Vous êtes sur la page 1sur 9

Developmental Psychology Copyright 1995 by the American Psychological Association, Inc.

1995, Vol. 31, No. 1,86-94


0OI2-1649/95/S3.0O

Developmental Changes in Relationship Quality in Gay and Lesbian


Cohabiting Couples
Lawrence A. Kurdek
Wright State University

Changes over 3 annual assessments were examined for both partners of 61 gay and 42 lesbian couples
on current levels of attachment, autonomy, and equality in the relationship; the importance of at-
tachment, autonomy, and equality in an ideal relationship; and relationship commitment. Lesbian
partners rated ideal equality as more important than did gay partners; for both gay and lesbian
partners, the importance of ideal attachment decreased over time; and for lesbian partners only, the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

importance of ideal equality changed curvilinearly over time. For both gay and lesbian partners,
This document is copyrighted by the American Psychological Association or one of its allied publishers.

changes in relationship commitment over time were explained by changes in the discrepancy be-
tween current and ideal levels of equality as well as between current levels of attachment and current
levels of autonomy.

One of the major developmental tasks facing adults is the ne- the pulls and pushes of a relationship are anchored to personal
gotiation of issues regarding intimacy (Erikson, 1963; Havig- needs in the domains of dyadic attachment, personal autonomy,
hurst, 1953). This article addressed one aspect of intimacy and equality. Specifically, it is proposed that changes in relation-
relationship commitmentin gay and lesbian couples. Al- ship commitment are linked to changes in (a) the discrepancy
though there is a growing body of work identifying factors between the importance attributed to dyadic attachment, per-
linked to relationship satisfaction and relationship stability for sonal autonomy, and equality in an ideal relationship and the
gay men and lesbians (see Kurdek, 1994, for a review), the study actual levels of dyadic attachment, personal autonomy, and
of the processes that regulate relationship commitment has equality perceived in the current relationship and (b) the rela-
been relatively neglected in this literature. This is somewhat sur- tive balance of currently experienced levels of dyadic attach-
prising in light of evidence that many lesbians and gay men are ment and personal autonomy. These propositions are addressed
in long-lasting cohabiting relationships (Blumstein & Schwartz, in turn.
1983; Bryant & Demian, 1994). Cochran and Peplau (1985) identified attachment, auton-
Although it is recognized that commitment is influenced by omy, and equality as core dimensions of relationship quality.
the larger social context in which relationships develop (Stanley Attachment refers to the value placed on having a close and
& Markman, 1992), the focus of this study was on the psycho- relatively secure love relationship with one person. Autonomy
logical context of commitment (Sternberg, 1988). Because gay concerns the boundaries between oneself as an individual and
and lesbian couples develop without institutional support and oneself as a partner in a relationship. Finally, equality deals with
often without the support and even knowledge of family mem- the extent to which partners share power in and responsibility
bers (Kurdek, 1988; Kurdek & Schmitt, 1987), the health and for the relationship.
stability of these relationships may be primarily tied to the level The importance attributed to attachment, autonomy, and
of dependence partners have on each other. Thus, gay and les- equality in an ideal relationship sets the comparison level for
bian couples provide a unique testing ground for theories of re- currently experienced levels of attachment, autonomy, and
lationship functioning based on notions of partner equality and exerts a direct effect on satisfaction with the rela-
interdependence. tionship (Baucom & Epstein, 1990; Rusbult, 1983). Although
Relationship commitment is described in interdependence Kurdek (1991) found that the overall discrepancy between the
theory (Kelley, 1983) in terms of a causal system that over time importance attributed to attachment, autonomy, and equality
and across situations supports membership in the relationship. in the ideal relationship and the actual levels experienced in
This system includes processes that either keep a person in the these domains was negatively related to relationship satisfac-
relationship ("pulls") or draw a person out of the relationship tion, the relation between discrepancies in each domain and re-
("pushes"; Duffy & Rusbult, 1986). The present study extends lationship commitment was not examined.
previous theoretical discussions of commitment by arguing that The contrast between currently perceived levels of attach-
ment and currently perceived levels of autonomy is important
in light of the view that the fundamental task in any close rela-
I thank the couples who participated in this study and Mark Fine for tionship is to balance attachment and autonomy (Raush, 1977).
comments on a draft of the manuscript. Although interdependence theory (Kelley, 1983) would predict
Correspondence concerning this article should be addressed to Law- that decreases in relationship commitment would occur when
rence A. Kurdek, Department of Psychology, Wright State University, the pushes of the relationship (i.e., autonomy) exceed the pulls
Dayton, Ohio 45435-0001. Electronic mail may be sent via Internet to of the relationship (i.e., attachment), no direct empirical exam-
LKURDEK@DES1RE. WR1GHT.EDU. inations of this proposition have been made.
86
SPECIAL ISSUE: GAY AND LESBIAN COUPLES 87

Most of the limited empirical work in the area of gay and Table 1
lesbian relationships was designed to account for variability in a Means and Standard Deviations for Time 1 Demographic
single assessment of relationship quality (e.g., Duffy & Rusbult, Scores for Partner 1 and Partner 2 in Gay and Lesbian
1986; Kurdek, 1991). However, because relationships them- Couples in the Longitudinal Sample
selves are developmental phenomena (McWhirter & Mattison, Lesbian
Gay
1984) and some relationships end in dissolution (Kurdek, (n = 42)
1992), relationship quality is likely to change over time.
In the present study, data regarding relationship commit- Score Partner 1 Partner 2 Partner 1 Partner 2
ment; the importance of attachment, autonomy, and equality
in the ideal relationship; and currently perceived levels of at- Age
M 42.90 42.11 41.26 40.76
tachment, autonomy, and equality were gathered from both SD 12.59 11.84 9.53 8.14
partners of gay and lesbian couples annually over a 3-year pe- Education
riod. It is important to note that the couples in this study had M 6.49 6.11 6.38 6.35
lived together at the first assessment an average of 10 years. SD 1.36 1.21 1.10 1.18
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Thus, changes in relationship commitmentparticularly de- Income


