ArticlePDF Available

Observing Gay, Lesbian and Heterosexual Couples' Relationships: Mathematical Modeling of Conflict Interaction

Authors:

Abstract and Figures

Two samples of committed gay and lesbian cohabiting couples and two samples of married couples (couples in which the woman presented the conflict issue to the man, and couples in which the man presented the conflict issue to the woman) engaged in three conversations: (1) an events of the day conversation (after being apart for at least 8 hours), (2) a conflict resolution conversation, and (3) a pleasant topic conversation. The observational data were coded with a system that categorized specific affects displayed. Data were weighted and two time-series created, one for the husband and one of the wife. The time series were modeled with nonlinear difference equations (Cook et al., 1995), and parameters were estimated that indexed uninfluenced steady state, influenced steady state, emotional inertia, repair effectiveness and threshold, and the power of positive and negative affect of one partner to affect the other partner.
Content may be subject to copyright.
Observing Gay, Lesbian
and Heterosexual Couples’ Relationships:
Mathematical Modeling
of Conflict Interaction
John Mordechai Gottman, PhD
University of Washington
Robert Wayne Levenson, PhD
University of California at Berkeley
Catherine Swanson
Kristin Swanson
Rebecca Tyson
Dan Yoshimoto
University of Washington
ABSTRACT. Two samples of committed gay and lesbian cohabiting
couples and two samples of married couples (couples in which the
woman presented the conflict issue to the man, and couples in which the
John Mordechai Gottman is affiliated with the University of Washington. Robert
Wayne Levenson is affiliated with the University of California at Berkeley. Catherine
Swanson, Kristin Swanson, Rebecca Tyson, and Dan Yoshimoto are affiliated with the
University of Washington. Correspondence may be addressed to: John Gottman, Uni-
versity of Washington, Department of Psychology, Box 351525, Seattle, WA 98195.
The authors are grateful for the support of NIMH grant R01 MH50841 to Robert
Levenson.
Journal of Homosexuality, Vol. 45(1) 2003
http://www.haworthpress.com/store/product.asp?sku=J082
© 2003 by The Haworth Press, Inc. All rights reserved.
10.1300/J082v45n01_04 65
man presented the conflict issue to the woman) engaged in three conversa-
tions: (1) an events of the day conversation (after being apart for at least 8
hours), (2) a conflict resolution conversation, and (3) a pleasant topic con-
versation. The observational data were coded with a system that categorized
specific affects displayed. Data were weighted and two time-series cre-
ated, one for the husband and one of the wife. The time series were mod-
eled with nonlinear difference equations (Cook et al., 1995), and parameters
were estimated that indexed uninfluenced steady state, influenced steady
state, emotional inertia, repair effectiveness and threshold, and the power of
positive and negative affect of one partner to affect the other partner. [Arti-
cle copies available for a fee from The Haworth Document Delivery Service:
1-800-HAWORTH. E-mail address: <docdelivery@haworthpress.com> Website:
<http://www.HaworthPress.com> © 2003 by The Haworth Press, Inc. All rights re-
served.]
KEYWORDS. Relationships, interaction, stability, dissolution, emo-
tions, humor, affection, criticism, contempt, prediction
Although there is a venerable history of research on gays and lesbi-
ans as individuals, relatively little descriptive scientific research exists
about gay and lesbian relationships. This is certainly not because gay
and lesbian relationships are rare. Even in research done at the Kinsey
Institute in the 1960s, Weinberg, and Williams (1974) reported that 71
percent of a sample of gay men between the ages of 36 and 45 were liv-
ing with a partner. In the 1970s Bell and Weinberg (1978) found that
82% of their lesbian sample were currently living with someone. Gay
men and lesbian women have always been very concerned about having
satisfying and stable relationships, and since the AIDS epidemic this in-
terest in the gay and lesbian communities in having stable committed
relationships has increased (Andriote, 1999).
In the 1980s an important book, American Couples (Blumstein &
Schwartz, 1983), reported the results of over 12,000 questionnaires and
more than 300 interviews with gay, lesbian, married heterosexual, and
cohabiting heterosexual couples on issues related to money, work,
power, and sex. The timing of this landmark study, meant that it pre-
ceded the AIDS epidemic, which was to have such a large impact on
gay and lesbian relationships. The Blumstein and Schwartz study pro-
vided one of the first opportunities to compare heterosexual and homo-
sexual relationships.
66 JOURNAL OF HOMOSEXUALITY
Since 1983 sophisticated longitudinal research has been conducted
on gay, lesbian and heterosexual married couples by Kurdek and his as-
sociates (e.g., Kurdek, 1998). Kurdek concluded that gay and lesbian
relationships operate on essentially the same principles as heterosexual
relationships, except that there are fewer barriers to leaving, and gay re-
lationships emphasize autonomy more than do heterosexual relation-
ships. He wrote, that by and large, the “correlates of relationship quality
have been found to be very similar for gay and lesbian couples
(Blumstein and Schwartz, 1983; Duffy and Rusbult, 1986; Kurdek and
Schmitt, 1986)” (Kurdek, 1992, p. 130). There were some differences
as well. Compared to married partners, gay partners reported more au-
tonomy, fewer barriers to leaving and more frequent relationship disso-
lution. Compared to married partners lesbian relationships reported
more intimacy, more autonomy, more equality and more frequent rela-
tionship dissolution. He wrote, “Overall, the strength with which the di-
mensions of relationship quality were linked to each relationship
outcome for married partners was equivalent to that for both gay and
lesbian partners” (Kurdek, 1998, p. 553).
Previous research on gay and lesbian relationships has provided im-
portant insights on the nature of these relationships; however, it has re-
lied entirely on self-report data derived from questionnaires (e.g.,
Kurdek, 1992), or questionnaires and interviews (e.g., Blumstein &
Schwartz, 1983). While these forms of data produce valuable informa-
tion, they are limited to people’s perceptions of their own relationships.
There is considerable evidence that people’s perception of their rela-
tionships may diverge quite markedly from their actual interaction. For
example, in an observational study of positive interaction at home, Rob-
inson and Price (1980) found that, compared to observers, distressed
couples underestimated positive interaction by 50%. Hence, there may
be valuable information to be gained by employing observational re-
search to study gay and lesbian relationships.
This paper reports the results of what, to our knowledge, is the first
observational study of gay and lesbian relationships. These observa-
tional data are used to support mathematical modeling of the nature of
interaction in these relationships. The mathematical modeling provides
both descriptive and theoretically-based data about interaction.
Two samples of heterosexual couples were also part of the experi-
mental design of this study. Analyzing the interaction of gay, lesbian,
and heterosexual couples presents a thorny issue in distinguishing be-
tween the two partners. With heterosexual couples, researchers usually dis-
tinguish “husbands” and “wives” but this cannot be done with same-sex
Gottman et al. 67
couples. In our research we addressed this issue by taking advantage of
the fact that our paradigm had couples discuss an area of conflict in their
interaction. We distinguished between partners in our analyses by not-
ing which partner initially presented the conflict. The samples of het-
erosexual couples were chosen so that in one group, the wife presented
the issue to the husband and in the other the husband presented the issue
to the wife. When all couples were combined, a 2 2 factorial design
resulted in which the first factor indicated who initiated the issue (a man
or a woman), and the second factor indicated whether the couple was
homosexual or heterosexual. In this manner, we could compare the in-
teraction when a man presented a relationship issue to a man he was
committed to the interaction when a man presented a relationship issue
to a woman he was committed to. Similarly, we could compare the in-
teraction when a woman presented a relationship issue to a woman she
was committed to with the interaction when a woman presented a rela-
tionship issue to a man she was committed to. This approach contrasts
with Kurdek’s, who randomly assigned each partner in the homosexual
relationship either to be “Partner 1” or “Partner 2” and then analyzed the
data twice. In our approach, one partner was designated “the initiator”
of the issue and the other was designated “the partner.”
Mathematical Modeling of Marital Interaction. Before presenting
our methods and findings, we will provide some background on mathe-
matical modeling intended to set the stage for what follows. For the past
seven years our laboratory has been constructing mathematical models
of marital interaction (Cook et al., 1995; Gottman et al., 1998). This
modeling involves estimating two linked nonlinear difference equa-
tions, one for the husband and one for the wife (described next). The de-
sire to create these mathematical models was inspired by General
System Theory (Von Bertalanffy, 1963). This book inspired many ma-
jor thinkers of family systems and family therapy, including Gregory
Bateson, Don Jackson, and Paul Watzlawick. Unfortunately, the mathe-
matics of General System Theory was not utilized by most of the social
scientists who influenced by von Bertalanffy’s work. Bateson and col-
leagues originally envisaged making their family systems theory math-
ematical (for an historical review see Gottman, 1979), but did not do so.
Hence, the nonmathematical work of these theorists of family interac-
tion kept their systems concepts at the level of metaphor. Even at the
level of metaphor these concepts were tremendously influential in the
field of family therapy (see Rosenblatt, 1994). However, they were
never quantified or subjected to experimental processes.
68 JOURNAL OF HOMOSEXUALITY
Von Bertalanffy clearly viewed his theory as mathematical. He be-
lieved that the interaction of complex systems with many units could be
characterized by a set of values that change over time, denoted Q 1,Q2,
Q3, and so on. Each of these Q’s were variables that indexed something
about a particular unit in the system, such as mother, father, and child.
He thought that the system could be best described by a set of ordinary
differential equations of the form:
dQ1/dt = f1(Q1, Q2,Q3, . . .)
dQ2/dt = f2(Q1, Q2,Q3, . . .)
and so on.
The terms on the left of the equal sign are time derivatives, that is,
rates of change of the quantitative sets of values Q 1,Q2, and so on. The
terms on the right of the equal sign are functions, f1,f
2,...,oftheQs.
