hypothesis 11: Greater internalized shame will be positively associated with
depression.

Table 25

.:1 O) O

   

 

 

 

90% 2-eided

Confidence Interval' , Odds Sig.

Group r’ SE. Lower Upper lP" Ratio‘ N (p < .05)
Heterosexual Men .83 .03 .78 .' .88 .99 99.00 112 sig
Gay Men (all) .87 .02 .83 .91 ' .99. 99.00 107 sig
HIV- .82 .04 .75 .90 .99 99.00 57 sig
HIV+ .90 .03 .86 .95 .99 99.00 50 sig
AIDS OX .82 .06 .72 .93 .99 99.00 31 sig
No AIDS Dx .96 .02 .92 .99 .99 99.00 17 sig

 

'Corrected for attenuation.
‘Inference Probability; the probability that the true population correlation is positive.
‘Ratio of the inference probability (IP) to the reverse probability (MP); the odds that the directional hypothesis is

Table 25 presents the correlation between internalized shame and depression for
each of the study subgroups. As the table indicates, shame was strongly positively
associated with depressed mood for each of the subgroups. This effect was particularly
pronounced for HIV-positive men without AIDS (r = .96).

Table 26 summarizes the results of the tests of the 11 bivariate hypotheses. A
hypothesis is considered to have been supported by the data if:

a) the result was statistically significant and/or

b) the odds of the true population parameter lying in the predicted direction were

estimated at 2:1 or greater (i.e., IP 2 .68).

100

 

 

 

 

 

 

 

 

 

 

 

 

Table 26
Summary of gsulg of bivarigtg mg.
H o- '
’9 Prediction Result Table
thesis .
1 The gay sample will eVlnce greater lntemallzed shame than the heterosexual Su I I 15
sample.
2 The gay sample will manifest more depressive symptomatology than the Supported for HIV+ gay 16
heterosexual sample. men only.
Within the gay sample. HIV-positive serostatus will be associated with greater
3 internalized shame. Supported ‘7
4 More physical symptoms will be associated with greater internalized shame. Supported 18
5 More physical symptoms will be associated with greater overall stress level. Supported 19
6 Greater perceived control will be associated with increased use of involvement Supported for gay men 20
coping. Mth AIDS only.
7 Lower perceived control will be assoaated with increased use of detachment Supported 21
coping.
8 Lower perceived control will be associated with greater intemalized shame. Supported 22
9 Greater use of involvement coping will be negatively associated with Supported for gay men 23
depression. with AIDS only.
10 Greater use of detachment coping will be positively associated with depression. Supported 24
1 1 Greater internalized shame will be positively associated with depression. Supported 25

 

 

 

 

 

 

101

MULTIVARIATE ANALYSES

The path models described in Chapter 1 were tested using Hunter and Hamilton’s
(1995) PA TH software, a least squares path analysis program. Results of these analyses
are shown in Figures 3 and 4. These figures depict statistically significant paths with solid
bold arrows. Paths that did not reach statistical significance but had odds ratios greater
than or equal to 2:1 (corresponding to an IP of .68 or greater) are depicted with bold two-
tone arrows. Figure 4 denotes a “missing link” with a bold dotted arrow.9 Non-significant
paths are shown with thin arrows.-

Results of the chi square test for goodness-of-fit, as well as the multiple R2 for
. depression, are given. The chi square test indicates deviation from goodness-of-fit;
therefore, non-significant values of chi square are desirable. 1R2 quantifies the proportion
of the variance in the dependent'variable (depression) accounted for by its antecedent
variables (involvement and detachment coping).

Finally, associated with each figure is a table showing the results for individual
links in the model. Because each path represents a directional bivariate hypothesis, a 90%
2-sided confidence interval is built around each path coefficient, and the inference

probability and odds ratio are calculated for each path coefficient.

 

9 The Hunter and Hamilton PA TH program provides a “missing link” analysis that evaluates whether any
paths not specified in the model would be significant if included in the model.

102

Figure 3 shows the results of Path Model 1, tested for gay men (N = 107). Income
and education were controlled for since these variables were found to be correlated with a
number of the variables in the path analysis (see Table 10). Table 27 shows analyses for

the individual links in the model.

Flgure3
l fP I amle'N=1 .

 

 

Overall Test fl Model:

Chi Square (11 (If) - 15.85
Statistically Significant Path p I .147
NS Path with Odds Ratio >- 2:1 (IP >- .68) R’ (Depression) - .76

 

> NS Path p < .001

As Figure 3 shows, Path Model 1 fits the data well overall (chi square 11 df =
15.85, p = .147). About 76% of the variance in depression is accounted for by

involvement and detachment coping among gay men in this sample (R2 = .76, p < .001).

103

 

 

 

Table 27 .
v i r 1 ° = 1
90% 2-sided

Predicted Path Confidence Interval Odds Sio-
Path Direction Coeff. SE Lower Upper (pb Ratio“ l2 < .05)
HIV-v Symptoms + .70 .06 .60 .80 .99 99.00 sig
HIV—b Shame + .09 .15 -.16 .34 .73 2.70 ns
HIV-b Stress + -.34 .18 -.64 -.04 .03 .03 sig“
Symmoms —> Shame + -.01 .16 -.27 .25. .48 .92 ns
Symptoms ->Stress + .69 .18 .39 .99 .99 99.00 sig
Symptoms —> Control - -.08 .17 -.36 .20 .68 2.13 ns

' Shame +Stress + .62 .09 .47 .77 .99 99.00 ' sig

Shame —p Control - -.20 .18 .50 .10 .87 6.69 ns
Shame —> Involvement - -.30 .18 -.60 .00 95 19.00 ns
Shame —> Detachment + .51 .14 .28 .74 99 99.00 sig
Stress -> Control - -.51 .22 -.87 -.15 .99 99.00 sig
Stress —9 Involvement - I .53 .25 .12 .94 .02 .02 sigd
Stress -§ Detachment + .49 .22 .13 .85 99 99.00 sig
Control -§ Involvement + -.03 .22 —.39 .33 44 .79 ns
Control -> Detachment - .13 .23 -.25 .51 28 .39 ns
Involvement ->' Depression - -.49 .23 -.87 -.11 .78 3.55 sig
Detachment -> Depression + 1.04 .20 .71 1.37 .83 4.88 sig

 

'Controlling for income 8. education.
°lnference Probability; the probability that the true population value lies in the predicted direction.
‘Ratio of the inference probability (IP) to the reverse probability (1-IP); the odds that the directional hypothesis is true.

“Relationship is in opposite direction of what was predicted.

Looking next at the results for individual paths, Table 27 shows that HIV -positive
serostatus was associated with increased physical symptoms, as expected (b = .70, p <
.05). However, contrary to expectations, being HIV-positive was negatively correlated
with stress (b = -.34, p < .05). The predicted positive relationship between HIV-positive
serostatus and shame was weak and did not reach statistical significance (b = .09, p <
.05), though the odds that the true population parameter is positive are nearly 3:1 (IP =
.73).

Physical symptoms were positively correlated with stress, as predicted (b = .69, p
< .05). However, the paths between symptoms and shame and between symptoms and

control were non-significant (b’s = -.01 and -.08, respectively), though the odds are about

104

2:1 (IP = .68) that the true population parameter for the relationship between symptoms
and control is negative in direction. '

Shame was significantly positively related to stress (b = .62, p < .05) and to
detachment coping (b = .51, p < .05). The path from shame to control (b = -.20) was not
statistically significant, but the odds are nearly 7:1 (IP = .87) that the true population ’
parameter is negative, as predicted. Likewise, the path between shame and involvement
coping (b = -.30) fell short of statistical significance, but the odds are 19:1 (IP = .95) that
the true population parameter is negative, as predicted.

Stress was negatively related to control (b = -.51, p < .05), and positively related
to both involvement (b = .53, p < .05) and detachment (b = .49, p < .05) coping. This
finding is somewhat contrary to expectations, as it was predicted that stress would be
positively related to detachment coping but negatively related to involvement coping.

The paths from control to involvement (b = -.03) and detachment (b = .13) coping
were non-significant.

The paths from involvement and detachment coping to depression were
significant, with involvement coping negatively related (b = -.49, p < .05) and detachment
coping positively related to depression (b = 1.04, p < .05), as predicted.

Finally, no “missing links” were indicated for direct paths between stress and
depression or between control and depression. In other words, the data do not indicate
that stress or control have any direct effect upon depression, but instead exert their effects
through the other variables in the model, as predicted. This was true for all four of the

path models tested.

105

Overall, these results provide strong support for Path Model 1. The main
exceptions were the failure of control to predict either type of coping, the positive
relationship between stress and involvement coping, and the negative relationship
between HIV-positive serostatus and stress.

The failure of appraisals of control to predict either type of coping is not entirely
unprecedented; both Folkman et al. (1993) and DeMarco et al. (1997) found that control
failed to predict detachment coping, and was only modestly related to involvement
‘ coping, among gay men. It was hoped that the more global measure of coping used in the
present study (i.e., measuring coping across all life domains rather than only the most
stressful domain) would improve its predictive value, but this was not the case.

Also unexpected was the positive relationship between stress and involvement
coping. Both Folkman et al. (1993) and DeMarco et al. (1997) found that increased stress
was related to increased use of detachment coping, but that stress and involvement coping
were unrelated. The present data indicate that increases in stress mobilized both types of
coping responses among gay men in this sample. This deviation from past findings may
be due to the fact that the present investigation employed the full length version of the
WOC questionnaire, while the earlier investigations cited above utilized shortened
versions of that instrument.

Finally, the significant negative relationship between HIV serostatus and stress is
counterintuitive. Among gay men in this sample (N = 107), these two variables have a
positive bivariate correlation of .25 (p = .004). The multivariate findings, however,
indicate that once other variables in the model (such as physical symptoms and shame)

are accounted for, being HIV -positive is actually related to decreased stress. These

106

findings suggest that emotionally well-adj usted HIV -positive gay men in good physical
health may actually experience less stress than their HIV -negative counterparts. One can
only speculate as to why this might be the case; perhaps HIV -positive men have been
forced to mobilize coping resources such as social support, which serves to reduce stress
as long as they remain healthy and emotionally well-adjusted. Or perhaps HIV-positive
men are able to abdicate certain social roles and obligations, thus reducing potential
sources of stress. This paradoxical and interesting finding warrants future investigation,

as it may prove fruitful in suggesting ways for HIV -positive men to reduce stress.

107

Figure 4 shows the results of Path Model 2, tested for heterosexual men (N =
112). Income was controlled for since it was found to be correlated with a number of the
variables in the path analysis (see Table 11). Table 28 shows analyses for the individual

links in the model.

Figure4
Re f? h I2h er xalsmI'N=11.

 

 

o ngrall Test of Model:
Statistically Significant Path - Chi Square (7 df) I 20.20
NS Path with Odds Ratio >= 2:1 (IP >I .68) p I .005

------- D ”Missing Link” R1 (“MIG") ' -3“
#- NS Path P < -01

 

As Figure 4 shows, Path Model 2 does not fit the data particularly well (chi square
7 df = 20.2, p = .005). About 31% of the variance in depression among the heterosexual
men is accounted for by the antecedent variables involvement and detachment coping (R2
' = .31, p < .001), considerably less than was the caseifor gay men. As the dotted arrow
indicates, there was a “missing link” between shame and depression; that is, the data

indicate that there was a significant positive direct effect of shame on depression for these

men that was not specified in the path model. This direct effect would have been

108

significant if included in the model, even after accounting for the effects of the other

 

 

variables.
Table 28
In Ivi Ilin l frPthM I h r. ml'N=112'.
90% 2-sided

Predicted Path anfidence Integal Oddli 3'9-
Path Direction Coeff. SE Lower Upper IE” gang (2 < .05)
Symptoms-b Shame + .15 .10 -.01 .31 .93 13.29 ns
Symptoms-P Stress + .27 .11 .09 .45' .99 99.00 sig
Symptoms-1b Control - -.04 .13 -.25 .17 .62 1.63 ns
Shame-b Stress + .25 .11 .07 .43 .99 99.00 sig
Shame-e Control - -.20 .11 -.38 -.02 .97 32.33 sig
Shame» Involvement - -.02 .13 -.23 .19 .56 1.27 ns
Shame—e Detachment + .55 .11 .37 .73 .99 99.00 sig
Stress—P Control - -.52 .12 -.72 -.32 .99 99.00 sig
Stress-9 Involvement - .40 .18 .10 .70 .01 .01 sig“
Stress-P Detachment + .48 .18 .18 .78 .99 99.00 sig
Control-P Involvement + .25 .20 -.08 .58 .89 8.09 ns
Control-b Detachment - .22 .21 -.12 .56 .15 .18 ns
Involvement? Depression - -.12 .14 -.35 .11 .81 4.26 ns
Detachment? Depression + .61 .12 .41 .81 .99 99.00 sig

 

‘Controlling for income.

l’Inference Probability; the probability that the true population value lies in the predicted direction.

‘Ratio of the inference probability.(IP) to the reverse probability (MP); the odds that the directional hypothesis is true.
‘Relationship is in opposite direction of what was predicted.

