ArticlePDF Available

Social Relationships and Ambulatory Blood Pressure: Structural and Qualitative Predictors of Cardiovascular Function During Everyday Social Interactions


Abstract and Figures

Healthy normotensive men and women (N = 102) underwent a 3-day ambulatory blood pressure (BP) assessment in which a BP reading was taken 5 min into each social interaction. After each interaction, participants completed a diary that included structural categorization of the relationship and ratings of the quality of the relationship with the interaction partner. Random regression analyses revealed that interactions with family members and spouses were associated with lower ambulatory BP. Interactions with ambivalent network members (characterized by both positive and negative feelings) were associated with the highest ambulatory systolic BP, an effect that was independent of the familial effects on BP. Although there were psychological correlates associated with both structural and functional aspects of relationships, no evidence was found that these mediated the primary findings involving ambulatory BP. These data highlight the influence of both structural and qualitative aspects of relationships on ambulatory BP and possibly health.
Content may be subject to copyright.
Social Relationships and Ambulatory Blood Pressure:
Structural and Qualitative Predictors of Cardiovascular Function
During Everyday Social Interactions
Julianne Holt-Lunstad
Brigham Young University
Bert N. Uchino, Timothy W. Smith,
Chrisana Olson-Cerny, and Jill B. Nealey-Moore
University of Utah
Healthy normotensive men and women (N 102) underwent a 3-day ambulatory blood pressure (BP)
assessment in which a BP reading was taken 5 min into each social interaction. After each interaction,
participants completed a diary that included structural categorization of the relationship and ratings of the
quality of the relationship with the interaction partner. Random regression analyses revealed that
interactions with family members and spouses were associated with lower ambulatory BP. Interactions
with ambivalent network members (characterized by both positive and negative feelings) were associated
with the highest ambulatory systolic BP, an effect that was independent of the familial effects on BP.
Although there were psychological correlates associated with both structural and functional aspects of
relationships, no evidence was found that these mediated the primary findings involving ambulatory BP.
These data highlight the influence of both structural and qualitative aspects of relationships on ambu-
latory BP and possibly health.
Key words: ambulatory blood pressure, social relationships, cardiovascular functioning, diary study,
ambivalence, health
Relationships with others occupy a central role in people’s
everyday lives, and evidence suggests that these relationships have
positive effects on physical health (Berkman, 1995; Cohen, 1988;
House, Landis, & Umberson, 1988). Reviews of prospective stud-
ies indicate that people who are socially isolated are at increased
mortality risk from a number of causes (Berkman, 1995) and that
the effects of social relationships are comparable with standard
risk factors such as smoking, exercise, and diet (House et al.,
1988). In fact, both structural (e.g., type of relationship or size of
social network) and qualitative (e.g., perceived helpfulness) as-
pects of social relationships are inversely related to the incidence
of the most common cause of death in industrialized nations—
coronary heart disease (CHD; Berkman, Leo-Summers, & Horow-
itz, 1992; Cohen, 1988; Orth-Gomer, Rosengren, & Wilhelmsen,
1993). One relatively unexplored pathway by which relationships
may influence CHD is through their impact on blood pressure (BP)
during daily life (Uchino, Cacioppo, & Kiecolt-Glaser, 1996).
Thus, the major aim of this study was to examine the association
of ambulatory BP levels with differing aspects of social
Studies of ambulatory BP are important because this index is a
predictor of cardiovascular outcomes. Although studies have con-
firmed that the risk of cardiovascular disease rises in a linear
fashion with increases in resting BP assessed in the clinic (Mac-
Mahon et al., 1990), clinic BP may not represent an individual’s
usual level or capture important fluctuations. Ambulatory assess-
ments may more closely characterize an individual’s BP because a
number of representative measurements are taken during everyday
life (see Stone & Shiffman, 1994). Importantly, studies suggest
that elevated ambulatory BP may be a stronger predictor of car-
diovascular outcomes, including severity of complications in es-
sential hypertension, organ damage such as left ventricular wall
thickness or hypertrophy, and overall morbidity and mortality than
are clinic BP readings (Perloff, Sokolow, & Cowan, 1983; Pick-
ering, Harshfield, Devereux, & Laragh, 1983; Prisant, Carr, Wil-
son, & Converse, 1990; Sokolow, Werdeger, & Hinman, 1966).