Von Bertalanffy thought that these functions, the fs, might generally be
nonlinear. The equations Von Bertalanffy selected have a particular
form, called “autonomous,” meaning that the fs have no explicit func-
tion of time in them, except through the Qs, which are functions of time.
Von Bertalanffy had no suggestions for what the Qs ought to measure,
nor what the fs ought to be. So his vision remained a creative one with-
out a quantitative science to back it up.
We obtained the Qs for our modeling from our ability to predict the
longitudinal course of marriages. Gottman and Levenson (1992) re-
ported that a variable that describes specific interaction patterns in
terms of the balance between negativity and positivity was predictive of
marital dissolution. In this work, we used a methodology for obtaining
synchronized physiological, behavioral, and self-report data in a sample
of 73 couples who were followed longitudinally during a four-year pe-
riod. Applying observational coding of interactive behavior with the
Rapid Couples Interaction Scoring System (RCISS, Krokoff, Gottman
& Hass, 1989), couples were divided into two groups, a low-risk group
and a high-risk group. This classification was based on a graphical
method originally proposed by Gottman (1979) for use with the Cou-
ples Interaction Scoring System, a predecessor of the RCISS. On each
conversational turn the total number of positive RCISS speaker codes
minus the total number of negative speaker codes was computed for
each spouse. Then the cumulative total of these points was plotted for
each spouse. The slopes of these plots, which were thought to provide a
stable estimate of the difference between positive and negative codes
over time, were determined using linear regression analysis. The deci-
sion to utilize the slopes in this way was guided by a balance theory of
Gottman et al. 69
marriage, namely that those processes most important in predicting dis-
solution would involve a balance, or a regulation, of positive and nega-
tive interaction. Low-risk couples were those for whom both husband
and wife speaker slopes were significantly positive; high-risk couples
had at least one of the speaker slopes that was either negative or not sig-
nificantly positive. We found that the high/low-risk distinction was able
to predict the “cascade” toward divorce, which consisted of: (a) marital
dissatisfaction, (b) persistent thoughts about divorce and separation,
and (c) actual separation and divorce. The ability to predict the longitu-
dinal course of marital relationships in this manner has now been found
in our laboratories in four separate longitudinal studies (see Gottman,
1993, 1994; Gottman, Coan, Swanson, & Carrere, 1998; Jacobson et al.,
1996).
In our first paper using mathematical modeling (Cook et al., 1995),
we made use of these speaker slopes as the Qs in our equations and used
them to develop a mathematical model that might explain the Gottman-
Levenson findings. Our own equations were very similar to the ones
that Von Bertalanffy had envisioned, except that we used discrete dif-
ference equations rather than differential equations. There is one addi-
tional difference. Von Bertalanffy thought that the equations had to be
linear. He presented a table in which these nonlinear equations were
classified as “Impossible” (von Bertalanffy, 1968, p. 20), referring to
the very popular mathematical method of approximating nonlinear
functions with a linear approximation. Unfortunately, linear equations
are not generally stable, so they tend to give erroneous solutions, except
as approximations under very local conditions near a steady state. Von
Bertalanffy was not aware of the mathematics Poincarè and others had
developed in the last quarter of the 19th century for the study of nonlin-
ear systems. It was actually no longer the case that these nonlinear sys-
tems were “impossible,” even in Von Bertalanffy’s day. This is even
more true today when the modeling of complex deterministic (and sto-
chastic) systems with a set of nonlinear difference or differential equa-
tions has become a productive enterprise across a wide set of phenomena
in a widerange of sciences. Thus, the use of nonlinear equations formed
the basis of our first attempts at modeling marital interaction (these
methods are described in detail in Cook et al., 1995).
This general method of mathematical modeling with nonlinear equa-
tions has been employed with great success in the biological sciences,
and many departments of applied mathematics now have a mathemati-
cal biology program (Murray, 1989). It is a quantitative approach that
allows the modeler to be able to write down, in mathematical form, on
70 JOURNAL OF HOMOSEXUALITY
the basis of some theory, the causes of change in the dependent vari-
ables. For example, in mathematical ecology, in the classic preda-
tor-prey problem, one writes down the rate of change in the population
densities of the prey and of the predator as some function of their cur-
rent densities (e.g., Murray, 1989). While this is a simple representation
of the predator-prey phenomenon, it has served well as an initial explor-
atory model. An advantage of nonlinear equations (in addition to their
possibility of stability) is that by employing nonlinear terms in the equa-
tions of change some very complex processes can be represented with
very few parameters. Unfortunately, unlike many linear equations,
these nonlinear equations are generally not solvable in closed func-
tional mathematical form. For this reason the methods are often called
“qualitative,” and visual graphical methods and numerical approxima-
tion must be relied upon. For this purpose, numerical and graphical
methods have been developed such as “phase space plots.” These visual
approaches to mathematical modeling can be very appealing in engag-
ing the intuition of a scientist working in a field that has no mathemati-
cally stated theory. If the scientist has an intuitive familiarity with the
data of the field, our approach may suggest a way of building theory us-
ing mathematics in an initially qualitative manner. The use of these
graphical solutions to nonlinear differential equations makes it possible
to talk about “qualitative” mathematical modeling. In qualitative math-
ematical modeling, one searches for solutions that have similarly
shaped phase space plots, which provide a good qualitative description
of the solution and how it varies with the parameters.
Another advantage of this kind of modeling is that it permits the sim-
ulation of the couple’s interaction under new conditions, with different
parameter values, such as when a partner begins the interaction much
more positively than has ever been observed. This possibility leads to natu-
ral experiments. There are many excellent introductions to this general ap-
proach to qualitative nonlinear dynamic modeling and its subtopics of
chaos and catastrophe theory. We refer the reader to a few sources here.
A general introduction to chaos theory was provided in a popular book
by Gleick (1987), and a clear introduction to its mathematics by the
physicists Baker and Gollub (1996). Qualitative modeling is introduced
by Morrison (1991). Introductions to the mathematics of this approach
(a system of nonlinear differential equations) can be found in Brauer
and Nohel (1969), and Jordan and Smith (1987). Introductions to the
extension from differential to difference equations can be found in
Goldberg (1986) and in Kelley and Peterson (1991). Applications to the
study of disease can be found in Glass and Mackey (1988), to biological
Gottman et al. 71
rhythms by Winfree (1990), to weather by Lorenz (1993), to the study
of turbulence by Berge, Pomeau, and Vidal (1984), to the study of capi-
tal markets by Peters (1991), to social and biological sciences by
Beltrami (1993), and to social psychology by Vallacher and Nowack
(1994). An introduction to catastrophe theory may be found in Arnold
(1986), Castrigiano and Hayes (1993), Gilmore (1981), and Saunders
(1990). The currently vast and expanding area of mathematical biology
is introduced by Murray (1989).
Our modeling of marital interaction using the mathematical methods
of nonlinear difference equations is an attempt to integrate the mathe-
matical insights of von Bertalanffy with the General Systems Theorists
of Family Systems (Bateson, Jackson, Haley, & Weakland, 1956) using
nonlinear equations. A basic concept in this modeling is that every sys-
tem of equations has one or more stable or unstable “steady states” or
“attractors.” These stable attractors are like the old family systems no-
tion of homeostatic set points of the system. These are values toward
which the system is drawn, and, if perturbed from the stable attractor,
returns the system back toward it. In our modeling there were both unin-
fluenced attractors and influenced attractors.
The Concept of Repair. In theorizing about mother-infant interaction,
Brazelton (e.g., Brazelton, Koslowski, & Main, 1974) had suggested
that “repair” was the sine qua non of human interaction, and that the
healthy norm was a cyclic shared rhythmicity and coordination of inter-
action. However, Gianino and Tronick (1988), in observing mother-in-
fant interaction, reported that 70% of mother-infant interaction was
actually miscoordinated. What they found was that some mothers no-
ticed the miscoordination and tried to repair the interaction, while other
mothers did not. The mother’s use of repair later predicted the infant’s
attachment security. Gianino and Tronick built a theory of interactive
regulation and repair based on their findings. In our own modeling we
studied couple’s repair of negativity, which, in our research on married
couples is the one central theoretical construct that has consistently dis-
criminated those couples whose marriages were dissatisfied and headed
for divorce from those whose marriages were satisfied and stable (e.g.,
see Gottman, 1999).
Parameters of the Modeling. Our initial modeling (Cook et al., 1995)
produced the following parameters. First, there was the uninfluenced
steady state for each partner, which represented the attractors reflecting
what that partner brought into the interaction before influence began.
This uninfluenced state turned out to be a function both of that partner’s
personality and the immediate past history of the relationship (Gottman
72 JOURNAL OF HOMOSEXUALITY
et al., 1998). Second were the inertia parameters, which assessed the
tendency of each person’s behavior to be predictable from that person’s
immediate past behavior. Third, were the influenced steady states,
which were where each partner was drawn to following the social influ-
ence process. Fourth, were the influence functions, which, for each af-
fect value, describe the average effect of that affect (over the entire
interaction) on the partner. These influence functions provide a more
detailed description of interpersonal influence or power. Power has
been an elusive construct in the study of relationships (for a critique and
review see Gray-Little and Burks, 1983). Power is defined in our mod-
eling as the ability of one person’s affect to move the other partner’s af-
fect.
Fifth, our initial model has now been modified to include a Repair
term, which is repair of negative interaction that is potentially triggered
at a particular threshold of a partner’s negativity and is effective at
pushing the data in a more positive direction. The two Repair terms
have two parameters, the threshold of the repair and its effectiveness.
METHOD
Participants
Gay and Lesbian Samples. Couples were recruited by placing adver-
tisements in the classified sections of Berkeley and San Francisco gay
newspapers, posting flyers, contacting various gay and lesbian groups,
and making public service announcements on Bay area radio stations.