Table 28 allows examination of the results of the individual link analyses for Path
Model 2. For the heterosexual men, physical symptoms were positively related to
increased stress as predicted (b = .27, p < .05), though this effect was not as strong as that
.observed for the gay men (see Table 27). Physical symptoms were also weakly related to
increased shame among the heterosexual men (b = .15, ns; IP = .93), which was not the
case for the gay men once HIV status was accounted for in the model. No relationship
was observed between physical symptoms and control for heterosexual men, unlike for
the gay men, who showed a weakly negative relationship between symptoms and control.

Shame was related to increased stress (b = .25, p < .05) and decreased feelings of

control (b = .20, p < .05) among heterosexual men, as it was for gay men, though the

109

latter reported a stronger relationship between shame and stress. For the heterosexual
men, as for the gay men, shame was related to increased use of detachment coping (b =
.55, p < .05); unlike for the gay men, however, shame was not related to decreased use of
involvement c0ping among the heterosexual men (b = -.02, ns; IP = .56).

Stress was positively related to both involvement (b = .40, p < .05) and
detachment (b = .48, p < .05) coping, and negatively associated with control (b = -.52, p <
i .05). The direction and magnitude of these effects were quite similar to what was found
for gay men. Once again, contrary to expectations, stress was positively correlated with
both kinds of coping, rather than being negatively related to involvement and positively
related to detachment as predicted.

Control was modestly related to increased use of involvement coping among the
heterosexual men (b = .25, ns; IP = .89), unlike for the gay men. However, control was
unrelated to detachment coping (b = .22, ns), which was the case for the gay men as well.

Finally, for the heterosexual men, involvement coping was negatively associated
with depression (b = -.12, ns; IP = .81), though this effect was weak compared to that
observed for the gay men. Detachment coping was positively related to depression (b =
.61, pi< .05) among the heterosexual men, though once again this effect was not as
pronounced as that observed for gay men.

In general, Model 2 does not fit the data for heterosexual men as well as Model 1
fits the data for gay men. In one important respect the two models are not directly.
comparable, since HIV serostatus was unknown among the heterosexual men. In addition,
the theoretical model tested was derived primarily from literature pertaining to HIV -

positive and -negative gay men. Nevertheless, there are some important similarities in

110

results between the two samples. For both samples, physical symptoms were related to
increased stress, which in turn was negatively related to feeling in control and positively
related to increased use of both involvement and detachment coping strategies. For both
heterosexual and gay men, involvement coping was associated with decreased depression,
and detachment coping with increased depression, with the latter effect being particularly
strong. Finally, for both samples, increased shame was related to higher stress, lower
control, and increased use of detachment coping, which as stated above leads to increased
depression.

Thus, for both gay and heterosexual men in this study, physical health has an
important impact on stress level, which in turn reduces feelings of control and leads to the
mobilization of various coping strategies. Some of these coping strategies are adaptive
with respect to emotional well-being, while others are detrimental. In general, coping
behaviors that involve active confi'ontation of problems and the utilization of social
support lead to better emotional adjustment, while those that involve detachment or
denial are associated with poorer adjustment. The results for both samples also suggest
that increased feelings of shame may exacerbate adjustment problems by increasing
stress, decreasing feelings of control, and increasing the use of detrimental detachment

coping strategies.

SUPPLEMENTAL ANALYSIS

One of the clearest findings emerging from the preceding multivariate analyses
was that greater use of detachment coping was associated with increased depression,

While greater use of involvement coping was related to decreased depression. Less clear,

111

however, were factors affecting the specific type of coping a person might tend to use.
Degree of perceived control, for example, was a poor predictor of coping among gay men,
and for heterosexual men showed only a modest positive correlation with involvement
coping. Level of internalized shame, on the other hand, was a fairly good predictor of
involvement versus detachment coping among gay men, but was associated only with
detachment coping among heterosexual men. Finally, increased stress was associated with
greater use of both types of coping, and therefore did not improve our ability to
differentiate which type of coping a person might tend to use.

To address this problem, a single measure of “coping preference” was created to
assess whether any factors were clearly associated with greater use of one type of coping
relative to the other. For each participant, a new variable was computed by subtracting
the involvement coping score from the detachment coping score. The resulting variable
therefore had a positive sign if the person’s detachment score was greater than his
involvement score, and a negative sign if the person preferred involvement coping over
detachment. In other words, the larger the value of the new variable, the greater was the
person’s preference for detachment coping relative to involvement coping.

Bivariate correlations were then calculated between the coping preference variable
and a variety of demographic and substantive variables. The results are shown in Table
29. Correlations are presented separately for heterosexual men, gay men, HIV-positive
gay men, and gay men with AIDS. Note that because of the way the coping preference
variable was calculated (detachment score minus involvement score), positive

correlations in Table 29 indicate factors related to greater use of detachment coping

112

relative to involvement coping, while negative correlations indicate factors related to

greater use of involvement coping relative to detachment coping.

Table 29

113

Correlg_tions between detachment coping preference " and

other variables in the study.

 

 

Group Variable r p
Heterosexual Age .20 .037
Men Caucasian Race .10 .304
(N = 112) Education .03 .778
Annual Income -.10 .321
Physical Symptoms .12 .209
Stress Quotient .11 .268
Control Quotient -.15 .106
Shame .36 .000
Depression .26 .006
Gay Men HIV Positive -.10 .290
(N =107) AIDS Diagnosis -.03 .742
Age -.09 .382
Caucasian Race .09 .369
Education .03 .780
Annual Income .01 .890
Physical Symptoms -.01 .906
Stress Quotient .13 .176
Control Quotient -.09 .367
Shame .41 .000
Depression .38 .000
HIV+ AIDS Diagnosis .01 .917
Gay Men Age -.06 .699
(N = 50) Caucasian Race .29 .043
' Education .04 .794
Annual Income -.07 .651
Physical Symptoms .11 .439
Stress Quotient .27 .063
Control Quotient -.26 .074
Shame ‘ .60 .000
Depression .54 .000
Gay Men Age .17 .375
with AIDS Caucasian Race .15 .407
(N = 31) Education .01 .963
Annual Income -.15 .409
Physical Symptoms .24 .185
Stress Quotient .50 .004
Control Quotient -.35 .055
Shame .57 .001
Depression .57 .001

 

‘Detachment score minus involvement score. Larger values Indicate
preference for detachment coping.

114

Table 29 indicates that, for heterosexual men, age, shame, and depression were
positively correlated with preference for detachment coping. The only surprising finding
here is the correlation between age and detachment, particularly given the small variance
in age within the heterosexual sample. It is unclear why slightly older college students
would indicate a significant preference for detachment coping behaviors.

Among gay men as a group, only shame and depression were significantly related
to coping preference; both of these variable were positively associated with a preference
for detachment coping. When considering only HIV-positive gay men, a slightly different
picture emerged. For these men, Caucasian ethnicity, as well as shame and depression,
were significantly correlated with detachment coping. Further, the correlations between
shame, depression, and detachment were larger for these men compared to gay men in
general. An unexpected finding here is that Caucasian men were more likely to prefer
detachment coping than ethnic minority men. It is unclear why white men would be more
likely than minority men to cope using detachmentvoriented strategies. Unfortunately, the
number of minority respondents in this sample is too small to permit meaningful analysis

.of the differences between white and non-white respondents.

Finally, for men with AIDS, stress, shame, and depression were all positively
correlated with a preference for detachment coping. In addition, the negative correlation
between control and detachment fell just short of statistical significance for these men (r
= -.35,p = .055).

These findings lend further support to the overall model proposed in this study,

particularly with respect to HIV-positive men and men with AIDS. For these at-risk men,

115

minimizing stress, increasing perceived control over stressful situations, and addressing
shame issues may be critical to maintaining healthy, involvement-oriented coping
behaviors, which in turn may result in better emotional adjustment and decreased

depression.

Chapter 4

DISCUSSION

METHODOLOGICAL CONSIDERATIONS

A number of methodological issues must be taken into account when interpreting
the results of this research. These are outlined in the following sections.

Sampling Issues

The use of non-random sampling techniques in this study may limit the
generalizability of the results. The heterosexual sample was a sample of convenience
recruited through the Psychology Department subject pool, and was homogenous‘with
respect to age, ethnicity, and educational attainment. This sample differed considerably
from the gay sample with respect to age, educational attainment, and annual income, and
therefore comparisons made on the basis of sexual orientation should be viewed with
caution since the two samples differedalong so many other dimensions as well. The HIV -
positive sample was a sample of convenience recruited through a local HIV/AIDS service
organization (LAAN), and was probably skewed toward more economically
disadvantaged men with more advanced HIV disease. Further, the response rate of 63%

- for LAAN clients, while adequate, nevertheless begs the question of whether there were
important differences between the men who decided to fill out the survey and those who
did not (or could not) do so. HIV-negative gay men volunteered for the study after

reading an advertisement or hearing about the proj ect from a friend or acquaintance.

116

117

Again, it is impossible to know whether men who volunteered to participate in this
research differed in important ways fi'om men not volunteering. Finally, both the gay and
heterosexual samples were ethnically homogenous and (by design) restricted to men;
thus, the results reported here should not be generalized to women or ethnic minorities.

Lack of HIV Information for the Heterosexual Sample

Because of the low base rate of HIV testing among heterosexual male college
students, it was not possible to recruit a sample of heterosexual men with known HIV
. serostatus using the Psychology Department subject pool. This precluded strict
comparisons between the gay and heterosexual samples, as well as the assessment of HIV
serostatus/sexual orientation interactions.

Limitations 01' Cross-Sectional Data

The cross-sectional design of this study precluded strict replication of Folkman et
al.’s (1993) prospective research, in which they were able to control for the effects of past
depressed mood. Since past emotional distress is perhaps the best predictor of current and
future distress, this represents a serious limitation of the current study. In addition,
inferences based on cross-sectional results are limited to correlation. Thus, although path
modeling was employed here to propose structural relationships among variables, it
should be understood that the results in no way demonstrate causal relationships among
these variables. Language such as “these results suggest that X may lead to Y, which in
turn affects Z ” is used merely to suggest possible interpretations of the results.

Scale Psychometrics

Confirrnatory factor analysis of the scales used in this study indicated that most of

them may not measure unitary constructs (see Table 2). Deletion of weaker items did not

118

significantly improve the scales’ performance on Hunter and Hamilton’s (1992) chi-
square tests of internal and external (consistency. These results are somewhat alarming
considering that several of these scales (such as the WOC and the CES-D) are well-
established, widely-used research instruments. Caution should therefore be used in
interpreting results based on these scales.

Possible Confounding of Shame and Depression

As shown in Table 25 (above), respondents’ ISS scores were highly correlated
with their scores on the CES-D depression measure. Corrected correlations between
shame and depression ranged from .82 to .96 (.78 to .90 uncorrected), which are
unusually large for behavioral science constructs (Hunter, class notes). Correlations of
this magnitude suggest possible confounding between the measures of shame and
depression used in this study (Neter, Wasserman, & Kutner, 1985). The Intemalized
Shame Scale (ISS; Cook, 1994) measures shame by asking respondents to endorse certain
experiences and beliefs about the self. It may be that the ISS actually measures depressive
thoughts and beliefs more than shame affect per se, thus confounding it with the measure
of depression.

This criticism is not unique to the ISS; it may be true of any attempt to measure
internal psychological phenomena (such as affect) indirectly through the use of paper-
and-pencil questionnaires. With the ISS, the degree of internalized shame is inferred from
the respondent’s endorsement of experiences and beliefs thought to be consistent with or
indicative of shame affect. It may certainly be the case that the experiences and beliefs
reflected in the ISS items are related to underlying shame; however, many of these items

could just as easily be related to underlying depression, which, while possibly related, is

119

not the same thing as shame. In testing terms, the ISS may demonstrate good sensitivity to
underlying shame, but at the same time demonstrate poor specificity in that it also
measures things other than shame (e.g., depression). This is consistent with the results of
the CFA (Table 2) which suggest that the ISS and CBS-D may not measure unitary

COHSII'UCIS.

W
Sexual Orientation, HIV, and Intemalized Shame
Gay men as a group reported higher levels of internalized shame than heterosexual
men in this study. However, the magnitude of this difference depended‘on the HIV
. serostatus of the gay men. HIV-positive gay men reported .42 SD greater shame than
heterosexual men, whereas HIV-negative gay men reported only a .14 SD difference.
Thus, the gap in reported shame between gay and heterosexual men is three times greater
for HIV-positive gay men than for HIV-negative. Further, within the gay sample, HIV-
positive men reported about .31 SD more internalized shame than HIV -negative gay men.
These results suggest that gay men in general, and HIV -positive gay men in
particular, may be especially vulnerable to shame, which was found here to be associated
with greater use of detachment Coping and increased'depression. This may reflect the
compound societal stigmatization of homosexuality and HIV infection (Herek & Glunt,
1988; Neisen, 1990), as well as the powerlessness and increasing dependency that may
accompany HIV infection (Kaufman, 1989). More about shame and powerlessness will

be discussed in the section on shame and control, below.