This document is copyrighted by the American Psychological Association or one of its allied publishers.

creasesthat occurred over the three annual assessments were M 7.91 7.36 8.14 7.97
SD 3.10 3.09 3.22 2.95
not likely to be due to normative changes in relationship func- Months living together
tioning experienced in the early phases of the relationship M 138.34 93.76
(McWhirter & Mattison, 1984) but instead were likely to be SD 107.08 64.68
due to idiosyncratic changes in the psychological processes that
regulate relationship functioning (Kurdek, 1992; Rusbult, Note. Maximum scores for education and personal income were 8 and
18, respectively.
1983).
Given the paucity of longitudinal data on gay and lesbian cou-
ples, the first purpose of this study was to provide descriptive
data on differences between gay and lesbian couples on develop- ticipants published in gay and lesbian periodicals and newsletters as well
mental changes in relationship commitment; in ideal levels of as through referrals of couples who had already participated in the
attachment, autonomy, and equality; and in currently perceived study. Partners were randomly assigned to a Partner 1 or a Partner 2
levels of attachment, autonomy, and equality. Because lesbian status.
partners have been noted to follow an ethic of equalityper- Data from each of three annual assessments were available for 61 gay
haps as a reaction to socialization influences that reduce het- couples and 42 lesbian couples. Most of these participants were White
(93%) and employed (83%). Attrition in sample size was due to separa-
erosexual women's equality in their relationships (Blumstein &
tion (4 gay couples and 3 lesbian couples), withdrawal from the study (4
Schwartz, 1983)it was expected that lesbian couples would gay couples and 3 lesbian couples), failure to return completed surveys
prize equality in their relationships more than would gay cou- (3 gay couples and 3 lesbian couples), and death of a partner (3 gay
ples. In the absence of previous longitudinal data on relatively couples).
durable relationships of gay and lesbian couples, no predictions To assess bias in the longitudinal sample of nonseparated couples, I
were advanced regarding the pattern of any longitudinal compared couples who had data for all three times of assessment (N =
changes or how type of couple (gay vs. lesbian) might moderate 103) with those who withdrew from the study or did not respond (N =
this pattern. 13) in a one-way (longitudinal vs. not longitudinal) multivariate analysis
Whereas the first purpose of this study focused on group of variance (MANOVA). The dependent variables were the three Time
trends, the second purpose of this study focused on individual 1 demographic scores for each partner (age, education, and income),
differences in changes in relationship commitment (cf. Belsky the Time 1 couple score (averaged over both partners) for number of
& Rovine, 1990). On the basis of the propositions advanced ear- months living together, and the seven Time 1 relationship-oriented
scores for each partner (relationship commitment; currently perceived
lier, it was predicted that changes in relationship commitment
levels of attachment, autonomy, and equality; and ideal levels of attach-
over the three annual assessments would be accounted for by ment, autonomy, and equality). The resulting effect was nonsignificant,
changes in the discrepancy between ideal and currently per- indicating that couples in the longitudinal sample were representative
ceived levels of attachment, autonomy, and equality as well as of those in the original Time 1 sample on these characteristics.
by changes in the discrepancy between currently perceived lev- For couples in the longitudinal sample, means and standard devia-
els of attachment and currently perceived levels of autonomy. tions for each partners' age, education, and income at Time 1 as well as
Specifically, decreases in commitment were expected to be the number of months the couple had lived together at Time 1 are pre-
linked to an increase in the extent to which the current relation- sented by type of couple in Table 1. A one-way (type of couple)
ship fell short of the ideal relationship as well as an increase in MANOVA yielded a nonsignificant effect, indicating that gay and les-
the salience of autonomy over attachment. Given that change in bian couples were equivalent on these demographic characteristics. Be-
commitment over three annual assessments could be examined, cause respondents tended to be well-educated (see Table 1), no claims
this hypothesis was examined for change in commitment from are made that this sample is representative of gay and lesbian couples.
Time 1 to Time 2 as well as that from Time 2 to Time 3.
Method Procedure
Participants At each annual assessment, each couple was sent two identical sur-
At Year 1, data were gathered from both partners of 75 gay and 51 veys. To promote honest responding, instructions in the surveys di-
lesbian cohabiting couples who were recruited through requests for par- rected partners to complete their surveys privately and not to discuss
LAWRENCE A. KURDEK