Advertisements and announcements asked for “volunteer couples, in-
cluding those with relationship problems, needed for a paid UC Berke-
ley research project on committed relationships.” Partners had to be
between the ages of 21 and 40 and living together in a committed rela-
tionship for at least two years. Respondents were paid $10.00 for com-
pleting a General Information Form and a modified version of the
Locke-Wallace (Locke & Wallace, 1959). Based on these data, 40 cou-
ples were invited to participate in the second phase of the study. To in-
sure a reasonable sampling of levels of relationship satisfaction, we
established a score of 115 or higher for the partner’s averaged Locke-
Wallace scores as indicating a “happy” couple and below 115 as indi-
cating an “unhappy” couple (the final sample consisted of 12 gay happy
couples, 10 gay unhappy couples, 10 lesbian happy couples, and 8 les-
bian unhappy couples. Other inclusion criteria were: (1) no more than a
Gottman et al. 73
10 year difference in ages between partners, (2) childless, (3) no previ-
ous committed (i.e., living together) heterosexual relationships, (4) dis-
crepancy between partners in modified Locke-Wallace relationship
scores of no more than 25 points, and (5) couple speaks English to one
another at home.
The second phase of the study consisted of filling out a larger battery
of questionnaires and coming to the Berkeley campus for three research
sessions in the laboratory. Both partners attended the first session to-
gether and each partner attended one additional session separately.
Each session lasted for two to three hours. Couples participating in this
second phase were paid $40.00 for completing the laboratory sessions.
Married Couple Sample. The comparison sample of married couples
was selected from a larger longitudinal study that recruited couples
from the environs of Bloomington, Indiana, beginning in 1983. Married
couples from this sample were matched in relationship satisfaction and
length of their relationship to the gay and lesbian samples. All couples
had participated in a marital interaction paradigm that included a dis-
cussion of a conflict area in the relationship (see below for details of this
procedure). Twenty married couples were selected where the husband
had presented the conflict issue, and 20 couples were selected where the
wife had presented the conflict issue. The disparity in geographical lo-
cation between the married couple sample and the gay and lesbian sam-
ples was also characteristic of Kurdek’s longitudinal study; in that study
the gay and lesbian samples were recruited nationally and the married
couple sample resided in Dayton, Ohio. Kurdek had the problem of how
to describe or assign same-sex partners to some condition such as the
roles of husband and wife. He randomly assigned same-sex partners to
“Partner 1” or “Partner 2,” and then re-analyzed his data two ways. For
our analysis we employed a psychological criterion to distinguish
among partners in all relationships. The partner who initiated and pre-
sented the major core relationship issue was called the “Initiator,” and
the other was called the “Partner.” This determination could be made in-
dependently by two observers with high reliability (100%). This created
a22 factorial design, with one factor being sexual preference (Ho-
mosexual/Heterosexual), and the other the sex of the initiator as either
male or female. We selected married couples from a longitudinal study
that had used procedures identical to those of the gay-lesbian study.
From the set of 79 couples we selected 20 in which the husband pre-
sented the marital issue to his wife and 20 in which the wife presented
the marital issue to the husband; these 40 couples were selected so that
they matched the gay-lesbian couples in age, relationship satisfaction,
74 JOURNAL OF HOMOSEXUALITY
and education and income. This selection produced the following 2 2
factorial experimental design:
The experimental design specifies two dimensions. The first dimen-
sion is who initiated the discussion of the relationship issue, a male or a
female. The second dimension is sexual orientation, whether the rela-
tionship is homosexual or heterosexual. This design makes it possible to
assess gender main effects (all males compared to all females) separate
from relationship-specific effects. Designs that do not include gay and
lesbian relationships confound gender with possible dominance effects.
Demographics. The lesbians were an average of 29.3 years old, and
the gay men were an average of 32.5 years old. The married couples for
which the women initiated the conflict discussion were an average of
28.7 years old, and the married couples for which the men initiated the
conflict discussion were an average of 29.6 years old. The mean
Locke-Wallace relationship satisfaction scores of the lesbians was
113.2, and the mean Locke-Wallace relationship satisfaction scores of
the gay men was 116.0. The mean Locke-Wallace relationship satisfac-
tion scores of the married couples for which the women initiated was
121.8, and mean Locke-Wallace relationship satisfaction scores of the
married couples for which the men initiated was 99.0.
Interaction Session. The procedures employed in this study were
modeled after those developed by Levenson and Gottman (1983) and
were used for both the heterosexual and homosexual couples. Couples
came to the laboratory after having not spoken to each other for at least
eight hours. After recording devices for obtaining physiological mea-
sures were attached, couples engaged in three conversations: (a) dis-
cussing the events of the day; (b) discussing an area of continuing
conflict and disagreement in their relationship; and (c) discussing a mu-
tually agreed upon pleasant topic. Each conversation lasted for 15 min-
utes and was preceded by a five-minute silent period. During the silent
Gottman et al. 75
Homosexual Heterosexual
Male Initiates Gay Male Married,
husband initiates
Female Initiates Lesbian Married,
wife initiates
periods and conversations, a broad sample of physiological measures
was obtained and a video recording was made of the interaction.
For the events of the day conversion, subjects were simply told to dis-
cuss what had happened during the day. Prior to initiating the conflict
area discussion, couples completed the Couple’s Problem Inventory
(Gottman, Markman & Notarius, 1977), in which they rated the per-
ceived severity of 10 relationship issues on a 0-to-100 scale. During the
conflict discussion partners were designated either “initiators” if they
were the one presenting the issue, or “partner” if they were the recipient
of the issue. Prior to initiating the pleasant topic discussion, couples
completed a similar inventory, in which they rated the enjoyment they
derived from 16 topics on a 0-to-100 scale. The experimenter used these
inventories to help couples select the topics that were used in these two
conversations.
For purposes of the present study, only data from the conflict area
discussion were utilized.
Observational Coding. The Specific Affect Coding System (SPAFF;
Gottman, McCoy, and Coan, 1996) was used to code the events of the
day and conflict area conversations of all couples. SPAFF focuses
solely on the specific affects expressed. The system draws on facial ex-
pression (based on Ekman and Friesen’s system of facial action coding;
Ekman & Friesen, 1978), vocal tone, and speech content to characterize
the emotions displayed. Coders categorized the affects displayed using
five positive affect codes (interest, validation, affection, humor, excite-
ment/joy), 10 negative affect codes (disgust, contempt, belligerence,
domineering, anger, fear/tension, defensiveness, whining, sadness,
stonewalling), and a neutral affect code. The dependent variables cre-
ated were the total number of seconds duration of each SPAFF code out
of the 900 seconds of the conflict area discussion. Every videotape was
coded in its entirety by two independent observers using a computer-as-
sisted coding system that automated the collection of timing informa-
tion; each coder noted only the onset of each code.
A time-locked confusion matrix for the entire videotape then was
computed using a 1-second overlap window for determining agreement
of each code in one observer’s coding against all of the other observer’s
coding (see Bakeman & Gottman, 1986). For the conflict segment, for
married couples the Cronbach alpha generalizability coefficients
summed over partners were: affection, .88; anger, .76; belligerence,
.89; contempt, .92; defensiveness, .99; disgust, .62; domineering, .96;
humor, .95; interest, .92; excitement/joy, .32; sadness, .82; stonewall-
ing, .64; fear/tension, .98; validation, .97; and whining, .86. For gays
76 JOURNAL OF HOMOSEXUALITY
and lesbians, for the conflict segment, the Cronbach alpha general-
izability coefficients summed over partners were: affection, .86; anger,
.86; belligerence, .91; contempt, .67; defensiveness, .97; disgust (a low
frequency code), .37; domineering, .84; humor, .96; interest, .75; ex-
citement/joy, .56; sadness, .72; stonewalling, .75; fear/tension, .95; val-
idation, .96; and whining, .81.
Weighting of the SPAFF Codes. For the mathematical modeling we
used a weighting scheme derived from previous prediction research
(Gottman, 1994). A numerical value was calculated for the SPAFF codes
for each 6-second time block separately for each partner by taking the sum
of positive codes minus the negative codes using the following weights:
Disgust = 3, Contempt = 4, Belligerence = 2, Domineering = 1, An-
ger = 1, Fear = 0, Defensiveness = 2, Whining = 1, Sadness = 1,
Stonewalling = 2, Neutral = 0.1, Interest = +2, Validation = +4, Affec-
tion = +4, Humor = +4, and Excitement/Joy = +4. This weighting yields
a potential score range of 24 to +24. For each couple this created two
time series, each with 150 data points, one series for the “initiator” and
one for the “partner.” For married couples, combined across partners,
the correlation for the two observers of total weighted SPAFF over the
conflict interaction was .90 (p < .001). For the gays and lesbians, com-
bined across partners, the correlation for the two observers of total
weighted SPAFF over the conflict interaction was .52 (p < .01).
Mathematical Modeling. A mathematical model using nonlinear dif-
ference equations was fit to the weighted SPAFF data during the con-
flict discussion for all couples. This model has been developed and
tested with a number of samples of married couples (Cook et al., 1995;
Gottman, Swanson, & Murray, 1999; Gottman, Murray, Swanson,
Tyson, & Swanson, in press). For this study there were two time-series,
the initiator’s and the partner’s. The initiator’s time series was denoted
Nt, and the partner’s time series was denoted Pt, where t is time. These
equations model both initiator and partner time series as follows:
Nt+1 = r1Nt+ a + IP-> N (Pt) + EPR R(KP) + EPD D(KP), and
Pt+1 = r2Pt+ b + IN-> P (Nt) + ENR R(KN) + END D(KN)
The parameters in these equations are defined in the following way. The
r’s are emotional inertia parameters that assess lag-one autocorrelation in
each series: r1assesses the lag-one autocorrelation for series Ntand r2as-
sesses the lag-one autocorrelation for series Pt. The a and b parameters
assess the initial uninfluenced states of the initiator and partner. Thus,
the (r1Nt+ a) term in the first equation and the (r2Pt+ b) term in the sec-
ond equation are the full autocorrelation terms. For those points when
Gottman et al. 77
only the autocorrelation functions (and there is no influence, repair or
damping), the uninfluenced steady states (neither N nor P are changing)
are computed from the equations:
N = r1N + a, or N = a/(1 r1), and
P = r2P + b, or P = b/(1 r2)
They are estimated using a least-squares fit.