120

Sexual Orientation, HIV, and Depression

Gay men as a group reported slightly greater (.17 SD) depressive symptomatology
than heterosexual men. However, as was the case with internalized shame, this difference
was affected dramatically by the HIV serostatus of the gay men. HIV -negative gay men
actually reported slightly less (.18 SD) depression than heterosexual men, while HIV-
positive gay men reported significantly more depression than either heterosexual men (d
= .52) or HIV-negative gay men (d = .66). These results indicate that, within this sample,
HIV serostatus was a much better predictor of depressed mood than sexual orientation,
since HIV -positive gay men were more different from HIV -negative gay men than from
heterosexual men with respect to depression level, while HIV-negative gay men were
similar to (and even slightly less depressed than) heterosexual men. These results contrast
with earlier findings that suggested that homosexuality per se is associated with increased
depression and poorer emotional adjustment (e.g., Bungener et al., 1993; Chuang et al.,
1992; Folkman et al., 1993; Krikorian et al., 1995; Rosenberger et al., 1993).

Physical Symptoms, Stress, and Intemalized Shame

Not surprisingly, physical symptoms have been cited as perhaps the most potent
stressor for people living with HIV/AIDS (Folkman et al., 1993; Hays et al., 1992;
Ostrow, Monjan, et al., 1989). The results of the present study lend further support to this
conclusion. For all of the groups studied here, physical symptoms were positively
correlated with increased stress, and this effect was especially strong for gay men
diagnosed with AIDS (r = .69). Interestingly, HIV-negative gay men reported a stronger
association between physical symptoms and stress than HIV-positive gay men without

AIDS. This finding may reflect fears among “worried well” HIV-negative gay men who

121

perceive themselves to be at risk for HIV infection, and who therefore experience greater
stress when physical symptoms are present.

As predicted, physical symptoms were also associated with increased shame for
all of the groups studied here. Once again, this relationship was particularly strong for gay
men diagnosed with AIDS (r = .47), and may reflect the stigmatizing and disfiguring
nature of many AIDS-related symptoms, as well as the increasing sense of pOwerlessness
' one experiences with declining health (Kaufman, 1989).

Appraisals of Control and Intemalized Shame

As predicted, feeling less in control of stressful situations was associated with
increased shame, particularly among HIV-positive gay men. Tomkins (1995), as well as
Kaufman (1989) and Kaufman and Raphael (1996), have written about the relationship
between powerlessness and shame. Kaufman (1989) describes the “affect dynamics” of
powerlessness in PWAs, for whom the diagnosis evokes a sense of acute helplessness and
uncertainty. According to Kaufman (1989) and Kaufinan and Raphael (1996), fear and
shame may become magnified into humiliation and terror for PWAs, and for HIV-
positive gay men may also reactivate or potentiate shame scenes involving sexuality.
Kaufman (1989) also warns of a "powerlessness-affect-stress” cycle that begins with
feelings of powerlessness or lack of control and culminates in increased stress and
physical symptoms. The results presented here are consistent with Kaufinan’s theory, as
decreased control was found to be associated both with increased stress and increased

shame, which in turn were both associated with increased physical symptoms.

122

Stress and Coping

The data presented here indicate that increased stress was associated with greater
use of both involvement and detachment coping strategies. While a positive relationship
between stress and detachment coping was predicted, the finding that stress was also
related to increased use of involvement coping was somewhat unexpeCted based on
previous findings. Specifically, Folkman et al. (1993) reported a weakly negative
bivariate relationship between these variables, which became non-significant in their
multivariate analysis. It must be remembered, however, that these studies are not directly
comparable, since F olkman and her colleagues used a shortened version of the WOC
while the present study used the full-length “gay male” version of the instrument. It is
difficult to say whether differences in the two versions of the scale account for the
differing relationships observed between stress and involvement coping. In any case, the
present findings indicate that among both heterosexual and gay men, increased stress was
associated with the mobilization of a wide variety of coping behaviors. In other words,
the effects of stress appeared to be non-specific with respect to the type of coping
behaviors (detachment versus involvement) elicited.

The above comments pertain when assessing detachment and involvement coping
as separate variables. When a single measure of relative coping preference was
calculated, however, it was shown that among gay men with AIDS, increased stress was
in fact related to relatively greater use of detachment as Opposed to involvement coping
strategies, as predicted. Thus, while heightened stress may be associated with increases in
a variety of different behaviors intended to help the person cope, the data suggest a

relatively greater use of detachment over involvement-oriented strategies as gay men with

123

AIDS become more stressed. This relative preference for detachment over involvement
c0ping was also associated with increased shame and depression among HIV-positive gay
men (regardless of AIDS diagnosis).

Appraisals of Control and Coping

Lazarus and F olkman’s (1984) theory/suggests that greater feelings of control over
stressors should lead to increased use of active, involvement-oriented coping behaviors,
while feeling less control over stressors should lead to greater detachment. The data
. presented here lend mixed support to this hypothesis. Multivariate analyses showed that
control was not significantly related to either type of coping for gay men as a group, while
among heterosexual men there were modest (non-significant) positive correlations
between control and both types of coping. When a relative measure of coping preference
was used, a trend (r = -.35, p = .055) was found for decreased control to be associated
with a relative preference for detachment over involvement coping, consistent with
Lazarus and F olkrnan’s theory. This moderate effect would have been statistically
significant if the N of gay men with AIDS (31) had been larger.

Coping and Depression

The results of this investigation are consistent with the hypothesis that the type of
behaviors used to cope with stressors affects emotional adjustment and distress. Active,
involvement-oriented coping strategies were associated with decreased depressive
symptomatology among all of the study subgroups, especially gay men with AIDS. On
the other hand, coping that emphasized detachment from or denial of problems was

associated with increased depression among all of the subgroups, with the strongest

124

effects observed for gay men in general and HIV-positive gay men in particular. These
findings hold important implications for clinical intervention, as will be discussed below.

Intemalized Shame and the Stress and Coping Model

One of the key questions addressed by this research was whether Lazarus and
Folkman’s (1984) cognitive-behavioral stress and coping model could be improved by the
addition of an affect component, i.e., shame. Bivariate analyses indicated that gay men in
general, and HIV-positive gay men in particular, reported more shame than heterosexual
men, and that increased shame was related to increased physical symptoms, increased
stress, decreased control, greater use of detachment coping, decreased use of involvement
coping, and increased depression. Further, the multivariate path analyses indicated that,
for gay men, shame was the best predictor of the use of involvement versus detachment
coping strategies. For these men, increased shame was associated with increased use of
detachment coping, which in turn was related to increased depression. On the other hand,
decreasedshame was associated with increased use of involvement coping, which was
related to decreased depression. Thus, level of reported shame differentiated between the
types of coping used, unlike control, Which was unrelated to either type of coping among
gay men, or stress, which was positively associated with both types of coping but did not
differentiate between them. Shame also proved important for heterosexual men in the
multivariate analysis, in that it was related to increased use of detachment coping, as well
as showing a direct positive effect on depression. Taken together, these findings suggest

that shame is an important addition to the stress and coping model.

125

IMPLICATIONS FOR CLINICAL INTERVENTION

The results of this investigation hold important implications for clinical practice
and intervention, particularly with respect to PWAs. The data presented here suggest that
HIV-positive gay men, especially those diagnosed with AIDS, may be particularly
vulnerable to the deleterious effects of stress. Increased stress, and the decrease in
perceptions of control associated with it, may lead to greater use of maladaptive coping
strategies involving detachment and denial, which in turn may be associated with
increased emotional distress.

In addition, the data presented here indicate that shame is related to the greater use
of maladaptive avoidance-oriented coping behaviors, which are related to increased
.depression. Psychologist Gershen Kaufman has focused his clinical practice on the
psychotherapy of “shame-based syndromes” (Kaufman, 1989) and has written extensively
on the deleterious effects of shame on mental health, particularly as this relates to gay
men and lesbians. Kaufman and Raphael (1996) argue that the societal stigmas associated
with homosexuality and AIDS have become inextricably intertwined, a situation which
may lead to compound shame for HIV-positive gay men:

The cultural shame and disgust about homosexuality have been displaced onto

AIDS and thus transferred directly to people with AIDS. But the reverse is also

happening: the shame, dissmell, and disgust about AIDS have been displaced onto

gays. Because of these two factors, being gay unfortunately equals having AIDS in
the minds of many people. By perceiving AIDS as a “gay disease,” our culture

equally repudiates gays and people with AIDS via dissmell and contempt (p. 102).

The present study suggests that mental health professionals working with HIV-

positive gay men should help these at-risk men better manage stress by teaching and

reinforcing active, involvement-oriented coping strategies, thereby helping them to

126

maintain or increase their sense of personal control, which has often been compromised
by their experiences living with the uncertainties of HIV /AIDS. In addition, addressing
shame issues with these men is critical, particularly due to the demonstrated relationship

between shame and coping.

SUGGESTIONS FOR FUTURE RESEARCH

Recruitment of research participants is challenging, and this is even more the case
when studying specialized populations such as gay men or persons living with
HIV/AIDS. The current study was fortunate to have access to both an agency sample of
gay men living with HIV and a community sample of HIV-negative gay men. Future
studies could benefit from including a non-agency related (i.e., community-based) sample
of gay men living with HIV to determine if these men differ in important ways from those
recruited through HIV/AIDS service organizations. Further, if large sampling frames of
potential respondents could be identified, random sampling fi'om within those frames
could improve the quality of the statistical inferences drawn from the data. Finally,
attempts should be made to recruit samples that are. similar to one another with respect to
demographic variables such as age, education, and income.

Future studies would also benefit from including other groups at increased risk for
HIV infection, such as women, ethnic minorities, sex trade workers, and IV drug users.
These groups have been underrepresented in the literature despite representing a growing
proportion of new HIV infections (CDC, 1997).

If feasible given available resources, prospective (longitudinal) research designs

should be utilized in future investigations to study changes in stress and coping processes

127

and emotional adjustment over time, and to allow cross-sectional analyses to control for
previous levels of the dependent variable.

Future investigations should use different assessment techniques for measuring
shame. Fortunately, a number of new measures of shame and guilt have been developed
in recent years (Tangney, 1996). It would be instructive to note the degree of
correspondence between different methods for measuring shame, as well as assess their
' relationship to other variables of interest such as depression.

Future studies should consider measuring coping within more than one domain. In
particular, investigations of coping among PWAs should assess both “general” coping
(i.e., coping with “typical” everyday stressors and life events) and coping with
HIV/AIDS-specific stressors such as physical symptoms, AIDS-related worries, losing a
loved one to AIDS, etc. (DeMarco et al., 1997). By assessing coping across domains, the
researcher can discern whether PWAs cope differently with AIDS-related stressors than
they do with non-AIDS-related stressors, as well as assess whether coping in one domain
has different implications for emotional adjustment than coping in other domains. This
knowledge would allow helping professionals to better focus their clinical interventions
with PWAs.

Researchers should make an effort to improve the quality of the measurement
scales used in social science research. The results of the confirmatory factor analysis of
the scales used in the present study clearly show that well-known, widely-used i
instruments with high coefficient alphas can nevertheless demonstrate poor intemal and

external consistency when subjected to stricter tests.

128

Finally, future research should investigate the effects of the new combination drug
therapies on the psychosocial functioning of HIV-infected persons. Since these new
treatments were introduced in 1996 there have been dramatic decreases in morbidity and
mortality attributable to HIV disease in the US (CDC, 1997). In many cases, persons
thought to be terminally ill have demonstrated a remarkable response to the medications.
Opportunistic infections disappear, and often the person’s viral load drops to undetectable
levels within the bloodstream. These treatment advances seem almost miraculous, and
offer renewed hope for an eventual cure for HIV disease. However, for persons “snatched
fi'om the jaws of dea ” by the new treatments, there may be new stressors to cope with as
well. These might include planning for an indefinite extension of life after having already
made arrangements for death, survivor’s guilt, existential issues, and the like. Future
research will need to map out this new territory so that our understanding keeps pace with

the growing complexities of living with HIV/AIDS.

APPENDICES

APPENDIX A

APPENDD( A

EXAMPLE OF WRITTEN NOTICE USED IN PARTICIPANT RECRUITMENT

PARTICIPANTS NEEDED for a research study at Michigan State University examining
the emotional well-being of gay men with and without HIV/AIDS. The study will focus
on how stress affects emotional wellness. Participation will involve filling out a
questionnaire that takes 30-45 minutes to complete. All responses to the research
questions will be kept strictly confidential. All respondents will receive a $10 Meijer gift
certificate for their participation. QUALIFICATIONS: Gay men 18 years of age or older
who have undergone HIV antibody testing and know their results. If you are interested in
participating, please contact Frank DeMarco by phone at (517) 355-9561 (leave name,
number, and the message “research project”), by electronic mail at ‘
demarcof@pilot.msu.edu, or by US mail c/o Department of Psychology, 129 Psychology
. Research Bldg., Michigan State University, East Lansing, MI 48824-1117.