Table 2 After rating current levels of attachment, autonomy, and equality,


Cronbach Alphas for Summed Composite Scores at Each participants rated analogous items in terms of how important (1 = not
Assessment Timefor Each Partner important, 9 = very important) they were in an ideal relationship. These
items are also presented in the Appendix. The stability of these three
Partner 1 Partner 2 summed composite scores over the 3 years was moderate: For Partner
1, the Pearson correlations for Time 1 -Time 2 and Time 1 -Time 3 were
Time Time Time Time Time Time .79 and .75 for attachment, .85 and .80 for autonomy, and .48 and .66
Score 1 2 3 1 2 3 for equality, respectively (all ps < .01); for Partner 2, the Pearson corre-
Commitment .89 .93
lations for Time 1-Time 2 and Time 1-Time 3 were .75 and .73 for
.93 .89 .88 89
Current relationship attachment, .72 and .70 for autonomy, and .64 and .66 for equality,
Attachment .72 .76 .75 .71 .73 77 respectively (all ps < .01).
Autonomy .64 .67 .76 .63 .59 .66 For descriptive purposes, Pearson correlations among scores for rela-
Equality .86 .86 .91 .87 .86 .89 tionship commitment; current levels of attachment, autonomy, and
Ideal relationship equality; and ideal levels of attachment, autonomy, and equality are pre-
Attachment .78 .82 .77 .75 .79 84 sented for each of the 3 years by Partner 1 and Partner 2 in Table 3. In
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Autonomy .83 .81 .85 .82 .81 .84 support of the validity of these scores, most correlations were
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Equality .86 .86 .81 .90 .89 .88 significant.


Note. TV = 103.
Discrepancy Between Current Attachment and Current
Autonomy and Between Ideal and Current Levels of
Attachment, Autonomy, and Equality
their answers with each other until the forms had been completed and
returned in separate postage-paid envelopes. Two sets of discrepancy scores were calculated for each partner. In the
first set, the balance between perceived levels of attachment and auton-
Demographic Variables omy in the current relationship was assessed by subtracting the current
autonomy score from the current attachment score. A positive score
Participants provided information regarding age, race, education indicated that attachment exceeded autonomy, whereas a negative score
(represented by eight intervals ranging from completion of less than indicated that autonomy exceeded attachment.
seventh grade to the award of a doctorate), employment status, and an- In the second set, for each of the three domains, the rating given to
nual personal income (represented by 18 intervals ranging from $5,000 the current version of the item was subtracted from that given to the
or less to $80,000 or more). They also reported the number of months ideal version of the item. The comparison level for each domain was the
they had lived with their partner. When needed, couple scores were cal- sum of these eight differences. Positive scores for attachment, auton-
culated by averaging partners' scores. omy, and equality comparison levels indicated that the current relation-
ship fell short of the ideal relationship, whereas negative scores indi-
cated that the current relationship exceeded the ideal relationship.
Relationship Commitment For descriptive purposes, Pearson correlations among scores for rela-
Relationship commitment was assessed by eight items from Stern- tionship commitment; discrepancies between ideal and current levels
berg's (1988) triangular assessment of love scale. Respondents indicated of attachment, autonomy, and equality; and the discrepancy between
how true (1 = not at all true, 9 = extremely true) each item was. Actual currently perceived attachment and currently perceived autonomy are
items are presented in the Appendix. As shown in Table 2, Cronbach presented for each of the 3 years by Partner 1 and Partner 2 in Table 4.
alphas were acceptable at each time of assessment for each partner for In support of the validity of these scores, most correlations were
this score and for all other summed composite scores used in this study. significant.
The relationship commitment score showed moderate levels of stability
over the 3 years of study; the Pearson correlations for Time 1-Time 2 Results
and Time 1-Time 3 were .68 and .65, respectively, for Partner 1 and
were .61 and .46, respectively, for Partner 2 (all ps < .01). Type of Couple and Time-of-Assessment Effects
Of interest here was whether relationship commitment; cur-
Current and Ideal Levels of Attachment, Autonomy, rent levels of attachment, autonomy, and equality; and ideal lev-
and Equality els of attachment, autonomy, and equality (a) differed for gay
Current and ideal levels of attachment, autonomy, and equality were and lesbian partners, (b) changed over the three annual assess-
assessed by a measure designed for this study that was derived from ments, and (c) changed over this time period similarly for gay
measures developed by Cochran and Peplau (1985) and Peplau and and lesbian partners. Because partner scores were not indepen-
Cochran (1981). Participants first rated how true (1 = not at all true, 9 dent of each other, separate analyses were conducted for Partner
= very true) each of 24 items was of their current relationship. There 1 and Partner 2, providing a within-study replication. Relevant
were eight items apiece for the domains of attachment, autonomy, and means are presented in Table 5.
equality. These are presented in the Appendix. The stability of these
three summed composite scores over the 3 years was moderate: For For each partner, the seven relationship-oriented scores listed
Partner 1, the Pearson correlations for Time 1-Time 2 and Time 1- earlier were submitted to a 2 (type of couple: gay or lesbian) X 3
' Time 3 were .75 and .72 for attachment, .78 and .78 for autonomy, and (time) M ANO\A with repeated measures on the time-of-assess-
.62 and .65 for equality, respectively (all ps < .01); for Partner 2, the ment factor. Because interest in the time effect focused on pat-
Pearson correlations for Time 1-Time 2 and Time 1-Time 3 were .67 terns of change over the three assessments, time effect was de-
and .73 for attachment, .69 and .62 for autonomy, and .67 and .58 for composed into a linear polynomial trend and a curvilinear
equality, respectively (allps < .01). polynomial trend. Findings are presented by each partner.
SPECIAL ISSUE: GAY AND LESBIAN COUPLES 89

Table 3
Correlations Among Commitment; Current Levels of Attachment, Autonomy, and Equality;
and Ideal Levels of Attachment, Autonomy, and Equality at Each Time ofAssessment for
Partner 1 and Partner 2
Current Ideal

Score 1 2 3 4 5 6 7

Time 1
1. Commitment .48** .01 .60** .23** -.10 .20*
Current .50** .63** -.25** .16
2. Attachment .53** -.15
3. Autonomy -.02 -.28** .07 -.13 .56** .05
4. Equality .68** .48** .02 .23** -.13 .41**
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Ideal
This document is copyrighted by the American Psychological Association or one of its allied publishers.