The Is are the influence functions, with IP- > N (Pt) denoting the influ-
ence of the partner on the initiator, as a function of the partner’s behav-
ior, and IN- > P (Nt) denoting the influence of the initiator on the partner, as
a function of the initiator’s behavior. The influence functions in each
equation are the cross-correlation terms. They specify the average
amount of influence each partner has on the other at a particular level of
weighted positive-minus-negative affect (over time). The functional
form of the influence function we have found to fit our data the best is
the bilinear form (Figure 1). In the bilinear function there are two
slopes, the influence of positive and negative affect ranges on the part-
ner. There are, of course, two influence functions. Differences in slopes
are our operational definition of power in the relationship, which is de-
fined separately for positive and negative affect.
In nonlinear dynamic mathematical modeling (e.g., see Murray,
1989) these influence functions determine what are called the “null
clines” of the equations, which are the curves in the N-by-P phase space
where the initiator is steady and the partner is steady. Where the null
clines intersect determine the influenced “steady states” of the system
of equations. Steady states can be stable or unstable. If the steady states
78 JOURNAL OF HOMOSEXUALITY
INFLUENCE ON PARTNER
Negative Affect Positive Affect
FIGURE 1. Bilinear Influence Function
are stable, the influenced steady states are also called the “attractors” of
the dyadic system. They may be compared to gravitational attractors.
As noted previously, the system is drawn to these states, and, if per-
turbed slightly away from an attractor, the system is pulled back to the
attractor. There can be more than one attractor for a dyadic system. Of-
ten there are both negative and positive attractors for each couple. These
attractors determine how the system moves in the long run.
We recently modified this mathematical model so that it became pos-
sible to estimate repair terms that represents a couple’s attempts to im-
prove their communication when it becomes sufficiently negative. We
also defined analogous “damping” terms that represent a couple’s at-
tempts to modulate the positivity of their influence on one another when
the interaction becomes sufficiently positive, but we do not yet fully un-
derstand the theoretical meaning of damping. We viewed repair as a
communication strength and damping as a communication deficit.
Adding these terms to the model improved the fit of the model by an or-
der of magnitude. The EPR R(KP) term in the first equation and the ENR
R(KN) term in the second equation represent the effectiveness of repair
and the threshold trigger of negativity at which repair is activated. The
R(.) function is a standard trigger function that makes the time series
jump to a higher value with amplitude E (the effectiveness of repair) at
the trigger threshold of negativity, K. The EPD D(KP) term in the first
equation and the END D(KN) in the second equation represent the effec-
tiveness and thresholds of damping functions that act as a mirror of re-
pair, reducing the positive effects of positive affect after K, a threshold
trigger of positivity is crossed with effectiveness E. Details of the algo-
rithms for estimating each model parameter and the influence functions
are available in Gottman, Murray, Swanson, Tyson, and Swanson (in
press).
RESULTS
Uninfluenced Steady States. For the uninfluenced steady state of the
initiator of the issue, there was a significant effect only for sexual orien-
tation, F(1,75) = 8.18, p = .005. The homosexual mean was 0.61, and
the heterosexual was .98. There was no significant difference be-
tween gay and lesbian relationships, t(77) = 1.09, ns, nor between het-
erosexual couples in which the man or woman initiated, t(77) = .45. For
the uninfluenced steady state of the partner, there was again only a sig-
nificant effect for sexual orientation, F(1,75) = 6.45, p = .013. The ho-
Gottman et al. 79
mosexual mean was 0.37, and the heterosexual mean was 1.09. There
was again no significant difference between gay and lesbian relation-
ships, t(77) = 1.37, ns, nor between heterosexual couples in which the
man or woman initiated, t(77) = .89. Thus, this means that the way the
issue is presented and received in the conflict interaction is positive for
homosexual couples and negative for heterosexual couples. These re-
sults suggest that, by analyzing observational data, homosexual rela-
tionships may be fundamentally different from heterosexual
relationships.
Subsequent Analyses of Specific SPAFF Codes. These analyses
showed that homosexual initiators of the conflict issue compared to
heterosexual initiators were characterized by less negative affect: less
belligerence, t(76) = 2.72, p < .01 (homosexual mean = 2.65, heterosex-
ual mean = 10.57); less domineering, t(76) = 2.38, p < .05 (homosexual
mean = 7.56, heterosexual mean = 33.18); less fear/tension, t(76) =
4.02, p < .001 (homosexual mean = 21.52, heterosexual mean =
121.76); less sadness, (t(76) = 3.89, p < .05 (homosexual mean = 6.87,
heterosexual mean = 30.21); less whining, t(76) = 1.97, p < .05 (homo-
sexual mean = 2.30, heterosexual mean = 10.29). The homosexual initi-
ators of the conflict also demonstrated more positive emotions when
compared with the heterosexual initiators: more affection, t(76) = 1.75,
p < .05 (homosexual mean = 1.82, heterosexual mean = .89); more hu-
mor, t(76) = 3.91, p < .001 (homosexual mean = 29.61, heterosexual
mean = 9.45); and more joy/excitement, t(76) = 2.34, p < .05 (homosex-
ual mean = .46, heterosexual mean = .007). For the partner (i.e., the re-
cipient of the conflict issue), homosexual compared to heterosexual
partners showed less negative affect: less belligerence, t(76) = 2.28, p <
.05 (homosexual mean = 1.61, heterosexual mean = 5.96); less domi-
neering, t(76) = 2.52, p < .05 (homosexual mean = 9.06, heterosexual
mean = 38.88); less fear/tension, t(76) = 7.60, p < .001 (homosexual
mean = 15.48, heterosexual mean = 94.71). The homosexual partners
also showed significantly more humor than the heterosexual partners,
t(76) = 3.86, p < .001 (homosexual mean = 29.71, heterosexual mean =
9.38). These results suggest that homosexual relationships may be dis-
tinguished from heterosexual relationships in the expression of specific
positive and negative affects during a conflict interaction.
Influenced Steady States. For the influenced steady states, we com-
puted the number of stable steady states in each of the four quadrants of
phase space. The only significant difference was in the quadrant where
both of their behavior is positive, and again there was a sexual orienta-
tion effect, F(1,75) = 5.25, p = .025. The mean number of influenced
80 JOURNAL OF HOMOSEXUALITY
stable steady states in this quadrant for the homosexual couples was .86
versus .32 for the heterosexual couples. There was also a marginally
significant interaction between sexual orientation and the gender of the
initiator, F(1,75) = 3.89, p = .052. Subsequent tests showed that the les-
bian mean of 1.14 stable steady states in the positive-positive quadrant
was significantly higher than the other three types of couples (gay male
mean = .57, man initiating to a woman = .50, woman initiating to a man =
.19). Once again, these results suggest that influence processes in ho-
mosexual committed relationships may be dissimilar to the influence
processes found in heterosexual couples. These findings imply that les-
bian couples, when compared with the gay and heterosexual couples,
are more likely to remain stable in their behavioral interactions when
both partners are positive in their communication content.
Repair Effectiveness. There were no significant main effects due to
sexual orientation or to who initiates the issue, for either the initiator’s
or the partner’s repair effectiveness. However, subsequent t-tests re-
vealed some interesting findings on partner’s repair effectiveness. The
mean was lower for gay male couples than in any of the other 3 groups
(gay mean = 2.78, man initiated to a woman mean = 3.98, woman initi-
ated to a man mean = 3.99, lesbian mean = 3.65). There was a signifi-
cant difference between homosexual and heterosexual couples where
the man initiates, t(69) = 1.70, p < .05. There was also a significant dif-
ference between a woman initiating to a man (the most common occur-
rence in heterosexual couples) and a man initiating to a man in a gay
couple, t(69) = 1.80, p < .05. These results suggest the following inter-
pretation of the data: when a gay man presents the conflict issue to his
partner and the initiator becomes too negative, his partner may not re-
pair very effectively. This finding suggests that interventions for gay
male couples in particular may do well focusing on the repair aspects of
conflict regulation.
Repair Threshold. There were no significant effects for the initia-
tor’s negative score threshold that triggers repair by their partner. How-
ever, there was a significant effect for the partner’s threshold that
triggered repair, and in this case the effect was due to who brought up
the issue to be discussed, F(1,69) = 5.14, p = .027. Recall that the thresh-
old for negativity is how negative the initiator’s behavior needs to get
for the partner to start trying to repair. The mean threshold of negativity
when a man initiated the issue was 4.64, compared to a mean of
6.04 when a woman initiated the issue. This means that when a man
initiates the issue, he detects negativity in his partner and attempts to re-
pair sooner (that is, before it gets too negative) compared to a woman
Gottman et al. 81
when she is presenting the issue. In addition, there was a marginally sig-
nificant interaction effect, F(1,69) = 3.03, p = .086. A subsequent t-test
showed that this was due to the fact that when a heterosexual man initi-
ates the issue, he notices and responds to negativity in his partner at a
much less negative threshold than in any other group, t(69) = 2.48, p <
.01.
