ICPR Abstracts: Session 4

---

Session 4: Symposium

Studying Relationships Using 
Intensive Repeated-Measures Designs

Methodological Overview

Niall Bolger, New York University
Deborah Kashy, Texas A&M University

In this presentation we will review the main types of 
intensive repeated measures designs that are used by 
relationships researchers: daily diary studies, interaction 
record studies, experience sampling studies, and multi-
wave panel studies.  The advantages of within-subject 
tests of hypotheses will be reviewed, and the capacity of 
these designs to estimate within-subjects effects will be 
demonstrated.  We will then show how these designs can 
be used to examine not only within-subject effects (e.g., 
the effect of interaction intimacy on well-being) but also 
between- subject differences in these effects (e.g., how the 
effect of interaction intimacy depends on attachment 
style).  Next, we will discuss the main types of statistical 
models that can be used to analyze data from these 
designs: conventional regression models, random intercept 
models, and random intercept and slope models.  We will 
finish by showing how the substantive presentations in the 
symposium fit into our analytic scheme. 


Does Relatedness Contribute to Daily Well-being?

Harry T. Reis, Kennon Sheldon, Joseph Roscoe 
and Richard Ryan
University of Rochester

It has become commonplace to begin articles and lectures 
about relationships with the observation that participation 
in satisfying social activities and networks is correlated 
with psychological and physical well-being.  Nearly all of 
the underlying evidence is limited to studies that examine 
differences between individuals; i.e., that show that 
persons who report high levels of social satisfaction also 
report high levels of psychological and physical well-
being.  Because of the reliance on such studies, an 
alternative explanation is also possible:  That certain 
individuals, for whatever reasons, are likely to evaluate 
their life circumstances more positively, whereas others 
view these circumstances more negatively. 
      A conceptually more interesting question that also 
eliminates this possible confound phrases the link between 
social life and well-being as a within-subjects question:  
Do individuals feel relatively healthy when they feel 
related to others, and unhealthy when they feel distant 
from others?  Or, in other words, do day-to- day variations 
in well-being correspond to day-to-day variations in social 
satisfaction? 
      In my talk, I will present results from a study that 
examines this question.  Specifically, we examined the 
relative impact of satisfaction of three basic needs -- 
autonomy, competence, and relatedness -- on several 
measures of well-being over a ten-day period.  
Additionally, we asked just what kind of social activities 
produce feelings of relatedness.  Although relationship 
researchers are well aware of the many varieties and forms 
that social activities may take, they have yet to 
demonstrate which forms are related to well-being and 
which are not.  In our study, we examined the prevalence 
and impact of seven types of social activity:  five positive 
(intimacy, conversation, fun, "hanging out," task-activity 
focus) and two negative (conflict, self- consciousness).  
Our findings demonstrate that well-being is indeed 
associated with daily variations in relatedness, and that 
some types of social experiences are more closely linked to 
feeling related and a sense of well-being than are others. 


Attachment as a Moderator of the Link Between
Partner Responsiveness and Personal Well-being

Lisa Feldman Barrett, Pennsylvania State University
Paula Pietromonaco, 
University of Massachusetts, Amherst

Most people benefit psychologically when they interact 
with a responsive partner.  Attachment theory suggests, 
however, that individuals who hold different expectations 
about attachment should be differentially sensitive to their 
partner's responsiveness.  We explored whether 
attachment style moderated the relationship between 
partner responsiveness and individuals' well-being 
following their social interactions. Secure, preoccupied, 
fearful- avoidant, and dismissing-avoidant participants 
provided immediate, on-line reports of their social 
interactions over a 1-wk period. We predicted that 
individuals' perceptions of their partners' responsiveness 
(i.e., expressions of positive or negative emotions, and 
disclosures) would be associated their (a) perceptions of 
intimacy (b) emotional responses, and (c) self- esteem, and 
(d) that attachment style would moderate the degree of 
these associations.  We performed a series of multi-level 
regression analyses with a weighted least squares 
estimation procedure (Kenny, Kashy, & Bolger, 1994) in 
which we included each of the partner responsiveness 
variables (e.g., partner's expression of positive emotion) as 
the lower-level predictor, and attachment group as the 
upper-level predictor.  Their cross-product term estimated 
the degree to which the relationship between the lower 
level predictor (e.g., partner responsiveness) and the 
criterion variable (e.g., self-esteem) was influenced by the 
upper level predictor (attachment). 
        As predicted, partner responsiveness was 
associated with more positive outcomes (e.g., greater 
perceptions of intimacy, more positive emotion, higher 
self-esteem), but these associations often were moderated 
by attachment style.  For example, (a) the self- esteem of 
preoccupied individuals depended more heavily on their 
perceptions of their partner's responsiveness than did the 
self- esteem of secure individuals, (b) the positive emotion 
experienced by preoccupied individuals depended less on 
their perceptions of partner responsiveness than did the 
emotions of secure individuals, and (c) fearful-avoidants 
reported more intimacy when they believed that their 
partner expressed more negative emotion, in contrast to 
secure individuals who showed the opposite pattern.  
Findings suggest that attachment style importantly 
moderates the link between partner responsiveness and 
personal well-being. 


Neuroticism, Negative Behavior, and the Trajectory of 
Marital Quality:  A Growth Curve Analysis

Benjamin R. Karney and Thomas N. Bradbury
University of California, Los Angeles

An inadequate understanding of the dependent variable in 
longitudinal research on marital quality renders even 
established findings difficult to interpret.  For example, 
two frequently studied independent variables in this 
research -- neuroticism and negative behavior during 
marital interactions -- have both shown consistent negative 
associations with marital quality longitudinally.  However, 
current methods of analyzing longitudinal data in marital 
research cannot specify whether these variables are 
associated with the level of marital quality over time or 
change in marital quality over time, leaving the true 
effects of these variables on the development of marital 
quality unclear.  In contrast, the analysis of growth curves 
allows the dependent variable in longitudinal research on 
marriage to be defined more precisely as the trajectory of 
marital quality over time, summarized by an intercept and 
a slope for each individual. 
        To refine our understanding of the longitudinal 
effects of neuroticism and negative behavior, the current 
study assessed the marital quality of 60 newlyweds every 
six months for four years (eight waves of data), and 
estimated intercepts and slopes summarizing the trajectory 
of each spouse.  Neuroticism was associated with the 
intercepts of husbands and wives but not with their slopes, 
i.e., the overall level of marital quality was affected but 
not the rate of change.  The negative behavior of wives 
was associated with the slopes of husbands and wives but 
not with their intercepts, i.e., the rate of change over time 
was affected but not the overall level of marital quality. 
        These findings demonstrate that current methods 
of analyzing longitudinal data in marriage may obscure 
important differences in the nature of effects on the 
development of marital quality.  The analysis of growth 
curves offers a useful means of disentangling these 
different effects and furthering an understanding of how 
marriages change.


Discussant

David Morgan
Portland State University