5. Attachment .27** .73** -.33** .15 .06 .39**


6. Autonomy -.14 -.27** .64** -.14 -.17* .31**
7. Equality .28** .29** .04 .38** .45** .17*

Time 2
1. Commitment .56** -.15 .59** .41** -.22* .30**
Current .68** .18*
2. Attachment .63** .18* .51** -.28**
3. Autonomy -.13 -.22* -.10 -.10 .71** .13
4. Equality .70** .59** -.06 .30** -.26** .42**
Ideal
5. Attachment .35** .74** -.23** .28** -.02 .38**
6. Autonomy -.05 -.27** .63** -.10 -.16 .20*
7. Equality .39** .38** -.14 .51** .45** .02
Time 3
1. Commitment .62** -.21* .69** .30** -.26** .20*
Current
2. Attachment .64** -.27** .64** .72** -.31** .19*
3. Autonomy -.24** -.33** -.08 -.09 .73** .08
4. Equality .71** .63** -.12 .27" -.23** .25**
Ideal
5. Attachment .31** .64** -.36** .22** .03 .42**
6. Autonomy -.14 -.27** .69** -.18* -.22* .20*
7. Equality .26* .21* -.01 .28** .34** .08

Note. Correlations below the diagonal are for Partner 1 and those above the diagonal are for Partner 2.
*p<.05. **p<X)\.

Partner I. Significant multivariate effects were obtained effect on ideal equality was significant for only lesbian partners,
only for type of couple, the linear trend on time, and the Type F( 1,41) = 5.69 (eta2 = . 12), p < .05. As can be seen in Table 5,
of Couple X Curvilinear Trend on Time interaction, F(7,95) = ideal equality for this group conformed to a U-shaped curve
2.67, 2.42, and 3.47, respectively, ps < .05. Univariate analyses over the three assessments. That is, scores held constant at Time
indicated that the effect of type of couple was due to differences 1 and Time 3 but dipped at Time 2.
between gay and lesbian couples on commitment, ideal equality, Partner 2. As with Partner 1, significant multivariate effects
and current autonomy, F(l, 101) = 4.29 (eta2 = .04), 5.12 (eta2 were obtained only for type of couple, the linear trend on time,
= .04), and 4.50 (eta2 = .04), respectively, ps < .05. As seen in and the Type of Couple X Curvilinear Trend on Time interac-
Table 5, compared with gay partners, lesbian partners had tion, F{1,95) = 2.62,1.99, and 1.97, respectively,ps < .05. Uni-
higher scores on each variable. variate analyses indicated that the effect of type of couple was
Univariate analyses indicated that the multivariate linear due to differences between gay and lesbian couples on ideal au-
time effect was due only to ideal attachment, F(\, 101) = 6.13 tonomy and ideal equality, F{\, 101) = 4.60 (eta2 = .04) and
(eta2 = .05), p < .05. As seen in Table 5, ideal attachment de- 11.04 (eta2 = .09), respectively, ps < .05. As seen in Table 5,
clined over the 3 years of study. compared with gay partners, lesbian partners had higher scores
Finally, univariate analyses indicated that the multivariate on each variable.
Type of Couple X Curvilinear Trend on Time interaction was Univariate analyses indicated that the multivariate linear
due only to ideal equality, F(l, 101) = 11.26 (eta2 = .10), p < time effect was due only to ideal attachment and current attach-
.01. Tests of simple effects indicated that the curvilinear time ment, F{\, 101) = 7.24 (eta2 = .06) and 8.19 (eta2 = .07), re-
90 LAWRENCE A. KURDEK

Table 4
Correlations Among Commitment; Discrepancy Between Current Levels ofAttachment and
Current Levels of Autonomy; and Discrepancy Between Ideal and Current Levels of
Attachment, Autonomy, and Equality at Each Time of Assessment for Partner 1 and Partner 2
Score 1 2 3 4 5

Time 1
1. Commitment .30** -.25** -.13 -.34**
2. Current attachment-autonomy .33** -.23** .09 -.17*
Ideal-current
3. Attachment -.31** -.08 .41** .52**
4. Autonomy -.16* .11 .25** .49**
5. Equality -.37** -.12 .59** .34**
Time 2
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

1. Commitment _ .46** -.09 -.14 -.24**


2. Current attachment-autonomy .48** -.21** -.07 -.32**
Ideal-current
3. Attachment -.31** -.11 .35** .47**
4. Autonomy .08 .14 .21* .37"
5. Equality -.41** -.14 .54" .23**
Time 3
1. Commitment .53** -.28** -.15 -.43"
2. Current attachment-autonomy .53** -.19* -.07 -.33"
Ideal-current
3. Attachment -.42** -.21* .34** .62"
4. Autonomy .09 .18* .07 .37"
5. Equality -.53** -.35** .59** .17*

Note. Correlations below the diagonal are for Partner 1 and those above the diagonal are for Partner 2.
*p<.05. "/?<.O1.