Damping Effectiveness. There was a significant effect for the initia-
tor’s effectiveness of damping, and the effect was due to sexual orienta-
tion, F(1,73) = 4.98, p = .029. The mean initiator’s effectiveness of
damping for homosexual couples was 4.85, as compared to 3.11 for
heterosexual couples. It is the magnitude and not the sign of this mea-
sure that is important (it is expected to be a negative number precisely in
order to have a damping effect on the underlying influence function). In
addition, there was a marginally significant effect for the partner’s ef-
fectiveness of damping, and the effect was also due to sexual orienta-
tion, F(1,70) = 3.91, p = .052. The mean partner’s effectiveness of
damping for homosexual couples was 5.03, as compared to 3.37 for
heterosexual couples. Thus, when either partner in the homosexual cou-
ples detected that the other person had become “too positive” (perhaps
treating a serious issue with too much levity), they were more effective
at damping this down than were the heterosexual couples. Damping is a
decrease in positive influence; it is not a dysfunctional part of the model
because we have discovered that damping may create a positive-posi-
tive stable steady state (“attractor”) where none existed before.
Affect and Influence Patterns. In every study we have conducted with
heterosexual couples using the mathematical modeling, the slope of the
influence functions in the negative affect range is steeper than the slope
in the positive affect range. Gottman (1994) referred to this as “the tri-
umph of negative over positive affect,” meaning that it is easier to hurt
one’s partner with negative affect than it is to have a positive influence
with positive affect. In the present study when we compared homosex-
ual relationships with heterosexual relationships we found some indica-
tion that for homosexuals there was a reversal of the pattern. The
variable used in this analysis was the slope for positive affect minus
slope for negative affect, so a negative value indicates the negative af-
fect slope is steeper (i.e., the usual finding). For the initiator’s influence
function there was a marginal interaction effect, F(1,66) = 3.05, p = .085
(man initiates to woman mean = .086, woman initiates to man mean =
.283, gay male mean = .331, lesbian mean = .122). The means
suggested a difference between the heterosexual men initiating to their
wives and all other groups, but a subsequent t-test did not find this to be
82 JOURNAL OF HOMOSEXUALITY
significant. However, for the partner’s influence function there was a
marginally significant effect for sexual orientation, F(1,68) = 2.83, p =
.097 (heterosexual mean = .19, homosexual mean = .002). Hence,
with increased power we may find a lessening of the effect of the tri-
umph of negative over positive affect in homosexual couples.
Gender Effects. In addition to the gender threshold effect for the re-
pair trigger previously noted, several other significant main effects for
gender were found. Regardless of sexual orientation, men were more
angry then women when presenting an issue, t(77) = 1.75, p < .05 (fe-
male mean = 1.55, male mean = 3.99) and women were more ex-
cited/joyful than men, t(77) = 2.86, p < .003 (female mean = .49, male
mean = .004). For the partner, regardless of sexual orientation, women
were more sad when receiving a conflict issue than men, F(1,76) = 6.03,
p = .016 (female partner mean = 16.64, male partner = 2.06).
Differences Between Gays and Lesbians. There was inadequate
power for the math modeling comparisons. However, there was enough
power to find significant differences in affect. The SPAFF coding re-
vealed, for the initiator of the issue, lesbians were more angry than gays,
t(76) = 1.66, p < .05 (gay mean = 4.97, lesbian mean = 15.10), used
more humor, t(76) = 2.15, p < .05 (gay mean = 21.88, lesbian mean =
37.33), and showed more excitement/joy, t(76) = 3.55, p < .01 (gay
mean = .007, lesbian mean = .86). For the partner, lesbians showed more
humor, t(76) = 1.76, p < .05 (gay mean = 23.29, lesbian mean = 36.14),
and lesbians showed more interest, t(76) = 1.95, p < .05 (gay mean = to
1.24, lesbian mean = 6.00). These results suggest that lesbians are more
emotionally expressive than gay men.
DISCUSSION
Because the mathematical modeling is unfamiliar to many readers,
we would like to take the reader through a nontechnical discussion of
the meaning of our findings. First, the model is able to decompose the
observational data into uninfluenced and influenced components.
Power, or influence, was defined in our modeling as the ability of one
person’s affect to move the other partner’s affect. These two compo-
nents may be thought of as (1) what each person brings into the interac-
tion at its start, and (2) how each partner influences that start value as
they talk. Using this particular experimental design (orientation by initi-
ator), we were able to determine, from an examination of the uninflu-
enced steady states that in homosexual relationships the initiator of the
Gottman et al. 83
conflict started positively, while in heterosexual relationships the initia-
tor started negatively. Other research suggests that women presenting
conflict issues to men is the most common pattern in heterosexual cou-
ples (Ball, Cowan, & Cowan, 1995; Oggins, Veroff, & Leber, 1993).
This gender pattern fits with the well-known female demand-male with-
draw pattern identified by Christensen and colleagues (e.g., Christensen
and Heavey, 1990). However, in our design we were able to include
male as well as female heterosexual initiators.
This homosexual/heterosexual pattern was echoed by the way the
partner received the issue. Again, the homosexual uninfluenced partner
affect mean was positive, while the heterosexual mean was negative.
Thus, this means that the way the issue is presented and received in the
conflict interaction is positive for homosexual couples and negative for
heterosexual couples. These results suggest that, by analyzing observa-
tional data, homosexual relationships may be fundamentally different
from heterosexual relationships.
This issue of “startup” is a critical issue in the management of con-
flict. It has made it possible for our laboratory to be able to predict di-
vorce in married couples from just the first three minutes of a conflict
discussion (Carrère et al., 2000).
A subsequent analysis of what specific affects contribute to this ef-
fect show that homosexual initiators were less belligerent and less dom-
ineering than heterosexual initiators. These findings are important in
light of the heightened sensitivity to equity in homosexual compared to
heterosexual relationships (Kurdek, 1998). For the partner (i.e., the re-
cipient of the conflict issue), homosexual compared to heterosexual
partners also showed less belligerence, less domineering, and less
fear/tension. Clearly the results suggest that not only is equity a greater
concern in homosexual relationships, but they behave in accordance
with these concerns in the way conflict is initiated. Consistent with
these results are the findings that there is less fear/tension, less sadness,
less whining in homosexual initiators that heterosexual initiators.
But the effects are not only the inhibition of negative affects in
startup of a conflict discussion. The data show that the homosexual ini-
tiators of the conflict also demonstrated more positive emotions when
compared with the heterosexual initiators: more affection, more humor,
and more joy/excitement. The homosexual partners also showed signif-
icantly more humor than the heterosexual partners.
Now let us consider the results about the influence process itself. For
the influenced steady states, we computed the number of stable steady
states in each of the four quadrants of phase space. These are called the
84 JOURNAL OF HOMOSEXUALITY
“attractors” of the interaction. The presence of an attractor means that
the interaction will tend toward that state, much as is the case for gravi-
tational attraction. We found that the only significant difference in the
attractors was in the quadrant where both partner’s behavior is positive,
and again there was a sexual orientation effect. The mean number of
positive-positive attractors for the homosexual couples suggested that
they tended to have an attractor in this quadrant, whereas this was far
less likely for the heterosexual couples (the means were .86 versus .32,
respectively). This was especially true for lesbians. Once again, using
this experimental design, these results suggest that influence processes
in homosexual committed relationships may be dissimilar to the influ-
ence processes found in heterosexual couples.
The results on repair were: for the partner’s repair effectiveness, gay
couples were less effective than all the other dyads in the study. These
results suggested that when a gay man presents the conflict issue to his
partner and the initiator becomes too negative, his partner may not re-
pair very effectively. This finding suggests that interventions for gay
male couples in particular may do well focusing on the repair aspects of
conflict regulation.
The other part of repair is the threshold at which repair attempts be-
gin. Recall that the threshold for negativity is how negative the initia-
tor’s behavior needs to get for the partner to start trying to repair. We
found that when a man initiates the issue, he detects negativity in his
partner and attempts to repair sooner (that is, before it gets too negative)
compared to a woman when she is presenting the issue; this is true re-
gardless of sexual orientation. It is particularly true that when a hetero-
sexual man initiates the issue, he notices and responds to negativity in
his partner at a much less negative threshold than any other group.
Damping is the converse of repair. It involves damping down the influ-
ence of one’s partner’s positive, rather than negative affect. Damping is a
decrease in positive influence; it is not a dysfunctional part of the model
because we have discovered that damping may create a positive-posi-
tive stable steady state (“attractor”) where none existed before. As far as
the initiator’s damping was concerned, homosexual couples were more
effective than heterosexual couples. In addition, there was a marginally
significant effect for the partner’s effectiveness of damping, and the ef-
fect was also due to sexual orientation. Thus, when either partner in the
homosexual couples detected that the other person had become “too
positive” (perhaps treating a serious issue with too much levity), they
were more effective at damping this down than were the heterosexual
couples.
Gottman et al. 85
Perhaps most exciting in the results is a suggestive rather than a de-
finitive result. As we noted, in every study we have conducted with het-
erosexual couples using the mathematical modeling, the slope of the
influence functions in the negative affect range is steeper than the slope
in the positive affect range. We call this “the triumph of negative over
positive affect” (Gottman, 1994), meaning that it is easier to hurt one’s
partner with negative affect than it is to have a positive influence with
positive affect. In the present study when we compared homosexual re-
lationships with heterosexual relationships we found some indication
that for homosexuals there was a reversal of the pattern. The variable
used in this analysis was the slope for positive affect minus slope for
negative affect, so a negative value indicates the negative affect slope is
steeper (i.e., the usual finding). For the partner’s influence function
there was a marginally significant effect for sexual orientation, and we
speculate that with increased power we may find a lessening of the ef-
fect of the triumph of negative over positive affect in homosexual cou-
ples.
The current design is ideal for examining gender effects, independent
of sexual orientation. In addition to the gender threshold effect for the
repair trigger previously noted, we found only that regardless of sexual
orientation, men were more angry then women when presenting an is-
sue, while women were more excited/joyful than men. Also, for the
partner, regardless of sexual orientation, women were more sad when
receiving a conflict issue than men.