Although there has been research on how psychosocial variables
such as mood (e.g., Gellman et al., 1990; Kamarck et al., 1998) and
personality (e.g., Porter, Stone, & Schwartz, 1999; Raikkonen,
Matthews, Flory, Owens, & Gump, 1999) may influence ambula-
tory BP, very little attention has been paid to whether character-
istics of social relationships predict ambulatory BP (Uchino et al.,
Julianne Holt-Lunstad, Department of Psychology, Brigham Young
University; Bert N. Uchino, Timothy W. Smith, Chrisana Olson-Cerny,
and Jill B. Nealey-Moore, Department of Psychology and Health Psychol-
ogy Program, University of Utah.
This research was generously supported by National Institute of Mental
Health Grant 1 R01 MH58690-01 awarded to Bert N. Uchino. We thank
Thomas Kamarck for his role as a consultant for this project, Ken Smith
and Robert MacCallum for their assistance with the statistical analyses, and
Kelly Anderson, Carol MacFarlane, and John Weir for their help running
participants through the protocol. We also thank John Cacioppo and Janice
Kiecolt-Glaser for their helpful comments on an earlier draft of this article.
Correspondence concerning this article should be addressed to Julianne
Holt-Lunstad, Department of Psychology, Brigham Young University,
1024 SWKT, Provo, Utah 84602–5543. E-mail: julianne.holt-lunstad@
Health Psychology Copyright 2003 by the American Psychological Association, Inc.
2003, Vol. 22, No. 4, 388–397 0278-6133/03/$12.00 DOI: 10.1037/0278-6133.22.4.388
1996). This is surprising considering social relationships have been
linked to significant physical health outcomes including endpoints
directly related to CHD (Berkman, 1995). Of the few studies that
have examined such an association, several do suggest its impor-
tance in predicting ambulatory BP. For example, in an examination
of more structural aspects of relationships (e.g., with family, with
friends, or with strangers), Spitzer et al. (1992) found that ambu-
latory BP levels were lowest when with a family member and
highest when with a stranger (also see Guyll & Contrada, 1998).
Although these few studies provide preliminary evidence of an
association between social relationships and ambulatory BP, many
important questions are raised by these data. First, Spitzer et al.
(1992) only provided a limited sampling of different relationship
types (i.e., family, friends, strangers). It was unclear if the BP
effects attributed to these higher order social contexts were due to
more specific interactions with spouses, coworkers, and others.
Thus, a comparison between other structural categories of rela-
tionships (e.g., coworkers, romantic relationships) is an important
unexplored issue. It should be noted that direct comparisons be-
tween more diverse structural aspects of social networks have been
difficult in prior research. Of the few studies that exist, all have
used a standard interval-contingent sampling that takes a BP
reading during a certain time point (e.g., once every 1520 min or
every hour). Unless extended over several days, it is unclear as to
whether such a method would adequately capture a comprehensive
assessment of daily social interactions. Even if a social interaction
was assessed, there may be significant variability in how long into
it that BP was taken using a random sampling procedure. There-
fore, in the present study, we capitalized on a well-established
method of sampling (i.e., event-contingent sampling; Reis &
Wheeler, 1991) used in social interaction research that has not thus
far been used in ambulatory BP studies. Event-contingent sam-
pling method uses explicit criteria to define an event (in our case,
a social interaction) that can then be linked to ambulatory BP. This
sampling method may provide an alternative test of whether dif-
fering social contexts predict ambulatory BP because any social
interaction that meets predefined criteria can potentially be sam-
pled. Consistent with the results of Spitzer and colleagues, we
predicted that interactions with family members would be associ-
ated with the lowest ambulatory BP. We further hypothesized that
interactions with spouses would also predict lower ambulatory BP
than interactions with others, given prior research on the protective
effects of marital relationships (Kiecolt-Glaser & Newton, 2001;
Stroebe & Stroebe, 1996).
A second major limitation of the existing studies is that no study
that we are aware of has assessed the effects of the quality of the
relationship on ambulatory BP during social interactions. General
views or schemas about specific relationships provide an important
cognitive lens through which the behaviors of others are encoded
and interpreted (Dunkel-Schetter & Bennett, 1990). However,
there appears to be considerably more heterogeneity in the quality
of individuals relationships than has been examined in prior
research. Most prior research has examined the positive and neg-
ative aspect of individuals social relationships from a bipolar
perspective, implicitly assuming that social relationships are char-
acterized as either primarily positive or primarily negative. Posi-
tive and negative aspects of social relationships, however, tend to
be separable dimensions (Finch, Okun, Barrera, Zautra, & Reich,
1989; Fiore, Becker, & Coppel, 1983; Kiecolt-Glaser, Dyer, &
Shuttleworth, 1988; Ruehlman & Karoly, 1991), and many indi-
viduals construe their relationships as having a mix of positive and
negative feelings (Uchino, Holt-Lunstad, Uno, & Flinders, 2001).