spectively, ps < .01. As seen in Table 5, both scores decreased autonomy, and equality and (b) the discrepancy between cur-
over the 3 years of study. rently perceived levels of attachment and currently perceived
Finally, univariate analyses indicated that the multivariate levels of autonomy. To assess these linkages, I used two hierar-
Type of Couple X Curvilinear Trend on Time interaction was chical multiple regressions to assess regressed change in each
due only to ideal equality, F(l, 101) = 2.99 (eta2 = .03),p < .05. partner (see Cohen & Cohen, 1983, pp. 417-423). In the first
Tests of simple effects indicated that the curvilinear time effect regression, the key issue was whether change in relationship
on ideal equality was significant for only lesbian partners, F( 1, commitment from Time 1 to Time 2 was related to change in
41) = 5.34 (eta2 = .11), p < .05. As can be seen in Table 5, the set of discrepancy scores from Time 1 to Time 2. In the
changes in ideal equality for this group conformed to the same second regression, the key issue was whether change in relation-
U-shaped curve that was described for Partner 1. ship commitment from Time 2 to Time 3 was related to change
in the set of discrepancy scores from Time 2 to Time 3.
Predicting Change in Relationship Commitment Predicting change in relationship commitment from Time 1
to Time 2 from change in discrepancy scores from Time 1 to
In contrast to the group trends highlighted earlier, the focus Time 2. To see if regressed change in Time 2 commitment was
here was on individual differences in changes in relationship linked to regressed change in both sets of Time 2 discrepancy
commitment over the three annual assessments. Although the scores, I ran a three-step hierarchical multiple regression for
modal difference between the Time 3 and the Time 1 relation- each partner. (Because preliminary analyses indicated that the
ship commitments was 0 (i.e., no change) for both Partner 1 standardized betas associated with the demographic scores [age,
(24% of the sample) and Partner 2(17% of the sample), the full education, income, months living together, and type of couple]
range for change in relationship commitment from Time 1 to were not significant in the final model, these scores were not
Time 3 was -25 to 32 for Partner 1 and -20 to 24 for Partner included.) At Step 1, each partner's Time 1 commitment score
2. Thus, for both Partner 1 and Partner 2, appreciable change was entered so that further steps focused on explaining re-
occurred in relationship commitment across the three annual gressed or residual change (Cohen & Cohen, 1983). At Step 2,
assessments. the Time 1 discrepancy scores for attachment, autonomy, and
Of interest here was whether change in relationship commit- equality were entered as well as the Time 1 attachment-auton-
ment was linked to change in two sets of discrepancy scores: (a) omy discrepancy score. Finally, in the critical Step 3, the Time
the discrepancy between ideal and current levels oattachment, 2 discrepancy scores for attachment, autonomy, and equality
SPECIAL ISSUE: GAY AND LESBIAN COUPLES 91

Table 5 that change in relationship commitment was linked to changes


Means and Standard Deviationsfor Relationship-Oriented in both sets of discrepancy scores would be supported if the
Scores by Type of Couple and Time of Assessment information added at Step 3 (regressed change in the discrep-
for Partner 1 and Partner 2 ancy scores) resulted in a significant increase in explained vari-
ance in regressed change in relationship commitment.
Gay Lesbian The findings from these analyses are summarized in the top
(n = 61)
i (n = 42) of Table 6. Overall, findings were similar for each partner: The
Time Time Time Time Time Time addition of the set of Time 2 regressed change discrepancy
Score 1 2 3 1 2 3 scores accounted for an additional 12% of the variance in re-
gressed change in relationship commitment for Partner 1 and
Partner 1 29% of the variance in regressed change in relationship com-
Commitment mitment for Partner 2. Furthermore, an inspection of the stan-
M 64.80 63.50 65.09 67.81 68.07 66.38 dardized beta weights associated with the full three-step equa-
SD 8.33 9.94 9.15 4.75 4.21 8.48 tion indicated that for both Partner 1 and Partner 2, unique
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Ideal importance variability in regressed change in relationship commitment was


This document is copyrighted by the American Psychological Association or one of its allied publishers.