We had very little power to effectively study differences between
gays and lesbians. The SPAFF coding revealed, that, for the initiator of
the issue, lesbians were more angry than gays, used more humor, and
showed more excitement/joy. For the partner, lesbians showed more
humor, and lesbians showed more interest than gays. These results sug-
gest that lesbians are more emotionally expressive, of both negative and
positive affect than gay men.
Previous questionnaire and interview studies have suggested that
committed gay and lesbian relationships are not fundamentally differ-
ent than committed married heterosexual relationships in terms of such
factors as the relationship between costs/benefits and relationship satis-
faction and the determinants of the progress toward commitment. In
fact, in a previous paper on the correlates of relationship satisfaction
and stability in gay and lesbian relationships we also concluded that the
correlates of relationship satisfaction do seem to be quite similar in het-
erosexual and homosexual committed relationships.
86 JOURNAL OF HOMOSEXUALITY
However, to summarize our findings, when we employed a 2 2
factorial study to compare the means of observed interaction between
homosexual and heterosexual groups, accounting for who initiates the
conflict issue, a different picture emerged. The mathematical modeling
revealed interesting differences between heterosexual and homosexual
couples in interactional dynamics when couples discussed areas of con-
flict in their relationships. In analyses of the uninfluenced steady states,
we found that homosexual couples began far more positively and far
less negatively in the way they presented an issue than heterosexual
couples. Homosexual couples were also more positive in the way they
received an issue from their partner than were heterosexual couples.
Then, after the social influence process proceeded, homosexual couples
were more likely to maintain a positive influenced steady state than
were heterosexual couples. Finally, homosexual couples were more
likely to have influenced states in the positive-positive quadrant of
phase space than heterosexual relationships.
Furthermore, our results suggest that gay and lesbian relationships
may operate on different principles than heterosexual relationships with
respect to power and affect. Although there was inadequate power to
test these results, the pattern of results suggest that homosexual couples
were more positive in their influence on the partner in the positive affect
ranges and less negative in their influence on the partner in the negative
affect ranges than were heterosexual couples.
This observational approach to studying gay and lesbian relation-
ships is important in its own right to determine the correlates (and even-
tually, the causes) of being able to maintain satisfying long-term
committed homosexual relationships. Other research has shown that the
uninfluenced steady state in and of itself is a significant predictor of di-
vorce in heterosexual married couples (Cook et al., 1995; Gottman et
al., 1998). Thus, based on our results, heterosexual relationships may have
a great deal to learn from homosexual relationships insofar as homosexual
relationships seem to have found a way to begin conflict discussions in a
more positive and less negative manner, and to continue to have a positive
rather than a negative influence on one another. Additional observational
research is clearly called for to study the correlates of successful gay and
lesbian relationships and parameters of interaction that are indicative of
long-term stability and happiness in these relationships.
Why would gay and lesbian committed relationships differ so much
from heterosexual couples both in the way they present and receive an
area of continuing disagreement and in their influence patterns? And
why would they differ by being so much more positive and less negative
Gottman et al. 87
than heterosexual couples? We can only speculate about potential differ-
ences. We suspect that it has to do with two facts: (1) homosexual couples
value equality far more than heterosexual couples, and (2) that there are
fewer barriers to leaving homosexual than heterosexual relationships. The
greater negativity and lowered positivity of heterosexual couples may have
to do with the standard status hierarchy between men and women, a pattern
that research has shown is largely absent in same-sex couples. It is well
known that the status hierarchy in heterosexual relationships breeds hostil-
ity, particularly from women, who tend to have less power than men, and
who also typically bring up most of the relationship issues. Because there
are fewer barriers to leaving homosexual compared to heterosexual rela-
tionships, homosexual couples may be more careful in the way they accept
influence from one another. Thus, we suggest that the process variables by
which they resolve conflicts may be the very glue that keeps these relation-
ships stable. Potentially, the reverse dynamic of the triumph of positive
over negative affect that may characterize homosexual versus heterosexual
relationships is a very exciting prospect.
Our findings on the lowered effectiveness of repair when the interac-
tion does become negative is interesting, particularly for gay men. It
suggests that interventions designed to help keep gay male relationships
stable be focused on processes that may operate to harm repair pro-
cesses. The most probable such process is physiological reactivity,
which our research with marriages has shown is more of an issue for
men than it is for women (Gottman & Levenson, 1988).
Subsequent research needs to expand the sample size of this investi-
gation and to move beyond correlational data to experiments that at-
tempt to improve relationships. These experiments are currently in the
planning stages in our laboratories.
REFERENCES
Andriote, J.M. (1999).Victory deferred: How AIDS changed gay life in America. Chi-
cago: University of Chicago Press.
Arnold, V.I. (1986). Catastrophe theory. Berlin: Springer-Verlag.
Bakeman, R., & Gottman, J. (1986). Observing Interaction: An Introduction to Se-
quential Analysis. New York: Cambridge University Press.
Baker, G.L., & Gollub, J.P. (1996). Chaotic dynamics: An introduction. New York:
Cambridge University Press.
Ball, F.L.J., Cowan, P., & Cowan, C.P. (1995). Who’s got the power? Gender differ-
ences in partner’s perception of influence during marital problem-solving discus-
sions. Family Process, 34, 303-321.
88 JOURNAL OF HOMOSEXUALITY
Bateson, G., Jackson, D.D., Haley, J., & Weakland, J. (1956). Toward a theory of
schizophrenia. Behavioral Science, 1, 251-264.
Bell, A.P., & Weinberg, M.S. (1978). Homosexualities: A study of diversity among
men and women. New York: Simon & Schuster.
Beltrami, E. (1993). Mathematical models in the social and biological sciences.
Boston, MA: Jones and Bartlett.
Berge, P., Pomeau, Y., & Vidal, C. (1984). Order within chaos: Toward a determinis-
tic approach to turbulence. New York: Wiley Interscience.
Blumstein, P., & Schwartz, P. (1983). American couples: Money, work, sex. New
York: William Morrow and Co.
Brauer, F., & Nohel, J.A. (1969). The qualitative theory of ordinary differential equa-
tions. New York: Dover.
Brazelton, T.B., Koslowski, B., & Main, M. (1974). The origins of reciprocity: The
early mother-infant interaction. In M. Lewis & L. Rosenbum (Eds.), The effect of
the infant on its caregiver (pp. 49-76). New York: Wiley.
Castrigiano, D.P.L., & Hayes, S.A. (1993). Catastrophe theory. New York: Addi-
son-Wesley.
Christensen, A., & Heavey, C.L. (1990). Gender and social structure in the de-
mand/withdraw pattern of marital conflict. Journal of Personality and Social Psy-
chology, 59, 73-82.
Cook, J., Tyson, R., White, J., Rushe, R., Gottman, J., and Murray, J. (1995). The
mathematics of marital conflict: Qualitative dynamic mathematical modeling of
marital interaction. Journal of Family Psychology,9, 110-130.
Duffy, S.M., & Rusbult, C.E. (1986). Satisfaction and commitment in homosexual and
heterosexual relationships. Journal of Homosexuality, 12, 1-23.
Ekman, P., & Friesen W.V. (l978). Facial Action Coding System. Palo Alto, CA: Con-
sulting Psychologist Press.
Gianino, A., & Tronick, E.Z. (1988). The mutual regulation model: The infant’s self
and interactive regulation and coping and defensive capacities. In T.M. Field, P.M.
McCabe, & N. Schneiderman (Eds.), Stress and coping across development
(pp. 47-70). Hillsdale, NJ: Lawrence Erlbaum Associates.
Gilmore, R. (1981). Catastrophe theory for scientists and engineers. New York: Do-
ver.
Glass, L., & Mackey, M.C. (1988). From clocks to chaos: The rhythms of life. Prince-
ton, NJ: Princeton University Press.
Gleick, J. 1987). Chaos: Making a new science. New York: Viking.
Goldberg, S. (1986). Introduction to difference equations: With illustrative examples
from economics, psychology, and sociology. New York: Dover.
Gottman, J., Murray, J., Swanson, C., Tyson, R., & Swanson, K. (in press). The mathe-
matics of marriage: Formal models. Cambridge, MA: MIT Press.
Gottman, J.M. (1999). The marriage clinic. New York: W.W. Norton.
Gottman, J.M. (1994). What predicts divorce? Hillsdale, NJ: Lawrence Erlbaum.
Gottman, J.M. (1993). The roles of conflict engagement, escalation, or avoidance in
marital interaction: A longitudinal view of five types of couples. Journal of Con-
sulting and Clinical Psychology, 61, 6-15.
Gottman et al. 89
Gottman, J. (1979). Marital interaction: Experimental investigations. New York: Aca-
demic Press.
Gottman, J.M., McCoy, K., Coan, J., & Collier, H. (1996). The Specific Affect Coding
System (SPAFF). In J.M. Gottman (Ed.), What Predicts Divorce: The Measures.
Lawrence Erlbaum Associates.
Gottman, J.M., & Levenson, R.W. (1992). Marital processes predictive of later disso-
lution: Behavior, physiology, and health. Journal of Personality and Social Psy-
chology, 63, 221-233.
Gottman, J., Markman, J., & Notarius, C. (1977). The topography of marital conflict: A
sequential analysis of verbal and nonverbal behavior. Journal of Marriage and the
Family, 39, 461-477.
Gottman, J., Coan, J., Swanson, C., & Carrère, S. (1998). Predicting marital happiness
and stability from newlywed interactions. Journal of Marriage and the Family, 60,
5-22.
Gray-Little, B., & Burks, N. (1983). Power and satisfaction in marriage: A review and
critique. Psychological Bulletin, 93, 513-538.
Harry, J. (1983). Gay male and lesbian relationships. In E.D. Macklin and R.H. Rubin
(Eds.), Contemporary families and alternative lifestyles (pp. 216-234). Beverly
Hills, CA: Sage Publications.