The implications of this ambivalence within relationships have not
been adequately considered in social support theory or research
(Coyne & DeLongis, 1986; Major, Zubek, Cooper, Cozzarelli, &
Richards, 1997; Uchino et al., 2001).
We have argued that the separability of positive and negative
aspects of social relationships may have significant conceptual
implications for their joint study (Uchino et al., 2001). As illus-
trated in Figure 1, such data suggest that any given social network
member may differ in his or her underlying positive and negative
basis (Cacioppo & Berntson, 1994). As depicted in the high-
positivity/low-negativity corner of Figure 1, there may be social
network members that are primarily sources of social support or
other pleasant interpersonal experiences (e.g., enjoyable friends).
The low-positivity/high-negativity corner reflects a network tie
that is primarily a source of negativity or what we label a socially
aversive tie (e.g., an unreasonable work supervisor). The low-
positivity/low-negativity corner would be a socially indifferent tie
and may represent network members who are characterized by
relatively low levels of social interactions (e.g., casual coworkers
or neighbors). A unique aspect of this conceptualization for the
social relationships and health literature is represented in the
high-positivity/high-negativity corner of Figure 1. We label such a
network member as a source of ambivalence. This refers specifi-
cally to network members who are a source of both positivity and
negativity (e.g., overbearing parent, volatile romance, competitive
friend). Unfortunately, the implications of ambivalent relation-
ships have not been adequately examined as most of the prior
research on social support has ignored the negative aspects that
may co-occur with the positive aspects of relationships (Coyne &
DeLongis, 1986; Rook & Pietromonaco, 1987; Uchino et al.,
We have found in our prior research that ambivalent ties, com-
pared with supportive ties, were associated with (a) greater inter-
personal stress and (b) higher cardiovascular reactivity during
acute stress in older adults (Uchino et al., 2001). Based on these
data, we predicted that interactions with ambivalent network mem-
bers during everyday life would be associated with the highest
ambulatory BP levels. Consistent with the literature linking more
Figure 1. General conceptual framework incorporating the positive and
negative aspects of social relationships on health.
positive aspects of relationships to better health outcomes, we also
predicted that daily interactions with supportive network members
(i.e., primarily positive relationships) would be associated with the
lowest ambulatory BP.
Social relationships presumably influence ambulatory BP by
means of relevant psychological processes. For instance, prior
research suggests that supportive relationships are associated with
a more positive psychological profile (Sarason, Pierce, & Sarason,
1990). In the present study, we examined a variety of psychosocial
factors that may be responsible for the association between social
relationships and these physiological processes. First, we were
interested in how the structural and qualitative relationship vari-
ables predicted experienced psychological states. On the basis of
prior research, we first predicted that supportive relationships
would be associated with increases in self-disclosure, intimacy,
and positive affect and decreases in negative affect compared with
the other relationship categories depicted in Figure 1. Our prior
work with older adults further suggests that interactions with
ambivalent ties would predict increases in negative affect. Second,
we were interested in whether these factors then served as medi-
ators of any obtained results. We thus examined if any significant
associations detected in our first set of analyses could account for
the observed associations between social relationships and ambu-
latory BP.
Forty-nine healthy men and 53 healthy women participated in this study.
Volunteers were recruited from introductory psychology courses and
through paid advertisement. Our participants were between the ages of 18
and 46 years (M 24 years), and 86% were employed.
Participants were
given extra credit or paid $50 for their time. Consistent with our prior
research (e.g., Cacioppo et al., 1995), the following self-reported inclusion
criteria were used to select healthy participants: no existing hypertension,
no cardiovascular prescription medication use, no past history of chronic
disease with a cardiovascular component (e.g., diabetes), no recent history
of psychological disorder (e.g., major depressive disorder), and no con-
sumption of more than 10 alcoholic beverages a week.
Participants volunteered to participate in a 3-day study in which an
ambulatory BP measurement was to be initiated 5 min into every social
interaction. Each participant wore an ambulatory BP monitor and carried a
folder with a number of diary sheets throughout the study. To get a broader
sampling of an individuals social interaction, we assessed 2 working or
school days and 1 nonworking or nonschool day. On each day, participants
arrived at the lab in the morning, and the monitor was attached by a trained
assistant. Participants returned to the lab each evening (after 8 p.m.), the
monitor was removed, completed diary sheets were collected, and protocol
accuracy for the day was assessed (i.e., diary accuracy scale; see below).
On the 1st day, participants received detailed instructions. They were
told to initiate a BP reading approximately 5 min into all social interactions
that lasted a minimum of that duration. Participants were also told that in
the event they forgot to initiate a BP reading after 5 min, they should still
take a reading and estimate how long into the interaction it was assessed.