131

APPENDIX B

- APPENDIX B

RESEARCH INSTRUMENT

Research Questionnaire

for a study of
Health, Stress, Coping,
And Emotional Well-Being

Principal Investigator:
Frank J. DeMarco, MA

Supervising Faculty Member:
Robert Caldwell, Ph.D.

Department of Psychology
Michigan State University

134

RESEARCH CONSENT STATEMENT

1. I have freely consented to take part in a scientific study being conducted by Frank
DeMarco, under the supervision of Professor Robert Caldwell (MSU Department of
Psychology).

2. I certify that I am 18 years of age or older.

3. I understand that this study concerns health, stress, and coping, and how they affect a
person’s emotional well-being.

4. I understand that my participation involves filling out the following questionnaire,
which takes approximately 20-30 minutes to complete.

5. Although I am encouraged to complete the entire questionnaire, I understand that I am
free to discontinue my participation in this study at any time without penalty.

6. I understand that my individual responses to this survey will be kept strictly
confidential and that I will remain completely anonymous. Study results will be
presented in summary form only, without reference to responses given by individual
participants. At no time will Frank DeMarco or anyone other than authorized
staff members of the agency that referred me to this study have access to my
name, address, or phone number unless I choose to reveal this information.

7. I understand that II will receive a ten dollar ($10) Meijer gift certificate for my
participation in this study. The gift certificate will be mailed to me by the agency
that referred me to this study as soon as my completed questionnaire is received by
Frank DeMarco. I understand that I must return the questionnaire directly to
Frank DeMarco, using the pre-addressed postage-paid envelope provided.

8. I understand that a summary of the study results will be made available to me at my
request once the study is completed (summer 1997), within the restrictions outlined in
section (6) above. I may request this smnmary by contacting the agency that referred
me to this study and stating that I participated in Frank DeMarco’s survey.

9. I understand that my participation in this study does not guarantee any benefits to me
beyond what is stated in sections (7) and (8) above.

10. I understand that, at my request, I can receive additional explanation of the study after
my participation is completed by contacting Frank DeMarco at (517) 355-9561 (MSU
Department of Psychology, 129 Psychology Research Bldg., East Lansing 48824-
1117)

By completing and returning this questionnaire you indicate
your understanding of this consent statement and
your voluntary agreement to participate in this study.

 

 

 

 

135

We all experience stress in our lives from time to time.

To what extent have each of the following areas of your life been stressful for you
during the past month? Please circle one number for each item.

IMPORTANT: If you feel that an item does not apply to you, please write “NIA”
next to that item and skip to the next one.

 

 

 

 

 

 

 

 

 

 

 

 

DURING THE PAST MONTH:
Not At All A Little Bit ' ' Quite A Bit Extremely
Stressful Stressful Stressful Stressful
a) Romantic relationship(s) 0 1 2 3
b) Relationships with friends 0 1 2 3
' c) Relationships with family 0 1 2 3
d) Work 0 1 2 3
c) School 0 1 2 3
0 Your fmances 0 1 2 3
g) Illness of someone close to you 0 l 2 3
h) Death of someone close to you 0 1 2 3
i) Your own health 0 1 2 3
j) Political issues 0 1 2 3
k) Other (please specify): 0 1 2 3

 

Now, please look over the list above. Which of these areas has been the MOST
stressful for you during the past month? If two or more seem equally stressful, pick
one.

 

Write your choice in here:
Most stressful area of my life in the past month

You will be asked to refer to this choice later in the questionnaire.

136

Now here is the same list of life areas that you just rated for stress.

To what extent have you felt that you had control over each of these areas of your
life during the east month? Please circle one number for each item.

IMPORTANT: If you feel that an item does not apply to you, please write “NIA”
next to that item and skip to the next one.

 

 

 

 

 

 

 

 

 

 

 

DURING THE PAST MONTH:
Almost Never Sometimes Often Almost Always
Felt In Felt In Felt In Felt In
Control Control Control Control
a) Romantic relationship(s) 0 1 2 3
b) Relationships with friends 0 l 2 3
c) Relationships with family 0 l 2 3
d) Work 0 1 2 ' 3
e) School 0 l 2 3
0 Your finances 0 1 2 3
g) Illness of someone close to 0 l 2 3
you
b) Death of someone close to 0 l 2 3
you '
i) Your own health 0 1 2 3
j) Political issues ' 0 l 2 3

 

k) Other (please specify): 0 1 2 3

 

 

137

On page 2 above you indicated which area of your life has been the most stressful
for you during the past month. Please review your choice, and write it in here:

Most stressful area of my life during the past month:

 

The following is a list of ways that people attempt to cope with stressful situations.

Please read each item below and indicate, by circling the appropriate number, are
artent to which you usgt_l_i_t in the situation you have just described as being the most
stressful area of your life during the past month. (If an item does not apply to your
situation, circle 0.)

 

 

 

 

 

 

 

 

 

 

 

 

Not Used Used Used
Used somewhat quite a bit a great deal

1. Just concentrated on what I had to do 0 1 2 3
next—the next step. _

2. I tried to analyze the problem in order to 0 l 2 3
understand it better.

.3. Turned to work or substitute activity to 0 l 2 3

take my mind off things.

4. I felt that time would make a difference— 0 1 2 ' 3
the only thing to do was to wait.

5. Bargained or compromised to get . 0 1 2 3
something positive from the situation.

6. I did something which I didn’t think would 0 1 2 3
work, but at least I was doing something. '

7. , Tried to get the person responsible to 0 l 2 3
change his or her mind.

8. Talked to someone to find out more about 0 1 2 3
the situation.

9. Criticized or lectured myself. 0 l 2 3

10. Tried not to burn my bridges, but leave 0 1 2 3
some things open.

11. Hoped a miracle would happen. ‘ 0 1 2 3

 

12. Went along with fate; sometimes I just 0 1 2 3
have bad luck. «

138

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Not Used Used Used
Used somewhat quite a bit a great deal

13. Went on as if nothing had happened. 0 l 2 3

14. I tried to keep my feelings to myself. 0 1 2 3

15. Looked for the silver lining, so to speak; 0 1 2 3
tried to look on the bflt side of things.

16. Slept more than usual. 0 1 2 3

17. 1 expressed anger to the person(s) who 0 1 2 3
caused the problem

18. Accepted sympathy and understanding 0 . 1 2 3
from someone.

19. I told myself things that helped me to feel 0 1 2 3
better.

20. I was inspired to do something creative. 0 1 2 3

21. Tried to forget the whole thing. 0 1 2 3

22. I got professional help. 0 1 2 3

23. Changed or grew as a person in a good 0 1 2 3
way.

24. I waited to see what would happen before 0 1 2 3
doing anything.

25. I apologized or did something to make up. 0 1 2 3

26. I made a plan of action and followed it. 0 1 2 3

27. I accepted the next best thing to what I 0 1 2 3
wanted.

28. I let my feelings out somehow. 0 l 2 3

29. I realized I brought the problem on myself. 0 l 2 3

30. I came out of the experience better than 0 l 2 3

when I went in.

 

 

139

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

get too serious about it.

Not Used Used Used
Used somewhat quite a bit a mat deal

31. I talked to someone who could do 0 1 2 3
something concrete about the problem

32. 1 got away from it for awhile; tried to rest 0 1 2 3
or take a vacation.

33. I tried to make myself feel better by eating. 0 1 2 3

34. I tried to make myself feel better by 0 1 2 3
drinking.

35. I tried to make myself feel better by using 0 1 2 3
recreational drugs (e.g., cocaine, crack,

, marijuana).

36. I tried to make myself feel better by using 0 l 2 3
prescribed mood-altering drugs (e.g.,
Valium, Halcion, Xanax).

37. I took a big chance or did something very 0 l 2 3
risky.

38. I tried not to act too hastily or follow my 0 1 2 3
first bunch.

39. I found new faith. 0 1 2 3

40. l maintained my pride and kept a stiff 0 1 2 3
upper lip.

41. I rediscovered what is important in life. 0 l 2 3

42. I changed something so things would turn 0 1 2 3
out all right.

43. I avoided being with people in general. 0 1 2 3

44. I didn’t let it get to me; I refused to think 0 l 2 3
too much about it.

45. I asked a relative or friend I respected for 0 l 2 3
advice.

46. I kept others from knowing how bad things 0 1 2 3
were.

47. I made light of the situation; I refused to 0 1 2 3

 

140

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Not Used Used . Used
Used somewhat quite a bit a great deal

48. I talked to someone about howl was 0 1 2 3
feeling.

49. I stood my ground and fought for what I 0 1 2 3
wanted.

50. To feel better, I had sex with my primary 0 1 2 3
partner.

51. I took it out on other people. 0 1 2 i 3 ’

52. I drew on my past experiences; I was in a 0 1 2 3
similar situation before.

53. I knew what had to be done, so I doubled 0 l 2 3
mxefforts to make things work.

54. I refused to believe that it had happened. 0 1 2 3

55. I made a promise to myself that things 0 1 2 3
would be different next time.

56. I came up with a couple of different 0 l 2 3
solutions to the problem

57. I accepted it, since nothing could be done. 0 1 2 3

58. I tried to keep my feelings from interfering 0 1 2 3
with other things too much.

59. I wished that I could change what had 0 1 2 i 3
happened or how I felt.

60. I changed something about myself. 0 1 2 3

61. I daydreamed or imagined a better time or 0 1 2 3
place than the one I was in.

62. I wished that the situation would go away 0 l 2 3
or somehow be over with.

63. I had fantasies or wishes about how things 0 l 2 3
might turn out.

64. I prayed. 0 l 2 3

65. I had anonymous sex to feel better. 0 l 2 3

 

 

 

141

 

 

 

 

 

 

 

 

 

 

 

Not Used Used Used
Used somewhat quite a bit a great deal
66. I had non-anonymous sex with someone to 0 l 2 3
feel better.
67. I prepared myself for the worst. 0 1 2 3
68. I went over in my mind what I would say 0 1 . 2 3
or do.
69. I thought about how a person I admire . .0 1 2 3
would handle this situation and used that
as a model.
70. I tried to see things from the other person’s 0 1 2 3 A
point of view. ,_
71. I reminded myself how much worse things 0 1 2 3
could be.
72. I jogged or exercised. 0 1 2 3
73. I meditated or used imagery. 0 1 2 3 J

 

142

Below is a list of statements describing feelings or experiences that you may have
from time to time. Most of these statements describe feelings and experiences that
are generally painful or negative in some way. We have all had some of these
feelings at some time. Read each statement carefully and circle the number to the
right of the item that indicates how frequently you find yourself feeling or
experiencing what is described in the statement.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SOME- ALMOST
NEVER SELDOM TIMES OFTEN ALWAYS

l. I feel like I am never quite good 0 1 2 3 4
enflh.

2. I feel somehow left out. 0 l 2 3 4

3. I think that people look down on me. 0 1 2 3 4

4. All in all, I am inclined to feel that I 0 1 2 3 4
am a success.

5. I scold myself and put myself down. 0 1 2 3 4

6. I feel insecure about others’ opinions 0 1 2 3 - 4
of me.

7. Compared to other people, I feel like 0 1 2 3 4
I somehow never measure up.

8. I see myself as being very small and 0 l 2 3 4
insignificant.

9. I feel I have much to be proud of. 0 l 2 3 4

10. I feel intensely inadequate and full of 0 l 2 3 4
self doubt.

11. I feel as if I am somehow defective as 0 1 2 3 4
a person, like there is something
basically wrong with me.

12. When 1 compare myself to others I 0 1 2 3 4
am just not as important.

13. I have an overpowering dread that 0 1 2 3 4
my faults will be revealed in front of
others.

14. I feel I have a number of good 0 l 2 3 4
qualities.

15. 1 see myself striving for perfection 0 1 2 3 4
only to continually fall short.

16. I think others are able to see my 0 1 2 3 4
defects.

17. I could beat myself over the head 0 1 2 3 4

with a club when I make a mistake.

143

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SOME- ALMOST
NEVER SELDOM TIMES OFTEN 'ALWAY

18. On the whole, I am satisfied 0 1 2 3 4 ‘
with myself.

19. I would like to shrink away 0 1 2 3 4
when I make a mistake.

20. I replay painful events over and 0 l 2 3 4
over in my mind until I am
overwhelmed.

21. Ifeelil am a person of worth at 0 l 2 ~ 3 4
least on an equal plane with
others.

22. At times I feel like I will break 0 1 2 3 4
into a thousandJieces.

23. I feel as if I have lost control 0 1 2 3 4
over my body functions and my '
feelings.

24. Sometimes I feel no bigger than 0 1 2 3 4
a pea.

25. At times I feel so exposed that I 0 1 2 3 4

, wish the earth would open up

and swallow me.

26. I have this painful gap within 0 1 2 3 4
me that l have not been able to
fill. 4

27. I feel empty and unfulfilled. ' o 1 2 3 4

28. I take a positive attitude toward 0 l 2 ‘ 3 4
myself.

29. My loneliness is more like 0 1 2 3 4
emptiness.

30. I feel like there is something 0 l 2 3 4

missing.