Attachment accounted for by regressed change in both the ideal versus cur-
M 48.96 49.37 48.32 49.54 47.88 46.59
SD 10.16 10.27 10.75 10.78 11.56 10.12 rent discrepancy in equality (-.32 for Partner 1 and -.41 for
Autonomy Partner 2, ps < .01) and the current attachment versus current
M 45.98 47.72 45.04 47.81 47.88 47.50 autonomy discrepancy (.32 for Partner 1 and .47 for Partner 2,
SD 13.00 11.72 13.16 10.82 10.75 13.89
ps < .05).
Equality
M 60.68 63.04 62.11 65.78 63.16 64.83 In additional analyses not reported here, interaction terms
SD 7.99 6.50 6.60 6.04 7.54 5.55 involving type of couple and the discrepancy scores were added
Current level in an extra step in the regressions. For both partners, effects
Attachment associated with these interaction terms were not significant.
M 52.50 52.34 52.34 52.90 54.07 51.83
SD 9.34 10.13 10.78 9.73 9.16 8.96 Thus, type of couple did not moderate the relations of interest.
Autonomy Predicting change in relationship commitment from Time 2
M 46.03 46.60 46.27 49.92 50.38 51.02 to Time 3 from change in discrepancy scores from Time 2 to
SD 10.55 10.57 12.68 8.71 9.26 9.88 Time 3. The regressions run to predict Time 3 commitment
Equality
M 61.90 61.01 61.15 64.90 63.85 63.50 were analogous to those run to predict Time 2 commitment
SD 7.52 8.45 9.38 8.07 7.51 10.11 from Time 1 discrepancy scores, with the exceptions that the
dependent variable was Time 3 commitment and the discrep-
Partner 2 ancy scores came from Time 2. Thefindingsfrom these analyses
Commitment are summarized in the bottom of Table 6. Again, findings were
M 65.67 64.77 65.55 66.76 67.28 65.09 similar for each partner: The addition of the set of Time 3 re-
SD 6.79 6.77 6.30 5.98 5.37 8.39 gressed change discrepancy scores at Step 3 accounted for an
Ideal importance additional 20% of the variance in regressed change in relation-
Attachment
M 47.32 47.80 46.16 47.81 46.38 44.38 ship commitment for Partner 1 and 13% for Partner 2. Further-
SD 9.03 10.10 12.22 10.53 10.11 13.20 more, an inspection of the standardized beta weights associated
Autonomy with the full three-step equation indicated that for both Partner
M 44.75 45.90 44.60 49.47 49.04 50.54 1 and Partner 2, unique variability in regressed change in rela-
SD 11.46 11.65 12.25 11.84 11.16 12.77
' Equality tionship commitment was again accounted for by regressed
M 58.60 58.77 58.34 64.47 62.19 63.64 change in both the ideal versus current discrepancy in equality
SD 10.19 8.09 8.86 6.99 8.66 6.30 (-.35 for Partner 1 and -.32 for Partner 2, ps < .01) and the
Current level attachment versus autonomy discrepancy (.38 for Partner 1 and
Attachment .36 for Partner 2, ps < .01). Finally, additional analyses indi-
M 53.39 52.93 52.29 52.21 51.88 49.33
SD 8.31 8.16 9.17 9.81 9.99 10.94 cated that type of couple did not moderate the relations of
Autonomy interest.
M 47.18 47.95 48.03 49.09 49.71 50.64
SD 9.58 9.23 9.74 9.06 8.18 9.51
Equality Discussion
M 50.24 61.34 60.82 62.71 62.38 61.23
SD 8.05 6.90 7.88 9.44 9.19 10.70 The basic premise of this study was that negotiating issues
regarding close personal relationships is a key developmental
task for many gay men and lesbians (Erikson, 1963; Havigurst,
1953). Two particular issues regarding close personal relation-
were entered as well as the Time 2 attachment-autonomy dis- ships were of interest. Thefirstissue was descriptive and focused
crepancy score. on group trends in how gay partners differed from lesbian part-
Because the Time 1 discrepancy values were controlled in ners in the nature of change over three annual assessments in
Step 2, Step 3 provided information on regressed change for the relationship commitment; in currently perceived levels of at-
discrepancy scores (Cohen & Cohen, 1983). The hypothesis tachment, autonomy, and equality; and in the importance at-
92 LAWRENCE A. KURDEK

Table 6 for lesbian couples, it may be an especially sensitive barometer


Summary ofHierarchical Multiple Regressions of changes in overall relationship quality. Future studies might
for Partner 1 and Partner 2 examine how appraisals of the importance of equality are
Partner 1
affected by degrees of conflict and harmony in the relationship.
Partner 2
Finally, for both gay and lesbian partners, the value attributed
2 2
Step Variable R F change R /"change to attachment in the ideal relationship decreased over time.
Thisfindingis consistent with accounts of relationship develop-
Predicting Time (T) 2 commitment ment (McWhirter & Mattison, 1984) that suggest that it is nor-
1 Tl commitment .47** 96.99** .29** 45.49** mative that partners in long-lasting relationships experience a
2 Tl attachment decline in the importance and intensity of the need to do things
Tl autonomy as a couple. As feelings of trust and security increase, partners
Tl equality may even begin to take each other for granted.
Tl balance .53** 3.27* .36** 2.95*
3 T2 attachment
T2 autonomy Predicting Changes in Relationship Commitment
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

T2 equality
This document is copyrighted by the American Psychological Association or one of its allied publishers.