Jacobson, N.S., Gottman, J.M., Gortner, E., Berns, S., & Shortt, J.W. (1996). Psycho-
logical factors in the longitudinal course of battering: When do the couples split up?
When does the abuse decrease? Violence and Victims, 11, 371-392.
Jordan, D.W., & Smith, P. (1987). Nonlinear ordinary differential equations (Second
edition). Oxford: Clarendon Press of Oxford University Press.
Kelley, W.G., & Peterson, A.C. (1991). Difference equations: An introduction with ap-
plications. New York: Academic Press.
Krokoff, L.J., Gottman, J.M., & Haas, S.D. (1989). Validation of a rapid couples inter-
action scoring system. Behavioral Assessment, 11, 65-79.
Kurdek, L.A. (1998). Relationship outcomes and their predictors: Longitudinal evi-
dence from heterosexual married, gay cohabiting, and lesbian cohabiting couples.
Journal of Marriage and the Family, 60, 553-568.
Kurdek, L.A., & Schmitt, J.P. (1986). Relationship quality of partners in heterosexual
married, heterosexual cohabiting, and gay and lesbian relationships. Journal of
Family Psychology, 51, 711-720.
Levenson, R.W., & Gottman, J.M. (1983). Marutal interaction: Physiological linkage
and affective exchange. Journal of Personality and Social Psychology, 45,
587-597.
Lorenz, E. (1993). The essence of chaos. Seattle, WA: University of Washington Press.
Morrison, F. (1991). The art of modeling dynamical systems: Forecasting for chaos,
randomness, & determinism. New York: Wiley Interscience.
Murray, J.D. (1989). Mathematical biology. Berlin: Springer-Verlag.
Oggins, J., Veroff, J., & Leber, D. (1993). Perceptions of marital interaction among
Black and White newlyweds. Journal of Personality and Social Psychology, 65,
494-511.
Peters, E.E. (1991). Chaos and order in the capital markets. New York: John Wiley.
90 JOURNAL OF HOMOSEXUALITY
Robinson, E.A., & Price, M.G. (1980). Pleasurable behavior in marital interaction: An
observational study. Journal of Consulting and Clinical Psychology, 48, 117-118.
Rosenblatt, P.C. (1994). Metaphors of family systems theory. New York: Guilford.
Saunders, P.T. (1990). An introduction to catastrophe theory. Cambridge, UK: Cam-
bridge University Press.
Vallacher, R.R., & Nowack, A. (1994). Dynamical systems in social psychology. New
York: Academic Press.
VonBertalanffy, L. (1968). General System Theory. New York: George Braziler.
Weinberg, M.S. & Williams, C.J. (1974). Male homosexuals: Their problems and ad-
aptations. New York: Oxford University Press.
Wiggins, J. (1977). Personality and prediction. New York: Addison Wesley.
Winfree, A.T. (1990). The geometry of biological time. Berlin: Springer-Verlag.
Gottman et al. 91
For FACULTY/PROFESSIONALS with journal subscription
recommendation authority for their institutional library . . .
Please send me a complimentary sample of this journal:
(please write complete journal title here–do not leave blank)
If you have read a reprint or photocopy of this article, would you like to
make sure that your library also subscribes to this journal? If you have
the authority to recommend subscriptions to your library, we will send you
a free complete (print edition) sample copy for review with your librarian.
1. Fill out the form below and make sure that you type or write out clearly both the name
of the journal and your own name and address. Or send your request via e-mail to
docdelivery@haworthpress.com including in the subject line “Sample Copy Request”
and the title of this journal.
2. Make sure to include your name and complete postal mailing address as well as your
institutional/agency library name in the text of your e-mail.
[Please note: we cannot mail specific journal samples, such as the issue in which a specific article appears.
Sample issues are provided with the hope that you might review a possible subscription/e-subscription with
your institution's librarian. There is no charge for an institution/campus-wide electronic subscr iption
concurrent with the archival print edition subscr iption.]
I will show this journal to our institutional or agency librar y for a possible subscription.
Institution/Agency Library: ______________________________________________
Name: _____________________________________________________________
Institution: __________________________________________________________
Address: ___________________________________________________________
City: ____________________
Return to: Sample Copy Department,The Haworth Press, Inc.,
10 Alice Street, Binghamton, NY 13904- 1580
State: __________ Zip: ____________________
... This lack of research means that we have a deficit in our understanding of the shared and unique strengths and challenges that SGD individuals experience with sexual communication. Research on non-sexual communication with couples suggests that same-gender/sex couples communicate more positively about global relationship issues than man-woman couples [80,81] and that women-women couples may value and exhibit more emotional expression than man-woman couples [80,82]. ...
... This lack of research means that we have a deficit in our understanding of the shared and unique strengths and challenges that SGD individuals experience with sexual communication. Research on non-sexual communication with couples suggests that same-gender/sex couples communicate more positively about global relationship issues than man-woman couples [80,81] and that women-women couples may value and exhibit more emotional expression than man-woman couples [80,82]. ...
Article
Full-text available
Purpose of Review The purpose of this review is to summarize the current knowledge on sexual communication among sexual and gender/sex diverse (SGD) groups. Complementing an existing review of the literature on safer-sex communication with SGD individuals (Parrillo & Brown, 2021), we focus on sexual communication related to promoting sexual satisfaction. Recent Findings The two-pathways model of sexual communication has yet to be generalized with SGD samples. Research comparing SGD with non-SGD individuals has varied in whether there are differences between groups. There is some evidence of differences between gender diverse and non-gender diverse groups in sexual communication. Emerging evidence of the unique strengths and challenges of sexual communication among gender/sex diverse groups highlights the importance of deepening gender/sex diverse-specific sexual communications research. Summary A lack of literature regarding sexual communication in SGD groups is reported. Results on whether there are differences between and/or within groups are mixed and confounded by inconsistent methodologies for measurement of demographic and sexual communication variables. Clearly, further research is needed to increase our understanding of sexual communication in SGD groups. As such, we provide recommendations for future research, specifically regarding inclusive demographic and analytical considerations.
... Future studies should include more diverse, nonbinary gender and relationship options in order to paint a more inclusive picture couples transitioning to parenthood. Given evidence that same-sex partnerships navigate conflicts with less domineering and belligerent behavior and more positive affect (Gottman et al., 2003), samesex couples may remain more supportive across the TTP in a way that mixed-sex couples fail to do and insights may be extracted from these partnerships to benefit all couples. Fourth, the selection of classes in combination with dropout of participants led to some rather small classes. ...
Article
How parents cope with stress as a couple (i.e., dyadic coping [DC]) is related to mental health problems in children. But little is known about DC within first-time parents and child mental health problems in early childhood. This study investigated subgroups in DC trajectories across the transition to parenthood (TTP) and examined subgroup differences in child mental health problems. Mothers' and fathers' self-report of positive and negative DC (n = 288 couples) at seven points of measurement (27th, 32nd week of pregnancy, 2nd, 14th, 40th week postpartum, 3- and 4-year postpartum) and children's emotional and behavioral problems from parent report (4-year postpartum) were used. Latent class growth analyses revealed that over half of the couples experienced a moderate decline in positive DC across the TTP (58%), whereas only fathers reported a decline among the remaining couples (42%). Fathers with a partner who maintained their level of positive DC reported more child emotional and behavioral problems than fathers whose partners' DC also decreased. Results for negative DC indicated two subgroups in which one partner maintained their initial level of negative DC (stable fathers: 10%, stable mothers: 23%), while the other increased. In most couples, both parents increased their negative DC (67%). Fathers reported more child emotional and behavioral problems if their negative DC increased across the TTP than if their negative DC remained stable regardless of the negative DC of their partner. The existence of different DC trajectory patterns needs to be considered in further research as well as prevention. (PsycInfo Database Record (c) 2022 APA, all rights reserved).
... Additionally, couples of diverse gender identities and sexual orientations face higher levels of stress due to living in heteronormative societal structures, which activates the attachment system, placing pressure on the attachment bond (Cook & Calebs, 2016). Notably, couples diverse in gender identity and sexuality appear to have particular strengths that may facilitate the therapeutic process in EFT, including responsiveness to partner influence, emphasis on equality, and growth experiences in the process of self-affirmation and expression that may expand and strengthen models of self and other (Gottman et al., 2003). As a model, EFT is well positioned to help diverse couples navigate the minority stress and buffer the effects of experiences with prejudice and discrimination these couples face, drawing on the resilience inherent in a secure attachment bond (Allan & Johnson, 2017). ...
Chapter
Emotionally focused couple therapy (EFT) has contributed substantially to the field of couple interventions. In particular, it has led the way in developing interventions that change emotion regulation and responses in ways that lead to increased emotional responsiveness and bonding interactions.
... Разрешение конфликтов между партнерами одного пола, которые и мир видят в относительно схожем ракурсе, оказывается более легким. Некоторые исследователи даже делали предположение, что отношения гомосексуальных и гетеросексуальных партнеров могут работать на разных принципах 20 . Дело в том, что в конфликте внутри пары, состоящей из мужчины и женщины, инициатор занимает позицию с выдающимися элементами властности и влияния: один партнер предъявляет требование, а другой -отступает. ...
Article
Full-text available
Интимная гомосексуальная пара — явление относительно новое. Как сложились практики постоянного партнерства, сожительства и брака в однополых союзах?
... Cependant, les couples de même sexe (notamment les couples lesbiens) semblent avoir un bon équilibre concernant le partage des responsabilités parentales, car ils échappent des comportements stéréotypés de genre 1 [21,22]. Les couples homosexuels résolvent et gèrent plus efficacement leur conflit, sont moins agressifs et utilisent davantage l'humour et compromis, ce qui contribue à améliorer la qualité de la relation de couple [23,24]. Les parents homosexuels sont suffisamment capables d'assurer un environnement familial stable et approprié pour le bon développe- ment de leurs enfants [25]. ...