 

144

We all have times when we feel mostly happy, and other times when we feel down or
blue.

Below is a list of some ways you might have felt or behaved recently. Please indicate
how often you have felt or behaved in these ways during the past week by circling the
appropriate number to the right of each item. Use the scale given below:

0 1 - 2 3
RARELY OR NONE SOME OR A LITTLE OCCASIONALLY . MOST OR
OF THE THWE OF THE TIME OR A MODERATE ALL OF THE
(Less than 1 day out of (1-2 days out of the AMOUNT OF THE TIME
the past week) past week) TIME (5-7 days)
(3-4 days)

Duringt hte past week.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1. I was bothered by things that usually don’ t bother me. 0 1 2 3
2. I did not feel like eating; my appetite was poor. 0 l 2 3
3. I felt I couldn’t shake off the blues, even with help from family/friends. 0 1 2 3
4. I felt that I was just as good as other people. 0 1 2 3
5. I had trouble keeping my mind on what I was doing. 0 l 2 3
6. I felt depressed. 0 1 2 3
7. I felt that everything I did was an effort. 0 1 2 3
8. I felt hopeful about the future. 0 1 2 3
9. I thought my life had been a failure. 0 1 2 3
10. I felt fearful. 0 1 2 3
11. My sleep was restless. 0 l 2 3
12. I was happy. 0 l 2 3
13. I talked less than usual. 0 l 2 3
' 14. I felt lonely. 0 l 2 3
15. People were unfriendly. 0 1 2 3
16. I enjoyed life. 0 1 2 3
17. I had crying spells. 0 1 2 3
18. I felt sad. 0 l 2 3
19. I felt that people dislike me. 0 1 2 3
20. I could not get “going.” 0 1 2 3

 

145

Please tell us about your health.

Have you ever been tested for exposure to tuberculosis (TB)?

Yes

No

Don’t know

If yes, have you ever tested positive fOr exposure to TB?

Yes

No

Don’t know __

Have you ever been tested for exposure to the human immuno-
deficiency virus (HIV)—the virus that is believed to cause AIDS?
Yes
No

Don’t know __
If you answered n_o or don ’t know, please go on to the next page. ->
If yes: How long ago was your most recent HIV test? Less than 3 months ago __
3 - 6 months ago

6 - 12 months ago

Over 12 months ago

Have you ever tested positive for HIV? Yes

No

Don’t know

If HIV-positive: Have you been diagnosed with AIDS? Yes
No

Don’t know

146

Please indicate which of the following physical symptoms, if any, you have
experienced during the past year (check all that apply):

Thrush, candida, or white patches in the mouth or throat for at least 2 weeks _
Unintentional weight loss of at least 10 pounds (not related to dieting) __
Diarrhea that lasted at least 2 weeks __
Sweating at night for at least 2 weeks
Hairy leukoplakia (a white coating) on the tongue _
A new or unusual dry cough lasting at least 2 weeks __
Persistent sore mouth or throat lasting at least 2 weeks
An unusual bruise, bump, or skin discoloration that lasted at least 2 weeks __
Persistent fatigue for at least 2 weeks _
Persistent or recurring fever of at least 100° for at least 2 weeks _
Tender or enlarged glands or lymph nodes (not counting groin) __
Herpes zoster (shingles) _
Persistent shortness of breath for at least 2 weeks __
A new skin rash that lasted at least 2 weeks __
Persistent, frequent, or unusual headaches for at least 2 weeks
Blurred vision, light flashes, or other unusual vision problems for at least 2 weeks __

Other (please describe):

 

 

 

147

Finally, please tell us about yourself.

What is your date of birth?

 

MM-DD- Y Y

Which of the following best describes your racial or cultural origin?
Asian/Pacific Islander __

Black/African American __

Latino/Hispanic __

Middle Eastern __

Native American __

White/Caucasian __

Other (please explain below)

 

Which of the following best describes your level of education?
Less than HS diploma __

HS diploma __

Some college _

4-yr. college degree _

Advanced degree

Which of the following best describes your sexual orientation?
Heterosexual
Bisexual

' Gay

148

Which of the following best describes your total ANNUAL household
income? (NOTE: If you depend on someone else for most of your
income, please indicate that person ’s total household income to the
best of your knowledge.)

Under $10,000 __
510,000-530,000 _
$30,001-S60,000_ !

$60,001-$100,000__ 3

Over $100,000

 

Which of the following best describes your use of computers?
Never __

Occasionally (several times a month) __

Frequently (several times a week) _

Daily __

Do you have access to the Internet?

Yes

No

Don’t know

Do you have access to a World Wide Web browser (such as Netscape“)?

Yes

No

Don’t know

149

If this survey had been available on the Internet or the World Wide
Web, would you have been more inclined to fill it out, less inclined to
fill it out, or would it have made no difference to you?

More inclined
Less inclined
No difference

Don’t know

Thank you very much
for participating in this study.

10.

150

RESEARCH CONSENT STATEMENT (for student participants)

 

I have freely consented to take part in a scientific study being conducted by Frank
DeMarco, under the supervision of Professor Robert Caldwell (MSU Department of
Psychology).

I certify that I am 18 years of age or older.

I understand that this study concerns health, stress, and coping, and how they affect a
person’s emotional well-being. '

I understand that my participation involves filling out the following questionnaire, !
which takes approximately 20-30 minutes to complete.

Although I am encouraged to complete the entire questionnaire, I understand that I am
free to discontinue my participation in this study at any time without penalty.

 

I understand that my individual responses to this survey will be kept strictly J
confidential and that I will remain completely anonymous. Study results will be ‘
presented in summary form only, without reference to responses given by individual

participants.

I understand that I will receive two (2) psychology experiment credits for my
participation in this study.

I understand that a summary of the study results will be made available to me at my
request once the study is completed (summer 1997), within the restrictions outlined in
section (6) above. I may request this summary by contacting Frank DeMarco at (517)
355-9561 (MSU Department of Psychology, 129 Psychology Research Bldg, East
Lansing 48824-1 117). ‘

I understand that my participation in this study does not guarantee any benefits to me
beyond what is stated in sections (7) and (8) above.

I understand that, at my request, I can receive additional explanation of the study after
my participation is completed by contacting Frank DeMarco at (517) 355-9561 (MSU
Department of Psychology, 129 Psychology Research Bldg., East Lansing 48824-
1117)

By completing and returning this questionnaire you indicate

your understanding of this consent statement and
your voluntary agreement to participate in this study.

 

 

APPENDIX C

APPENDIX C
EXPANDED DISCUSSION OF
AFFECT THEORY VS. PSYCHOANALYTIC APPROACHES
TO UNDERSTANDING SHAME
THE PSYCHOANALYTIC PERSPECTIVE ON SHAME: THE WORK OF
HELEN BLOCK LEWIS

Definition of Shame El

Helen Block Lewis is widely recognized as the foremost modem proponent of the

 

psychoanalytic perspective on shame. Shame is defined by Lewis (1971) as an affective _
state evoked by the internal self-monitoring regulatory agency (i.e., the superego) to U
control the drives, especially sexual, and to maintain or repair attachment bonds to others.

According to Lewis (1987), shame arises out of a failure of the central attachment
bond between self and other. This failure evokes humiliated rage directed both at the
other for the perceived scorning and rejecting of the self, and toward the self for being
unworthy of the other’s esteem, for caring so deeply about the other’s opinion of the self,
and for being passive and dependent. Additional rage is then deflected toward the self
due to guilt about directing “unjust” rage at the other.

Lewis distinguishes between shame and guilt as being diflerent affects, tied to
different aspects of the superego. For Lewis (1971), the structure of the superego
incorporates both a “castration threat” (threat of punishment) which makes possible the
sense of guilt, and an “ego ideal,” which makes possible the sense of shame (or pride)
about failing (or succeeding) to live up to an internalized standard. Shame and guilt can

thus be viewed as equally advanced, but different, superego functions. An overly harsh

152

 

 

153

superego has long been implicated as pathogenic in psychoanalytic thought. Freud (1923)
contended that a malfunctioning or "archaic" superego is connected to neurotic
symptoms. Kohlberg (1968) wrote that the neurotic has trouble maintaining a balanced
self-evaluation due to an overly harsh superego.

Lewis contends that the one-sided emphasis in psychoanalytic thought and writing
on the castration threat and guilt-related aspects of the superego has led to the neglect of
shame. Lewis (1987) refers to shame as the “sleeper” of psychopathology.
Phenomenology of Shame and Guilt

In guilt, one’s value system is the referent or focal point of awareness. Guilt is the
affect that accompanies a transgression of rules. In shame, it is the internalized "other"
that serves as the referent. The shame experience is one of feeling exposed as falling
short "in the eyes of" the other, i.e., of failing to live up to the ego ideal. The "other" is
thus experienced as rejecting the self because the self is defective or unworthy.
"Shame...involves more self-consciousness and more self-imaging than guilt" (1971, p.
30). The self in general is the focus of shame; a specific behavior is the focus in guilt.
With shame, the source of the negative evaluation of the self is localized as "out there"; in
guilt, it is localized as originating from within the self, as a violation of internal
standards. The shame experience involves feedback from all perceptual modalities and
considerable autonomic stimulation; thus, there is much greater body awareness in the
shame state than in the guilt state. Shame is therefore more autonomic, whereas guilt is
more cognitive. Because shame is less rational than guilt, it can be evoked by a

“rationally trivial” stimulus which feels disproportionately painful.

 

 

 

154

Dissipation of Shame

Shame and guilt are acutely. disturbing affects and must be discharged or
dissipated as quickly as possible in order for the individual to return to a state of
emotional equilibrium. Shame & guilt thus evoke defenses which force compromises of
affect and behavior in an attempt to maintain or reestablish emotional homeostasis, self-
esteem, and relationships with others. Lewis refers to this as a "righting tendency" (1971, h

p. 26) which serves to dissipate shame and guilt. For example, shame reactions can be

 

dissipated through some gesture of kindness or reassurance fiom another, while guilt can - U i
be dissipated by some action that makes amends for the transgression. I
Shame, however, is often difficult to discharge. First, it is often difficult to

recognize the shame state. Shame is ofien fused and confused with guilt; the more
autonomic nature of shame as compared to the more cognitive guilt state makes shame
less accessible, and leads to denial or by-passing of the shame affect. Shame is therefore
difficult to identify, and thus is not easily dissipated. This split between cognition and
affect, and the global generalization of shame to the entire self, make shame difficult to
resolve.

Second, shame leads to a “malfunctioning” of the self. The self is reduced in size '
and efficiency of functioning during shame. The shamed self seeks to hide, but is also at
the center of the experience, painfully exposed and self-aware. Shame feels involuntary
and irrational, and the self is experienced as passive. Thus the shamed self feels impotent

and is less able to mobilize mature ego defenses to dissipate the shame.

 

 

155

Third, hostility evoked by the shame reaction is difficult to discharge because it is
directed toward a valued other and thus is often redirected back toward the self as guilt.
This inward-directed hostility leads to symptoms such as depression, both directly and via
the strangulation of shame affect.

Shame and Psychological Symptoms

Lewis writes that “states of shame and/or guilt which cannot be discharged or
righted lead into ‘primary process’ transformations or psychic symptoms” (p. 27). Shame
and guilt that cannot be dissipated by more mature defenses (such as intellectualization,
for example) are often driven out of awareness by more primitive defenses (such as
‘ repression). These affects are thereby repressed or “strangulated.” Strangulated affect
finds compromise expression in symptoms. Feelings such as worry, anxiety, and
depression can appear without awareness of the underlying states of guilt or shame that
evoke them.

Lewis (1987) describes a pattern in which shame leads quickly to humiliated fury
and retaliatory urges, which then evoke guilt over “unjust” or “irrational” rage. This rage
is redirected back against the self, leading to a further lowering of self-esteem which
makes the self more susceptible to subsequent shaming and leads to psychological
‘ problems such as depression. This cycle is referred to by Scheff ( 1987) as the “shame-
rage spiral.”

This cycle is further amplified by the fact that people feel shame about being

ashamed. Adult life, with its emphasis on rationality, leaves little room for the shame

 

 

 

l
l
156

experience, which relies largely on imagery and autonomic reactions. Thus the feeling of
shame itself has stigma attached to it due to its “irrational” and out-of-control nature.
Lewis argues that shame has been largely neglected in psychoanalytic writing, and
where it is discussed it has often been misunderstood. Most psychoanalytic (as well as
behavioral) theory has viewed human nature as essentially individualistic or narcissistic. ' ..
Lewis feels that greater emphasis needs to be placed on crucial attachments to others 1

across the life span. Although shame is filndamentally “about the self,” it is intimately

 

other-connected as well—it arises in relation to the other when attachment bonds are U
threatened and when the self is experienced as falling short of the other’s expectations.

Another way in which shame has been misunderstood in classical psychoanalytic
theory according to Lewis is that it has been considered a more primitive emotion than
guilt. Shame has been viewed as pre-oedipal, and guilt as oedipal. Lewis argues that
there is no evidence that shame is a more primitive affect than guilt.