T2 balance .65** 9.01** .65** 20.78** Consistent with predictions derived from interdependence
Predicting Time 3 commitment theory (Kelley, 1983; Rusbult, 1983), changes in ideal versus
current levels of equality and changes in the balance between
1 T2 commitment .45** 84.15** .34** 53.70** current levels of attachment and current levels of autonomy ex-
2 T2 attachment
T2 autonomy plained unique portions of change in relationship commitment.
T2 equality This pattern is of particular note because it explained changes
T2 balance .48** 1.19 .43** 3.73** in relationship commitment from Time 1 to Time 2 as well as
3 T3 attachment from Time 2 to Time 3, replicated across Partner 1 and Partner
T3 autonomy 2, and was not moderated by type of couple.
T3 equality
T3 balance .68** 14.36** .56** 7.24** Previous analyses of the first wave of data obtained from this
sample found that an aggregate measure of ideal versus current
*p<.05. **p<.01. discrepancies was linked to relationship satisfaction for both
gay and lesbian partners (Kurdek, 1991). This study extends
that link to relationship commitment and documents that de-
tributed to attachment, autonomy, and equality in the ideal re- creases in relationship commitment occur when currently per-
lationship. The second issue focused on individual differences ceived levels of equality increasingly fall short of their ideal lev-
in changes in relationship commitment. On the basis of inter- els and when the relative importance of autonomy over attach-
dependence theory (Kelley, 1983; Rusbult, 1983), changes in ment increases. In short, the change patterns found here are
relationship commitment were expected to be linked to changes consistent with the position that developmental changes in rela-
in the discrepancy between ideal and current levels of attach- tionship commitment are linked to corresponding developmen-
ment, autonomy, and equality as well as to changes in the dis- tal changes regarding the causal conditions that keep a person
crepancy between currently perceived levels of attachment and in or draw a person away from a relationship (Kelley, 1983).
currently perceived levels of autonomy. Findings from each is- The presentfindingshave two sets of theoretical implications
sue are discussed in turn. for the general study of close relationships. First, with one major
exception (Drigotas & Rusbult, 1992), the discussion of com-
Differences Between Gay Men and Lesbians in Changes parison levels in interdependence theory has been unidimen-
in Interpersonal Intimacy sional. That is, there has been limited consideration that the
discrepancy between actual and ideal relationship dimensions
Because data were available from both partners, separate may vary from dimension to dimension and that comparison
analyses for Partner 1 and Partner 2 yielded within-study repli- levels for particular dimensions may be especially critical for
cations. Threefindingsgeneralized across both partners. First, relationship commitment and even relationship stability.
compared with gay partners, lesbian partners attributed greater Thefindingsof this study highlight the salience of compari-
importance to equality in their ideal relationships. This finding son levels for the dimension of equality. Because gay and espe-
is consistent with previous evidence that lesbian couples are es- cially lesbian relationships are more egalitarian than are hetero-
pecially likely to follow an ethic of equality (Kurdek, 1994), per- sexual relationships (Kurdek, 1994), the comparison level for
haps in reaction to socialization pressures that typically reduce this dimension is understandably important. Future studies
women's power in intimate relationships (Blumstein & could include a sample of heterosexual couples to assess
Schwartz, 1983). whether the comparison level for equality is indeed more im-
Second, compared with gay partners who showed no change portant for changes in relationship commitment (as well for
in the importance attributed to equality in the ideal relation- predicting relationship stability) for gay and lesbian than for
ship, lesbian partners showed a U-shaped curvilinear pattern. heterosexual couples.
That is, relative to assessments obtained at Times 1 and 3, the Second, a major proposition of interdependence theory is
assessment at Time 2 was lower. It is unclear how to explain this that relationship commitment depends on the balance of the
"dip." It is possible that if equality is a core relationship ideal pushes and pulls to the relationship, that is, their joint effect.
SPECIAL ISSUE: GAY AND LESBIAN COUPLES 93

However, previous empirical tests of the theory have considered Duffy, S. M., & Rusbult, C. E. (1986). Satisfaction and commitment in
only the independent effects exerted by satisfaction and invest- homosexual and heterosexual relationships. Journal ofHomosexual-
ity, 12, 1-23.
ments (pulls) and alternatives (pushes; Rusbult, 1983). This
Erikson, E. H. (1963). Childhood and society (2nd ed.). New York:
study found that changes in the balance between currently per-
Norton.
ceived levels of attachment and currently perceived levels of au- Havighurst, R. (1953). Human development and education. London:
tonomy reliably predicted changes in relationship commit- Longmans, Green.
ment. Thus, future refinements and tests of interdependence Kelley, H. H. (1983). Love and commitment. In H. H. Kelley, E.
theory might include components that directly address not only Berscheid, A. Christensen, J. H. Harvey, T. L. Huston, G. Levinger,
the effects of pushes and pulls on relationships but also the rel- E. McClintock, L. A. Peplau, & D. R. Peterson (Eds.), Close relation-
ative balance of these effects. ships (pp. 265-314). New York: Freeman.
Finally, it is of note that type of couple did not moderate the Kurdek, L. A. (1988). Perceived social support in gays and lesbians in
cohabiting relationships. Journal of Personality and Social Psychol-
relation between changes in relationship commitment and ogy, 54, 504-509.
changes in the sets of discrepancy variables. Because both gay Kurdek, L. A. (1991). Correlates of relationship satisfaction in cohabit-
partners and lesbian partners were fairly well-educated and ing gay and lesbian couples: An integration of contextual, investment,
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

were participating in a longitudinal study that involved public and problem-solving models. Journal of Personality and Social Psy-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

disclosure of being gay or lesbian, these findings cannot be chology, 61,910-922.