Article
Résumé En France, plus de trois millions de personnes se revendiquent homosexuelles ou bisexuelles et de 200 à 300 000 enfants auraient un parent homosexuel. L’homoparentalité est au centre de nombreux débats sociopolitiques. On repère également une intensification de la recherche scientifique, dans ce domaine, au cours de ces dernières décennies. Cette revue systématique de la littérature visait deux principaux objectifs : 1) identifier les effets de l’homophobie intériorisée sur la santé mentale et l’adaptation sociale des parents homosexuels ; 2) identifier les effets de l’homophobie intériorisée sur le développement psychologique et social des enfants de parents homosexuels. La recherche a été effectuée sur un total de 87 840 publications scientifiques, portant sur l’homoparentalité, extraites des bases de données scientifiques informatisées : SciELO, PubMed, PsycINFO, SCOPUS, Medline, ProQuest et Web of Science. Dans cette étude, 79 études empiriques et expérimentales, publiées entre 1971 et 2020, ont été retenues. L’analyse des données montre que l’homophobie intériorisée est à l’origine de perturbations psychologiques chez les parents homosexuels et chez leurs enfants. L’orientation sexuelle des parents n’a pas d’effets négatifs, sur le développement cognitif et émotionnel de l’enfant, ni sur son orientation sexuelle. L’intériorisation de l’homophobie est un obstacle à l’adaptation sociale et à la santé perçue des familles homosexuelles. De nouvelles études, dans ce domaine, sont nécessaires afin de mieux identifier les facteurs qui favorisent l’intériorisation de l’homophobie.
... Extensive research finds that the similarities between same-sex and opposite-sex couples, on most relational factors, such as trust, intimacy, and satisfaction, outweigh the differences (Peplau & Fingerhut, 2007). For instance, Gottman et al. (2003) conducted a 12-year study of gay, lesbian, and opposite-sex couples and found levels of relationship satisfaction and quality to be equivalent. These findings were supported by Roisman et al. (2008), who found attachment styles, relationship satisfaction, and interaction quality during conflict, among gay, lesbian, and oppositesex couples to be comparable. ...
Article
Full-text available
IntroductionGay men form and maintain romantic relationships in the face of stigma and discrimination, negatively impacting their well-being. Early experiences with caregivers likely influence well-being and may later impact the satisfaction felt in the romantic relationships of gay men.Method Following the 2017 legalization of same-sex marriage in Australia, 198 self-identified gay men were surveyed between June and July2019, to examine the role of well-being in the association between attachment style and relationship satisfaction.ResultsResults revealed securely attached men demonstrated higher levels of well-being and relationship satisfaction. Men with higher levels of anxiety and avoidant attachment had lower well-being and relationship satisfaction. Well-being partially mediated the relationship between attachment style and relationship satisfaction.Conclusions Results highlight the significant contributing role of well-being among gay men in the association between attachment processes and the satisfaction felt in their romantic relationships. Our findings suggest that insecure attachment styles are associated with reduced well-being and may place gay men at risk for poorer relationship outcomes.Policy ImplicationsOur findings support the idea for researchers and clinicians to be mindful of the influence of attachment processes on well-being when working with gay male couples. Ongoing practitioner training should include a focus on these predictors of relationship quality for gay men. As social and public policy continues to move forward, it will be important to monitor how marriage equality evolves and the impact it continues to have on same-sex relationship outcomes.
... El segundo aspecto, se relaciona con los hallazgos de investigaciones como la de Kurdek (2005) en la que como parte de sus hallazgos, logró establecer que las parejas homosexuales y heterosexuales se asemejan en aspectos de la interacción entre sus miembros, como la expresión del afecto, la comunicación y la solución de problemas, este último aspecto, de acuerdo con Gallego y Barreiro de Motta (2010), Gottman et al. (2003) y Torres (2011), mayormente caracterizado en el caso de las parejas homosexuales por utilizar como estrategia principal el diálogo y búsqueda de acuerdos. Lo anterior cobra gran importancia principalmente porque esas semejanzas podrían indicar que la TCIP que ha mostrado ser efectiva para mejorar la satisfacción marital entre parejas heterosexuales, dada la semejanza existente, podría ser efectiva también para mejorar el manejo del conflicto, el intercambio de reforzadores y el malestar emocional entre las parejas homosexuales (Christensen, et al. 2004;Jacobson et al. 2000;Wimberly, 1997). ...
Thesis
Full-text available
Efecto de la terapia conductual integrativa sobre el manejo del conflicto, el intercambio de reforzadores y el malestar emocional en parejas homosexuales.
Article
Relationship health has a strong influence on physical and emotional health, and with reported rates of divorce at 40‐50%, relationship health is a critical public health issue. Thus, it is important to identify mechanisms that encourage healthy relationship functioning. This study measures the impact of engaging in discussions of challenging relationship patterns, or Relationship Pattern Labeling (RPL). Such discussions are embedded in existing interventions, and yet there is no literature, to our knowledge, that examines the impact of these discussions independently. Our results indicate significant small increases in intimacy (p = 0.002), acceptance (p < 0.001), and relationship satisfaction (p = 0.003) when participants engaged in RPL. We present preliminary evidence that supports the current applications of RPL and indicates that it may be successfully extended into independent online interventions.
Book
Full-text available
This volume provides a novel platform to re-evaluate the notion of open-ended intimacies through the lens of affect theories. Contributors address the embodied, affective and psychic, sensorial and embodied aspects of their ongoing intimate entanglements across various timely phenomena. This fascinating collection asks how the study of affect enables us to rethink intimacies, what affect theories can do to the prevailing notion of intimacy and how they renew and enrich theories of intimacy in a manner which also considers its normative and violent forms. This collection brings together a selection of original chapters which invite readers to rethink such concepts as care, closeness and connectivity through the notion of affective intimacies. Based on rigorous research, it offers novel insights on a variety of themes from austerity culture to online discussions on regretting motherhood, from anti-ableist notions of health to teletherapies in the era of COVID-19, and from queer intimacies to critiques of empathy. Lively and thought-provoking, this collection contributes to timely topics across the social sciences, representing multiple disciplines from gender studies, sociology and cultural studies to anthropology and queer studies. By so doing, it advances the value of interdisciplinary perspectives and creative methodologies for understanding affective intimacies.
Article
Objective: The present study aimed to examine the role of positive problem solving in the relation between perspective taking and relationship satisfaction. Participants: One hundred and four college students participated in the present study, if they had been in a romantic relationship for a minimum of six months. Methods: Participants completed measures of relationship satisfaction, perspective-taking relative to romantic couples, and of positive problem-solving in couples. Results: As predicted, both perspective-taking and positive problem-solving were significantly related to relationship satisfaction. In addition, positive problem-solving emerged as a significant partial mediator of the relation between perspective-taking and relationship satisfaction. Conclusions: The benefits of skills training in the areas of perspective-taking and problem-solving in college student health and functioning are discussed.
Article
Perceptions of marital interactions were gathered from a representative sample of urban newlywed couples (199 Black and 174 White). A factor analysis of the reports found 6 factors common to husbands and wives: Disclosing Communication, Affective Affirmation, Negative Sexual Interaction, Traditional Role Regulation, Destructive Conflict, and Constructive Conflict. Avoiding Conflict was specific to men and Positive Coorientation was specific to women. Wives reported fewer constructive and more destructive conflict behaviors. Compared with Whites, Blacks reported more disclosure, more positive sexual interactions, and fewer topics of disagreement. They also more often reported leaving the scene of conflict and talking with others more easily than with the spouse. As hypothesized, perceptions that marital interactions affirm one's sense of identity strongly predicted marital well-being. Although regression analyses predicting marital happiness yielded few interactions with race or gender, those that are significant, coupled with race and gender differences in perceiving interaction, suggest taking a contextual orientation to the meaning of marital interaction.
Article
Seventy-three couples were studied at 2 time points 4 years apart. A typology of 5 groups of couples is proposed on the basis of observational data of Time 1 resolution of conflict, specific affects, and affect sequences. Over the 4 years, the groups of couples differed significantly in serious considerations of divorce and in the frequency of divorce. There were 3 groups of stable couples: validators, volatiles, and avoiders, who could be distinguished from each other on problem-solving behavior, specific affects, and persuasion attempts. There were 2 groups of unstable couples: hostile and hostile/detached, who could be distinguished from each other on problem-solving behavior and on specific negative and positive affects. A balance theory of marriage is proposed, which explores the idea that 3 distinct adaptations exist for having a stable marriage.
Article
1. Introduction 2. Some helpful tools 3. Visualization of the pendulum's dynamics 4. Toward an understanding of chaos 5. The characterization of chaotic attractors 6. Experimental characterization, prediction, and modification of chaotic states 7. Chaos broadly applied Further reading Appendix A. Numerical integration - Runge-Kutta method Appendix B. Computer program listings References Index.
Article
Data from partners of 236 married, 66 gay cohabiting, and 51 lesbian cohabiting couples were used to assess if members of married couples differed from those of either gay couples or lesbian couples on five dimensions of relationship quality (intimacy, autonomy, equality, constructive problem solving, and barriers to leaving), two relationship outcomes (the trajectory of change in relationship satisfaction and relationship dissolution over 5 years), and the link between each dimension of relationship quality and each relationship outcome. Relative to married partners, gay partners reported more autonomy, fewer barriers to leaving, and more frequent relationship dissolution. Relative to married partners, lesbian partners reported more intimacy, more autonomy, more equality, fewer barriers to leaving, and more frequent relationship dissolution. Overall, the strength with which the dimensions of relationship quality were linked to each relationship outcome for married partners was equivalent to that for both gay and lesbian partners.