Thus, in a reformulation of the role of shame within psychoanalytic theory, shame
affect functions not only to control drives but to maintain attachments, according to
' Lewis. Shame is the universal human response to loss of love or other failure of the
central attachment bond to the other. Shame is so disturbing and so disorienting that it
motivates one to repair or reestablish damaged or severed attachment bonds (Lewis,
1987)

In this schema depression results from a sequence in which shame is followed by
humiliated fury which, in turn, is guiltily redirected back at the self. This results in a

further lowering of self-esteem and increased feelings of worthlessness, helplessness,

 

157

impotence, passivity, and ultimately depression. Women in Lewis’s view may be more
prone to depression in our culture because they are less free to openly express hostility or
aggression (therefore increasing the likelihood that aggression will be redirected against
the self) and because they place greater value on interpersonal attachment bonds. Women
are therefore more disturbed when those bonds are compromised.

' THE AFFECT THEORY PERSPECTIVE ON SHAME: THE WORK OF
SYLVAN TOMKINS

A General Introduction To Affect Theory

The perspective on shame offered by Silvan Tomkins is rooted in his seminal
work on affect theory. Tomkins views affect as the primary motivating factor in human
behavior. This view differs markedly from the traditional psychoanalytic perspective in
which the drives are primary in governing behavior.

Tomkins was the first theorist to conceive of drives and affects as constituting
separate systems (Kaufman & Raphael, 1996). The drive system, according to Tomkins,
firnctions to meet physiological needs for food, water, oxygen, sleep, warmth, and sex,
and the avoidance of pain. The innate affect system consists of six primary aflects (three
positive and three negative), one aflect auxiliary, and two drive auxiliaries. All but the

drive auxiliaries are described by a pair of terms representing a range of intensity.

 

158

Positive Primary Affects:

Interest—Excitement, Enjoyment—Joy, Surprise—Startle

Negative Primary A fleets:

Distress—Anguish, Fear—Terror, Anger—Rage
A fleet Auxiliary:
Shame—Humiliation

Drive Auxiliaries:

Dissmell and Disgust (innate defensive responses tied to the oxygen,

hunger, and thirst drives)
The positive affects are innately rewarding and we are motivated to maximize them. The
negative affects are innately punishing and we are motivated to minimize them. Negative
affect magnifies more quickly and easily than positive affect, and will often overwhelm
positive affect (Kaufinan & Raphael, 1996). It is easy to see how this arrangement is
adaptive; for instance, survival requires that fear of a potentially dangerous stimulus be
able to quickly overcome interest in that stimulus.

In Tomkins’s schema, the affect system is considered primary over the drive
system because the drives require amplification by affects in order to function. For
example, the sex drive must be fused with excitement affect in order to function properly;
any impediment to excitement, such as shame, disgust, or fear, will quickly disrupt the
sex drive and in many instances make it impossible for individuals to function sexually

until the impediment is removed. Likewise, the hunger and thirst drives can be quickly

 

 

 

159

overridden by disgust, which functions to prevent the ingestion of toxic or noxious
substances (Kaufman & Raphael, 1996).

Both the primary and auxiliary affects are considered by Tomkins to be innate, not
learned; they are biologically endowed, and therefore part of our evolutionary inheritance.
However, various combinations of the innate affects, whiCh constitute our more complex
emotional reactions such as contempt, depression, and jealousy, are learned through
experience (Kaufinan & Raphael, 1996). Similarly, the association of certain affects or
combinations of affects with particular stimuli (situations, objects, events, etc.) is also
learned through experience.

A distinction is made in affect theory between affects, drives, and interpersonal
needs. Kaufman defines an interpersonal need as “a filndamental patterning of
interpersonal interaction that becomes amplified by and fused with affect, positive or
negative, and that becomes stored within the self as an interpersonal scene” (1989, p. 66).
Kaufman (1989) has identified seven innate primary interpersonal needs: the need for
relationship, the need for touching/holding, the need for identification, the need for
differentiation, the need to nurture, the need for affirrnation, and the need for power or
control over one’s life.

Combinations of affects, along with the stimuli or situations that trigger them
(represented by imagery and/or language), are stored in memory as scenes according to
Tomkins. The more intense the affect attached to the original object or event, the more
memorable and powerful the scene. Because negative affects often overwhelm positive

affects, negative affect scenes are usually more powerful and enduring than positive ones.

160

Scenes involving similar affects can become interconnected and magnified, thereby
increasing their psychological “footprint” and exerting a greater influence over one’s
psychic life.

Scenes in turn generate scripts, which Tomkins defines as rules for action and
cognition that allow one to predict, interpret, control, and respond to specific scenes
(Kaufman & Raphael, 1996). The affect theory concept of script appears similar to the
notion of coping strategies found in the work of Lazarus & Folkman. Also like Lazarus
& Folkman, the affect theory paradigm is a dynamic, process-oriented approach. The
scripts that emerge from particular scenes subsequently help to determine (via‘approach
or avoidance behaviors, for example) the likelihood that other scenes will be experienced.
For instance, if sexuality becomes fused with shame, one might avoid sexual contact and
intimacy, thereby avoiding the trauma of further sex-shame scenes, but at the same time
be deprived of beneficial love-intimacy scenes.

Scenes eventually disappear from full consciousness until they are later
reactivated by a similar scene, or by language that evokes the old scene. When scenes are
reactivated, a substantial portion of the original scene may remain inaccessible to
conscious experience. In these instances, one may be aware only of the affect associated
with the old scene, or of the imagery, or of the language. As long as the scene in its
entirety (i.e., combining affect, imagery, and language) remains inaccessible, it resists

control or modification, and one is prone to relive the scene again and again.

 

 

 

161

The Affect Of Shame
Tomkins defines shame as “an innate affect auxiliary response and a specific
inhibitor of continuing interest and enjoyment....The innate activator of shame is the
incomplete reduction of interest or joy....” (Tomkins, 1995, p. 84, emphasis in original).
Thus, according to Tomkins, shame is evoked whenever one experiences a partial
reduction in the positive affects of interest—excitement or enjoyment—joy. In other
words, shame is triggered when there is a breach or interruption in the self’ 5 experience
of positive affect vis-a-vis self or'other, but only when the breach or interruption is
incomplete, leaving vestiges of the positive affect intact. This creates a sense of longing
for reinstatement or reparation of the relationship or situation that existed prior to the
breach (Kaufman, 1989). If the positive affect is completely reduced—if there is no
longing for a continuation or reinstatement of the previous positive affective state—then
shame will not be experienced. There needs to be an ongoing investment in the
relationship in order for shame to be evoked. Tomkins, in characteristic fashion,
describes this idea beautifully and poetically:
“The experience of shame is inevitable for any human being insofar as desire
outruns fulfillment sufliciently to attenuate interest without destroying it. ‘I want,
but...’ is the essential condition for the activation of sharne....Insofar as human
beings are excited by or enjoy their work, other human beings, their bodies,
selves, and the surrounding inanimate world, they are vulnerable to a variety of
vicissitudes in the form of barriers, lacks, losses, accidents, imperfections,

conflicts, and ambiguities that will impoverish, attenuate, impair, or otherwise

 

 

162

prevent total pursuit and enjoyment of work, of others, of sexuality and other

drive satisfactions, and of the surrounding physical and social world.” (Tomkins,

1995, p. 406, emphasis added).

This idea can very readily be applied to interpersonal relationships and affective
bonds. Kaufman (1980; 1989) refers to the affective bond between two people as the
“interpersonal bridge.” Whenever this bridge is severed or damaged, whether due to a
fight, betrayal, or other failure, strong negative emotions ensue, including shame. The

same occurs if a person one cares about offers criticism, insult, blame, or belittlement

 

(Kaufman & Raphael, 1996). The deeply disturbing feeling of shame strongly motivates
one to restore or repair the breach to the interpersonal bridge. Thus shame functions to
motivate the reparation and maintenance of interpersonal bonds.
Shame Vs. Guilt
Unlike Lewis, Tomkins does not differentiate shame and guilt as different affects
per se. Rather, they represent different “affect complexes” consisting of shame as the
core affect, but with differing “causes and consequences.” The same may be said of
'shyness, inferiority, embarrassment, etc. All are considered by Tomkins to be variants of
the shame experience.
“Shyness, shame, and guilt are identical as affects, though not so experienced
because of differential coassembly of perceived causes and consequences.
Shyness is about strangeness of the other; guilt is about moral transgression;
shame is about inferiority; but the core affect in all three is identical...” (Tomkins,

1995,p.85)

 

 

163

Kaufman clarifies this point in his discussion of the variants of shame.

“Variants of shame become manifest in a broad range of interpersonal

contexts....Because of the differential coassembly of perceived causes and

consequences, including perceptions, cognitions, and intentions, these [inner]

states are actually experienced, overall, as quite different. Yet their core affect is

identical” (1989, p. 22).
That core affect is shame. Discouragement, for example, is shame about temporary
defeat. Self-consciousness is shame about being seen, being scrutinized, in the context of
performance. Embarrassment is social shame, shame before an audience. Shyness is
shame in the presence of strangers. Guilt is shame over moral transgression. And shame
itself, in its “purest” sense, involves a loss of face, dishonor, a sense of being small,
diminished, found wanting. Shame can be felt in the presence of another, or in the
presence of the self alone. The self can and does shame the self (Kaufman, 1989).
Phenomenology Of Shame

“Shame....is an affect auxiliary to the affect of interest—excitement. Any
perceived barrier to positive affect with the other will evoke lowering the eyelids and loss
of tonus in the face and neck muscles, producing a head hung in shame” (Tomkins, 1995,
p. 85). This image of shame as “loss of face” is similar to the phenomenology of shame
described by Lewis (1971). Both affect theory and psychoanalytic theory view shame as a
deeply disturbing experience in which one feels profoundly defective, diminished,
inferior. These feelings are accompanied by hanging the head and lowering the eyes,

along with increased autonomic activity and heightened self-awareness. Kaufman (1989)

 

 

 

164

describes shame as feeling “seen in a painfully diminished sense” (p. 17). ‘It is the
experience of being exposed as inferior, unworthy, lacking. Shame is experienced as an
interruption of the smooth functioning of the self, and as a breach in the interpersonal
bond between self and other. Shame thus divides the self from the self and the self from
the other, producing an intensely alienating experience. Shame is so disturbing precisely
because it is so central to the self and to the formation of identity and self-concept, as
well as to interpersonal relationships. It is also disturbing because of the ambivalence it

engenders toward the perceived source of that shame. There exists a longing for a

 

reestablishment of the interpersonal bridge between self and other, a yearning to heal of
the rift within the self (Kaufman, 1989), while simultaneously there may be rage and
hatred experienced toward the shaming other as well as the shamefirl self. The emphasis
on the significance of shame to interpersonal affective bonds is another similarity
between Lewis’s psychoanalytic approach and Tomkins’s affect theory approach to
shame. In an evolutionary sense, shame may be seen as adaptive because it motivates
individuals to-maintain and repair interpersonal relationships, which contribute to
survival. I
Shame And Mental Health

Shame, though always disturbing, is not necessarily harmful in and of itself, and
in fact plays the important role of motivating one to reestablish a positive emotional
connection with the source of one’s shame.

However, shame, like all affects, can be “magnified in frequency, duration, and

intensity” according to Tomkins (1995), and thereby becomes psychologically harmful or

 

165

“malignant.” The shame experience may also be made malignant by combining it with
other affects such as distress; when if prolonged, that combination of affects results in
depression. Shame also evokes additional emotional responses such as distress or anger,
and even shame about shame. Shameymay also recruit disgust or dissmell about aspects
of the self, which then can result in total rejection, and splitting off of those parts of the
self.
“Magnification of shame occurs not only by combining multiple affects about the
same scene, as when a rape victim may experience not only shame, but disgust,
dissmell, anger, and distress as well as terror, but also by combining multiple
sources of shame about the same scene. Thus an impotent failure of sexuality
may generate multiple feelings of deep shame suchthat the individual not only
feels the shame of sexual inferiority, but of a totally inferior self, along with
shyness, along with discouragement, guilt, defeat, and alienation. These are all
the same affect of shame but to different aspects of the same scene” (Tomkins,
1995, pg. 403, emphasis in original).
This process of psychological magnification or interconnection of shame scenes, in
combination with the internalization of shame through shame binds, results in a self that
is profoundly and malignantly bound by shame (Kaufman, 1989).
Intemalized shame scenes develop around affects, drives, and interpersonal needs.
Affective expression (positive or negative), drive expression, and expression of
interpersonal needs can all be shamed. When the shaming is sufficient in frequency,

duration, and intensity, then scenes surrounding these affects, drives, or interpersonal

 

166

needs become imprinted with shame. Kaufman refers to such internalized linkages as
binds (1989, p. 61), and differentiates affect-shame binds, drive-shame binds, and
interpersonal need-shame binds. These constitute three general classes of scenes that can
become infused with shame and thereby profoundly hinder the self s functioning by
impeding the expression of affect, gratification of drives, and satisfaction of needs
(Kaufman, 1989).