claimed to generalize to the diverse array of gay and lesbian cou- Kurdek, L. A. (1992). Relationship stability and relationship satisfac-
ples. Despite the nonrepresentative nature of the sample, the tion in cohabiting gay and lesbian couples: A prospective longitudinal
finding that changes in relationship commitment are similarly test of the contextual and interdependence models. Journal ofSocial
and Personal Relationships, 9, 125-142.
explained for both lesbian partners and gay partners is consis-
Kurdek, L. A. (1994). Lesbian and gay couples. In A. R. D'Augelli &
tent with previous findings that the predictors of relationship C. J. Patterson (Eds.), Lesbian and gay identities over the lifespan:
satisfaction and relationship stability are similar for gay and les- Psychological perspectives on personal, relational, and community
bian couples (Kurdek, 1994). processes (pp. 243-261). New York: Oxford University Press.
Kurdek, L. A., & Schmitt, J. P. (1987). Perceived emotional support
from family and friends in members of gay, lesbian, and heterosexual
References cohabiting couples. Journal ofHomosexuality, 14, 57-68.
McWhirter, D. P., & Mattison, A. M. (1984). The male couple: How
Baucom, D. H., & Epstein, N. (1990). Cognitive-behavioral marital relationships develop. Englewood Cliffs, NJ: Prentice Hall.
therapy. New York: Brunner/Mazel. Peplau, L. A., & Cochran, S. D. (1981). Value orientations in the inti-
Belsky, J., & Rovine, M. (1990). Patterns of marital change across the mate relationships of gay men. Journal ofHomosexuality, 6, 1-20.
transition to parenthood: Pregnancy to three years postpartum. Jour- Raush, H. (1977). Orientations to the close relationship. In G. Levinger
nal ofMarriage and the Family. 52, 5-19. & H. Raush (Eds.), Close relationships: Perspectives on the meaning
Blumstein, P., & Schwartz, P. (1983). American couples. New York: Wil- of intimacy (pp. 163-188). Amherst: University of Massachusetts
liam Morrow. Press.
Bryant, S., & Demian (1994). Relationship characteristics of American Rusbult, C. E. (1983). A longitudinal test of the investment model: The
gay and lesbian couples: Findings from a national survey. Journal of development (and deterioration) of satisfaction and commitment in
Gay and Lesbian Social Services, 1, 101-117. heterosexual involvements. Journal of Personality and Social Psy-
Cochran, S. D., & Peplau, L. A. (1985). Value orientations in heterosex- chology, 45, 101-117.
ual relationships. Psychology of Women Quarterly, 9, 477-488. Stanley, S. M., & Markman, H. J. (1992). Assessing commitment in
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation personal relationships. Journal ofMarriage and the Family, 54, 595-
analysis for the behavioral sciences. Hillsdale, NJ: Erlbaum. 608.
Drigotas, S. M., & Rusbult, C. E. (1992). Should I stay or should I go? Steinberg, R. J. (1988). Triangulating love. In R. J. Sternberg & M. L.
A dependence model of breakups. Journal of Personality and Social Barnes (Eds.), The psychology oflove (pp. 119-138). New Haven, CT:
Psychology, 62, 62-87. Yale University Press.

(Appendixfollows on next page)


94 LAWRENCE A. KURDEK

Appendix

Items for Commitment and Current Versus Ideal Ratings of Attachment, Autonomy, and Equality
Commitment 3. I have a close friend other than my partner/Having a close friend
other than my partner.
1. I am committed to maintaining my relationship with my partner. 4. My sense of being an individual is separate from my sense of being
2. Because of my commitment to my partner, I would not let other part of a couple./Keeping my sense of being an individual separate from
people come between us. my sense of being part of a couple.
3. I have confidence in the stability of my relationship with my 5. I make most decisions on my own without checking with my part-
partner. ner./Being able to make most decisions on my own without checking
4. I view my commitment to my partner as a solid one. with my partner.
5. I cannot imagine ending my relationship with my partner. 6. I maintain the position that if I had to, I could really "make it on
6. I view my relationship with my partner as permanent. my own."/Maintaining the position that if I had to, I could really "make
7. I plan to continue in my relationship with my partner.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

it on my own."
8. Even when my partner is hard to deal with, I remain committed
This document is copyrighted by the American Psychological Association or one of its allied publishers.

7. Although I can depend on my partner, the only one I can really


to our relationship. count on is myself/Knowing that, although I can depend on my part-
ner, the only one I can really count on is myself.
Attachment 8. I manage my ownfinances./Managingmy own finances.

Current Version/Ideal Version Equality


1. Sex with my partner is one of the most intimate forms of commu- Current Version/Ideal Version
nication that we have./Seeing sex as one of the most intimate forms of
communication that we have. 1. My partner and I have equal power in the relationship./Knowing
2. I spend as much time with my partner as possible./Spending as that my partner and I have equal power in the relationship.
much time with my partner as possible. 2. My partner shows as much affection to me as I think I show to him
3. I do as many activities with my partner as possible./Sharing as [her]./Having my partner shown as much affection to me as I think I
many activities with my partner as possible. show to him [her].
4. My partner and I have built an identity as a couple./Building an 3. My partner and I invest equal amounts of time and energy in the
identity as a couple. relationship./Having my partner and I invest equal amounts of time
5. I get so close to my partner that I'm not sure where he [she] begins and energy in the relationship.
and I end./Getting so close to my partner that I'm not sure where he 4. My partner and I are equally committed to working out problems
[she] begins and I end. that occur in our relationship./Knowing that my partner and I are
6. My partner is a very important part of how I see myself./Having equally committed to working out problems that occur in our
my partner be a very important part of how I see myself. relationship.
7. I think in terms of "we" and "us" instead of "I" and "me."/Think- 5. All things considered, my partner and I contribute an equal
ing in terms of "we" and "us" instead of "I" and "me." amount to the relationship./Believing that, all things considered, my
8. I can never get too close to my partner./Feeling that I could just partner and I contribute an equal amount to the relationship.
never get too close to my partner. 6. My partner and I deal with each other as equals./Having a sense
that my partner and I deal with each other as equals.
7. My partner treats me and respects me as an equal./Having my
Autonomy partner treat me and respect me as an equal.
8. My partner depends on me as much as I depend on him [her]./
Current Version/Ideal Version
Having my partner depend on me as much as I depend on him [her].
1. I have major interests of my own outside of the relationship./Hav-
ing major interests of my own outside of the relationship. Received October 12, 1992
2. I have a supportive group of friends separate from my partner./ Revision received May 13, 1993
Having a supportive group of friends separate from my partner. Accepted May 26, 1993

Vous aimerez peut-être aussi