All internalized shame scenes consist of three interrelated components, according
to Kaufinan (1989). First, each scene has an aflect-belief component. This refers to a
cognitive label or self-appraisal that originated in an earlier scene and was infused with
shame affect (e.g., “You’re stupid” or “Hey, fatso”). This appraisal is reexperienced
along with recurring shame whenever the earlier scene is reactivated. In the original
scene the shame affect may be fused with a verbal message (e.g., being called “stupid,”
“fatso,” or “lazy”) and with imagery such as a disgusted, contemptuous, or mocking face.
Once the scene is internalized, the visual aspects fade, but the shaming “voice” remains
and typically is reexperienced again'and again.

Second, every scene contains an image of interaction patterns that mirror the
actual interpersonal relations that the person encountered within his family. These
images are more complex than simple cognitive self-appraisals; they involve repeated
patterns of interaction between the person and a significant other that become fused with
affect and language. Kaufman uses the example of a child who is repeatedly blamed for

errors in judgment and eventually internalizes the image of a blaming parent; once this

 

167

image is internalized it is likely that he will grow up learning to blame himself whenever
things go wrong.

Third, each scene includes an identification image or “internalized other” usually
based on one or both parents. This identification image results from a coalescence of
various scenes. It is originally experienced as affectively chargedimagery (e.g., the
shocked or disappointed face of the parent) combined with an auditory voice
admonishing the self. However, as with the affect-belief component, the visual aspect
fades over time while the shaming voice remains conscious.

Kaufman (1989) posits that shame in combination with other negative affects will
' lead to a variety of shame-based syndromes. For the purposes of the present study, the
relationship of shame to depressive syndromes is most relevant. Both Kaufman and
Tomkins view depression as a prolonged state of shame combined with distress. Unlike
traditional views that consider depression to be a direct result of inwardly-directed anger,
Kaufman suggests that while this dynamic does exist, it is secondary to the more
important shame-distress dynamic. Self-blame and self-contempt identity scripts are also
common according to Kaufman. An identity script is a higher-order script that is built up
from numerous lower-order scripts, all of which originate in scenes. Identity is defined
. by Kaufman as “the conscious experience of [the] self together with the active, living
relationship the self comes to have with the self” (1989, p. 104). According to Kaufman,
self-blame identity scripts recruit anger and direct it at the self in an accusatory manner.
These scripts develop within blame-oriented families that are more concerned with

determining fault than with reparation of mistakes; In the self-blame script, the self is

 

 

 

168

repeatedly accused by the self for real or imagined transgressions, with the blaming
accompanied by angry denouncement and humiliation of the self. The self-contempt
identity script, on the other hand, actually rejects the self entirely. Contempt combines
dissmell and anger, two innate affects, to produce a violent, rejecting affect blend which
in its most extreme form is akin to a lynching. Contempt manifests through hypercritical
attitudes directed toward others as well as the self. When directed against the self,
contempt creates a split within the self: one part becomes the offending object of
contempt while the other becomes brutally critical and punishing. Self-blame and self-
contempt scripts reactivate and intensify powerful shame-distress scenes and thereby
exacerbate depression.
Comparison Of The Psychoanalytic And Affect Theory Perspectives

The psychoanalytic and affect theory perspectives on shame are similar in many
respects. Both view shame as a critical variable for understanding the organization of the
self. Both consider shame to be central to certain psychopathological syndromes. The
two perspectives are also similar in their description of the phenomenology of shame.
'Finally, both perspectives share the view that shame functions as a signal that critical
interpersonal bonds have been broken or compromised, and as a motivating factor in the
restoration of those bonds.

There are, however, important differences in these two perspectives. Lewis’s
conceptualization of shame, rooted in psychoanalytic language, does not differentiate the
affects from the drives, and considers the drive system to be the primary motivator of

behavior. In this conceptualization, shame is a superego function tied to the control of

 

169

drives. In contrast, the affect theory perspective considers the affect system to be distinct
from, and primary over, the drive system. According to this schema, the drives require
fusion with affect in order to function. The affect system is viewed as the universal,
innate biological motivating mechanism which gives urgency to both drives and
interpersonal needs.

Another important difference between Lewis’s and Tomkins’s perspectives is the
relationship between shame and guilt. According to Lewis, guilt and shame are different
affects, tied to different facets of the superego. Tomkins, in contrast, considers guilt to be
merely one of several variants of shame—it is immorality shame, shame felt over moral
transgression. Each of the variants of shame (guilt, embarrassment, shyness, etc.) is
phenomenologically distinct because of differing causes and consequences, but the core
affect in each is shame.

This distinction between the psychoanalytic perspective on shame and guilt,
which views them as distinct affects tied to different superego fimctions, and affect
theory, which views guilt as a variant of shame, is fundamental. In the psychoanalytic
approach, shame and guilt arise from the superego, which itself emerges to restrain and
channel the drives. Hence, the drives are primary. In affect theory, the affect of shame is
innate, and experiencing shame in connection with specific attempts at drive satisfaction
or interpersonal interaction, or in combination with other affects, leads to distinct -
affective states such as guilt or embarrassment. In this schema, affect is primary.

The two perspectives also differ in their conceptualization of the dynamics by

which shame influences psychological functioning and mental health. Lewis posits that

 

170

shame is difficult to recognize and therefore to discharge, causing shame affect to remain
unacknowledged and in some cases not be experienced at all (“bypassed shame”). This
unacknowledged or bypassed shame is problematic because the rage (“humiliated firry”)
which invariably accompanies shame is “strangulated,” leading to a “shame-rage” spiral
and symptom formation. Tomkins, as well as Kaufman, agrees that shame affect is often
difficult to recognize as such (largely due to shame about shame itself), and that this
failure to recognize or acknowledge shame allows it to have a continuing influence over
psychological existence. However, rather than positing an inevitable shame-rage
sequence, affect theory assumes a more flexible approach rooted in the concept of
internalized “scenes.” Shame-anger sequences or scenes can and do exist, but shame can
be bound to any affect, positive or negative, as well as to drives and interpersonal needs.
Shame can also be followed by any affect, not just anger. This conceptualization allows
for more precise predictions about the types of psychopathological symptoms and
syndromes that will result from varying combinations of shame with interpersonal needs,
drives, and other affects.

Affect theory has been selected’as the theoretical viewpoint which can best guide
the present investigation for several reasons. First, as Kaufman (1989) argues, affect
theory differentiates the affect system from the drive system, and thereby provides a more
precise language with which to describe and predict emotional experience. Secondly,
Tomkins and Kaufman argue persuasively for the primacy of the affect system over the

drive system. “It is affect that gives texture to experience, urgency to drives, satisfaction

to relationships, and motivating power to purposes envisioned in the future. The affect

 

171

system and drive system are distinct, interrelated motivators. They empower and direct
both behavior and personality, but the drives must borrow their power from affect”
(Kaufman, 1989, p. 61). Affect theory subsumes and integrates classical psychoanalytic
theory and object relations theory within a general theory of the emotions, one in which
innate drives and relationships with both “real” and internalized objects derive their
power over behavior and personality through their fusion with affect. Thirdly, from a
practical standpoint, affect theory, unlike Lewis’s approach, has led directly to the
development of standardized instruments for measuring internalized shame. Hence,
rather than attempting to infer shame from clinical material as Lewis has, instruments
such as the Intemalized Shame Scale (ISS; Cook, 1994) allow for more precise
investigations of the impact of internalized shame on psychological health. Affect theory
may therefore prove to be a more useful perspective through which to advance the

empirical investigation of shame.

 

REFERENCES

REFERENCES

Aldwin, C. (1991). Does age affect the stress and coping process? Implications of
age differences in perceived control. Journal of Gerontology, 46, 174-180.

Atkinson, J. H., Grant, 1., Kennedy, C. J ., Richman, D. D., Spector, S. A., &
McCutchan, J. A. (1988). Prevalence of psychiatric disorders among men infected with
human immunodeficiency virus. Archives of General Psthiatrv. 4; 859-864.

Barret, R. L. (1989). Counseling gay men with AIDS: Human dimensions. Journal
of Counseling; and Development, 61 573-575.

Batchelor, W. F. (1984). AIDS: A public health and psychological emergency.
American Psychologist, 39, 1279-1284.

Billings, A., & Moos, R. (1981). The role of coping responses and social
resources in attenuating the impact of stressful life events. Journal of Behavioral
Medicine, 4, 139-157.

Bradley, L. N. (1991). The place of psychotherapy in the treatment of people who
are HIV positive. British Journal of Psychotherapy. 8, 26-35.

Bungener, C., Kosmadakis, C. S., Jouvent, R., & Widlocher, D. (1993).
Emotional and psychopathological disturbances in HIV-infection. Progress in Neuro-
Psychophamacologymd Biological Psychiatry, 17, 927-937.

Carr, E. W. (1989). Psychosocial issues of AIDS patients in hospice: Case studies.
Hospice Journal. 5. 135-151.

Catalan, J ., Klimes, 1., Day, A., Garrod, A., Bond, A., & Gallwey, J. (1992). The
psychosocial impact of HIV infection in gay men: A controlled investigation and factors
associated with psychiatric morbidity. British Journal of Psychiafl, 161, 774-778.

Cazzullo, C. L., Gala, C., Martini, S., Pergami, A., Rossini, M., & Russo, R.
(1990). Psychopathologic features among drug addicts and homosexuals with HIV
infection. International Journal of Psychiatry in Medicine, 20, 285-292.

Centers for Disease Control and Prevention (1997). HIV/AIDS Surveillance
Rgport, 9(1).

173

 

174

Chuang, H. T., Jason, G. W., Pajurkova, E. M., & Gill, M. J. (1992). Psychiatric
morbidity in patients with HIV infection. Canadian Journal of Psychiatry, 37, 109-115.

Chung, J. Y., & Magraw, M. M. (1992). A group approach to psychosocial issues
faced by HIV-positive women. Hospital and Community Psychiatry. 43, 891-894.

Cohen, J. (1992). A power primer. Psychological Bulletin. 11; 155-159.

Conwell, Y. (1994). Suicide and terminal illness: Lessons from the HIV '
pandemic. Crisis 15 57-58, 64. fi

 

Cook, D. R. (1994). Intemalized Shame Scale: Professional manual. Menomonie,
WI: Channel Press.

DeMarco, F., Ostrow, D., DiFranceisco, W., & Halrnan, J. (1997). General and
AIDS-specific stress, coping, and mental health within the Coping & Change cohort of
gay men. Paper presented at the meeting of the Midwestern Psychological Association, L
Chicago, IL, May 1997.

 

Derogatis, L., & Melisaratos, N. (1983). Brief Symptom Inventory (BSI): An
introductory report. Psychological Medicine. 13. 595-605.

Dew, M. A., Ragni, M. V., & Nimorwicz, P. (1990). Infection with human
immunodeficiency virus and vulnerability to psychiatric distress: A study of men with
hemophilia. Archives of General Psychiatry, 47, 737-744.

Dilley, J. W., Ochitill, H. N., Perl, M., & Volberding, P. A. (1985). Findings in
psychiatric consultations with patients with acquired immune deficiency syndrome.
American Journal of Psychiatry, 142, 82-86:

Folkman, S., Chesney, M. A., Cooke, M., Boccellari, A., & Collette, L. (1994).
Caregiver burden in HIV+ and HIV- partners of men with AIDS. J om] of Consultig
and Clinical Psychology. 62. 746-756.

Folkman, S., Chesney, M., Pollack, L., & Coates, T. (1993). Stress, control,
coping and depressive mood in human immunodeficiency virus-positive and -negative
gay men in San Francisco. Journal of Nervous and Mental Disease, 181. 409-416. ,

Folkman, S., & Lazarus, R. S. (1985). If it changes it must be a process: Study of
emotion and coping during three stages of a college examination. J oumaL of Personam
and Social Psychology, 48, 150-170.

Folkman, S., & Lazarus, R. S. (1988). The ways of copingquestionnaire. Palo
Alto, CA: Consulting Psychologists Press.

175

Folkman, S., Lazarus, R. S., Pimley, S., & Novacek, J. (1987). Age differences in
stress and coping processes. Psychology and Ag’ng, 2, 171-184.

Forsythe, C. J. & Compas, B. E. (1987). Interaction of cognitive appraisals of
stressful events and coping: Testing the goodness of fit hypothesis. Cognitive Therapy
and Research, 11, 473-485.

Gala, C. Pergami, A. Catalan, J ., & Riccio, M. (1992). Risk of deliberate self-
harm and factors associated with suicidal behaviour among asymptomatic individuals-
with human immunodeficiency virus infection. Acta Psychiatrica Scandinavica, 8Q 70-
. 75.

Gala, C., Pergami, A., Catalan, J ., Durbano, F., Musicco, M., Riccio, M.,
Baldeweg, T., & Invernizzi, G. (1993). The psychosocial impact of HIV infection in gay
men, drug users and heterosexuals. British Journal of Psychiatry, 163, 651-659.

Gorman, J. M., Kertzner, R., Todak, G., Goetz, R. R., Williams, J. B. W., Rabkin,
J ., Meyer-Bahlburg, H. F. L., Mayeux, K, Stern, Y., Lange, M., Dobkin, J ., Spitzer, R., &
Ehrhardt, A. A. (1991). Multidisciplinary baseline assessment of homosexual men with
and without human immunodeficiency virus infection: I. Overview of study design.
Archives of General Psychiatry. 4& 120-123.

Hays, R. B., Turner, H., & Coates, T. J. (1992). Social support, AIDS-related
symptoms, and depression among gay men. Journal of Consulting and Clinical
Psychology. 60. 463-469.

Herek, G., & Glunt, E. (1988). An epidemic of stigma: Public reactions to AIDS.
Special issue: Psychology and AIDS. American Psychologist, 43, 886-891.

Hestad, K., Aukrust, P., Ellersten, B., & Klove, H. (1994). Psychological
difficulties related to human immunodeficiency virus-1 infection in intravenous drug
users.‘Acta Psychiatrica Scandinavica 90 25-31.

 

Hunter, J ., & Hamilton, M. (1992). CFA: A program in BASICA [computer
software].

Hunter, J ., & Hamilton, M. (1995). Path: A program in BASICA [computer
software].

Hunter, J ., & Schmidt, F. (1990). Methods of meta-analysis: Correctingerror and
bias in research findings. Newbury Park: Sage.

J acobsen, P. B., Perry, S. W., & Hirsch, D. A.(1990). Behavioral and
psychological responses to HIV antibody testing. Special Series: Acquired Immune
Deficiency. Journal of Consulting and Clinical Psychology, 58, 31-37.

 

 

176

James, M. E., Rubin, C. P., & Willis, S. E. (1991). Drug abuse and'psychiatric
findings in HIV -seropositive pregnant patients. Generzi Hospital Psychiatry. 13. 4-8.

Johannet, C., & Muskin, P. R. (1990). Mood and behavioral disturbances in
hospitalized AIDS patients. Psychosomatics. 31, 55-59.

Jones, J. M. (1988). Scientific lssues: Sectionintroduction. Special lssue:
Psychology and AIDS. American Psychologist 43,899.

Kaufirlan, G. (1989). The psychology of shame. New York: Springer. fl

Kaufman, G. (1992). Shame: The power of caring (3rd ed.). Rochester, VT:
Schenkrnan Books.

Kaufman, G., & Raphael, L. (1996). Coming out of shame: Transforming gay and
lesbian lives. New York: Doubleday.

 

Kessler, R. C. (1991). Comorbidity of substance use disorders and other
psychiatric disorders: An ADAMHA report on prevalence, etiology, implications for
prevention, course of illness and research. Unpublished manuscript, University of
Michigan, Ann Arbor.

Krikorian, R, Kay, J. & Liang, W. M. (1995). Emotional distress, coping, and
adj ustment in human immunodeficiency virus infection and acquired immune deficiency
syndrome. Journal of Nervous and Mental Disease 183 293- 298.

Lackner, J. B., Joseph, J. G., Ostrow, D. G., Kessler, R. C., Eshleman, S.,
Wortman, C. B., O'Brien, K., Phair, J. P., & Chmiel, J. (1993). A longitudinal study of
psychological distress in a cohort of gay men: Effects of social support and coping
strategies. Journal of Nervous and Mental Disease, 18; 4-12.

Lazarus, R. S., & Folkman, S. (.1984). Stress, appraisal, and coping. New York:
Springer.

Levenson, A. I. (1988). Psychiatric aspects of AIDS. 55th annual meeting of the
National Association of Private Psychiatric Hospitals (Phoenix, Arizona). Psychiatric

Hospital, 19, 109-113.

Lewis, H. B. (1971). Shame and gilt in neurosis. New York: International
Universities Press.

Lewis, H. B. (1987). Introduction: Shame—the “sleeper” in psychopathology. In
H. B. Lewis (Ed.), Role of shame in symptom formation. Hillsdale, NJ: Lawrence
Erlbaum Associates.

 

177

Lippmann, S. B., James, W. A., & Frierson, R. L. (1993). AIDS and the family:
Implications for counseling. Special Issue: The family and HIV disease. AIDS CareLS.
71-78.

Lipsitz, J. D., Williams, J. B. W., Rabkin, J. G., Remien, R. H., Bradbury, M., e1
Sadr, W., Goetz, R., Sorrell, S., & Gorman, J. M. (1994).'Psychopathology in male and
female intravenous drug users with and without HIV infection. Amerifl J ournal_of
Psychiatry. 151. 1662-1668. '

Lynd, H. M. (1958). On shame and the search for identity. New York: Harcourt,
Brace.

Martin, M., & Henry-Feeney, J. (1989). Clinical services to persons with AIDS:
The parallel nature of the client and worker processes. Clinical Social WorLlournal. 17.
337-349.

McKegney, F. P., & O’Dowd, M. A. (1992). Suicidality and HIV status. VI
International Conference on AIDS (1990, San Francisco, California). American Journal
of Psychiatry, 149, 396-398.

Michels, R., & Marzuk, P. M. (1993). Progress in psychiatry: II. New England
Journal of Medicine. 329, 628-638.

Michigan Department of Public Health (1995). Statistics. Michigan HIV Report.
m, 8-9.

Millan, F ., & Ivory, L. I. (1994). Group therapy with the multiply oppressed:
Treating Latino, HIV-infected injecting drug users. Special Issue: The challenge of AIDS.
Group, 18, 154-166.

Miller, D., & Riccio, M. (1990). Non-organic psychiatric and psychosocial
syndromes associated with HIV-1 infection and disease. AIDS 4 381-388.

 

Morin, S. F. (1988). AIDS: The challenge to psychology. Special Issue:
Psychology and AIDS. American Psychologist, 43, 838-842.

Morin, S. E, Charles, K. A., Malyon, A. K. (1984). The psychological impact of
AIDS on gay men. American Psychologist, 39, 1288-1293.

Narnir, S., Wolcott, D. L., Fawzy, F. I., & Alumbaugh, M. J. (1987). Coping with
AIDS: Psychological and health implications. Special Issue: Acquired Immune
Deficiency Syndrome (AIDS). Journal of Applied Social Psychology, 17, 309-328.

Nathanson, D. L. (1994). Shame and pride: Affect, sex, and the birth Of the self.
New York: Norton.

 

 

178

Neisen, J. H. (1990). Heterosexism: Redefining homophobia for the 19905.
Journal of Gay & Lesbgn Psychotherapy, 1, 21-35.

Nichols, S. E. (1983). Psychiatric aspects of AIDS. Psychosomatics. 2i 1083-
1089.

Odets, W. (1995). Introduction. In W. Odets & M. Shernoff (Eds.), The second
decade of AIDS: A mental heflh practice pandbocfi (pp. 1-10). New 'York: Hatherleigh.

O'Dowd, M. A., & McKegney, F. P. (1990). AIDS patients compared with others
seen in psychiatric consultation. General Hospital Psychiatry. 12. 50-55.

O'Dowd, M. A., Biderrnan, D. J ., & McKegney, F. P. (1993). Incidence of
suicidality in AIDS and HIV-positive patients attending a psychiatry outpatient program.
Psychosomatics. 34. 33-40.

O'Dowd, M. A., Natali, C., Orr, D., & McKegney, F. P. (1991). Characteristics of
patients attending an HIV-related psychiatric clinic. Hospital and Community 'Psychiafl,
fl, 615-619.

Ostrow, D. G., Joseph, J. G., Kessler, R., Soucy, J ., Tal, M., Eller, M., Chmiel, J .,
& Phair, J. P. (1989). Disclosure of HIV antibody status: Behavioral and mental health
correlates. AIDS Education and Prevention. 1. 1-11.

Ostrow, D. G., Monjan, A., Joseph, J ., VanRaden, M., Fox, R., Kingsley, L.,
Dudley, J ., & Phair, J. (1989). HIV -related symptoms and psychological firnctioning in a
cohort of homosexual men. American J oumal of Psychiatry, 146, 737-742.

Pakesch, G., Loimer, N., Grunberger, J ., Pfersmann, D., Linzmayer, L., &
Mayerhofer, S. (1992). Neuropsychological findings and psychiatric symptoms in HIV -1
infected and noninfected drug users. Psychiatry Researcha41, 163-177.

Parker, J. D., Endler, N. S., & Bagby, R. M. (1993). If it changes, it might be
unstable: Examining the factor structure of the Ways of Coping Questionnaire.
Psychological Assessment, 5, 361-368.

Perry, S., J acobsberg, L. B., F ishman, B., Frances, A., Bobo, J ., & J acobsberg, B.
K. (1990). Psychiatric diagnosis before serological testing for the human
immunodeficiency virus. American Journal of Echiaatry. 147. 89-93.

Perry, 8., Jacobsberg, L. B., Card, C. A., & Ashman, T. (1993). Severity of
psychiatric symptoms after HIV testing. American Journal of Psychiaaw. 150, 775-779.

Rabkin, R., & Rabkin, J. (1995). Management of depression in patients with HIV
infection. In W. Odets & M. Shernoff (Eds), The second decade of AIDS: A mental
health practice handboplg (pp. 11-25). New York: Hatherleigh.

179

Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for
research in the general population. Applied Psychologiaal Measurement. 1. 385-401.

Rajs, J ., & Fugelstad, A. (1992). Suicide related to hmnan immunodeficiency
virus infection in Stockholm. Acflsychiatrica Scandinavica, 8; 234-239.

Reed, G. M., Taylor, S. E., & Kemeny, M. E. (1993). Perceived control and
psychological adjustment in gay men with AIDS. J Gupta] of Applied 3ch Psychology,
23, 791-824.

Rosenberger, P. H., Bomstein, R. A., Nasrallah, H. A., Para, M. F., Whitaker, C.
C., Fass, R. J ., & Rice, R. R., Jr. (1993). Psychopathology in human immunodeficiency
virus infection: Lifetime and current assessment. Comprehensive Psychiatry, 34. 150-
158.

Sacks, M., Burton, W., Dermatis, H., Looser-Ott, S., & Perry, S. (1995). HIV -
related cases among 2,094 admissions to a psychiatric hospital. Psychiatric Services. 46.
131-135.

Schaffirer, B. (1994). Psychotherapy with HIV-infected persons. New Directions
for Mental Health Services, 48, 5-20.

Shernoff, M. (1990). Why every social worker should be challenged by AIDS.
Socgl Work, 35, 5-8.

Shernoff, M. (1995). Counseling chemically dependent clients with HIV infection.
In W. Odets & M. Shernoff (Eds), The second decade of AIDS: A mental health practice
handbook (pp. 27-46). New York: Hatherleigh.

Tangney, J. P. (Ed. ) (1995). Self-conscious emotions: The psychology of sham___e_,
gpilt, embarras sment, and pride. New York: Guilford Press.

Tangney, J. P. (1996). Conceptual and methodological issues in the assessment of
shame and guilt. Behayiour Research yd Therapy. 34. 741-754.

Tomkins, S. S. (1995). Role of the specific affects. In E. V. Demos (Ed.),
- Exploring affect: The selected writipgs of Siflan S. Tomkins (pp. 68-85). Cambridge,
UK: Cambridge University Press.

Tomkins, S. S. (1995). The varieties of shame and its magnification. In E. V.

Demos (Ed), Exploring affect: The selected writings of Silvan S. Tomkins (pp. 397-410).
Cambridge, UK: Cambridge University Press.

' Tross, S., & Hirsch, D. A. (1988). Psychological distress and neuropsychological
complications of HIV infection and AIDS. (Special Issue: Psychology and AIDS.)
American Psychologist, 43, 929- 934.

 

180

Volberding, P. A., & Cohen, P. T. (1990). Clinical spectrum of HIV infection. In
P. T. Cohen, M. A. Sande, & P. A. Volberding (Eds), The AIDS knowledge base: A
textbook on HIV disease from the University of California San Francisco. and the Sa_n
Francisco General Hospital (section 4.1.1). Waltham, MA: Medical Publishing Group.

 

Williams, J. B. W., Rabkin, J. G., Remien, R. H., Gorman, J. M., & Ehrhardt, A.
A. (1991). Multidisciplinary baseline aSsessment of homosexual men with and without
human immunodeficiency virus infection: 11. Standardized clinical assessment of current
and lifetime psychopathology. Archives of General Psychiatry. 48. 124-130. I

Wolf, T. M., Balson, P. M., Morse, E. V., Simon, P. M., Gaumer, R. H., Dralle, P.
W., & Williams, M. H. (1991). Relationship of coping style to affective state and
perceived social support in asymptomatic and symptomatic HIV-infected persons:
Implications for clinical management. Jouml of Clinical Psychiatry. 5;, 171-173.

World Health Organization (1997). The Current Global Situation of the
HIV/AIDS Pandemic as of 15 December 1995. Joint United Nations Programme on
HIV/AIDS [On-line]. Available: http://www.us.unaids.org/highband/document/epidemio/
report97.html.

Wurrnser, L. (1981). The mask of shame. Baltimore, MD: Johns Hopkins
University Press.

 

"‘lllllllllllllllESm