Criteria for a social interaction were defined as any activity in which
participants were mutually engaged with another individual, such as a
conversation (Reis & Wheeler, 1991). The mere presence of another
individual was neither necessary (e.g., telephone conversations) nor suffi-
cient (merely being in the same room without an interaction). Diary sheets
were then explained, and participants were told to fill out one diary sheet
immediately following each social interaction.
On the final day of the protocol, all of the participants were compen-
sated, debriefed, and thanked for their participation. Because of the poten-
tially sensitive nature of these interpersonal assessments, strict assurances
of confidentiality were given to all participants. They were also given the
opportunity at the end of the study to withdraw and take their diaries with
them without any penalty, although no participant in the present study
chose this option.
Ambulatory BP monitor. The Accutracker II (Suntech Medical Instru-
ments, Raleigh, NC) was used to estimate ambulatory readings of systolic
blood pressure (SBP), diastolic blood pressure (DBP), and heart rate (HR).
The Accutracker II was designed specifically for ambulatory assessments
and is well-validated as readings correspond with intra-arterial BP assess-
ments during rest, isometric exercise, and bicycle exercise (White, Lund-
Johansen, & Omvik, 1990; also see Light, Obrist, & Cubeddu, 1988). The
Accutracker II assessed HR by means of three electrocardiogram sensors
attached to the chest, whereas SBP and DBP were assessed with the
auscultatory method using a microphone and occluding cuff. The sensors
and the cuff were worn under the participants clothing, and only a small
control box (approximately 5.0 in. 3.5 in. 1.5 in., or 12.7 cm 8.89
cm 3.81 cm) attached to the participants belt was partially exposed.
Participants were instructed to press an easily identified start button on
the monitor to initiate a BP reading. The monitor recorded and saved all BP
and HR readings, as well as the times and the date they were taken in
memory. Each evening a trained research assistant downloaded the data
that was obtained during the day.
The Accutracker II uses a number of codes that may signify problems
with the estimation of the ambulatory cardiovascular assessment. On the
basis of prior research (see Kamarck et al., 1998), we deleted readings
associated with test codes 2 (weak Korotkoff sounds), 3 (microphone
difficulties), and 4 (air leaks). Outliers associated with artifactual readings
were also identified using the criteria by Marler, Jacobs, Lehoczky, and
Shapiro (1988). These included the following: (a) SBP less than 70 mm Hg
or greater than 250 mm Hg, (b) DBP less than 45 mm Hg or greater than
150 mm Hg, (c) HR less than 40 beats per minute (bpm) or greater than 200
bpm, and (d) SBP/DBP less than 1.065 (.00125 DBP) or greater
than 3.0. Compared with prior research, 17% of the ambulatory cardiovas-
cular readings were deleted on the basis of the above criteria, resulting
in 1,568 valid ambulatory readings (e.g., Raikkonen et al, 1999). These
criteria resulted in the loss of 4 participants who started with few readings.
As a result, our final sample included 98 participants.
Ambulatory diary. Participants were instructed to complete a one-page
diary sheet with each ambulatory cardiovascular assessment. It was de-
signed to be easy to complete (about 12 min) to maximize cooperation
and was divided into four general sections. The first consisted of general
information, such as the date and time of the interaction. It also assessed
the duration of the interaction when the cuff was inflated. The second
section assessed information on basic variables that might influence car-
diovascular function (see Guyll & Contrada, 1998; Kamarck et al., 1998).
These include items such as posture (lying down, sitting, standing); activity
level (1 no activity, 4 strenuous activity); location (work, home,
other); talking (yes, no); temperature (too cold, comfortable, too hot); prior
consumption of nicotine, caffeine, alcohol, or a meal; and prior exercise
(no, yes). The third section of the ambulatory diary was adapted from the
Rochester interaction record (see Reis & Wheeler, 1991). Participants were
The University of Utah is a commuter school and comprises many
nontraditional students; therefore, many of our participants may have been
working and attending school.
asked to list the initials of the primary person they were interacting with
and their relationship to the person (e.g., friend, mother). This section also
included assessments of perceived characteristics of the interaction, includ-
ing positive affect, negative affect, intimacy, self-disclosure, and social
influence (1 not at all, 6 extremely). The final section of the diary was
adapted from our social relationship index (Uchino et al., 2001) and
assessed how generally positive and negative the participant typically felt
toward the primary interaction partner (1 not al all, 6 extremely).
Diary Accuracy Scale. The Diary Accuracy Scale (DAS) was created
on the basis of interview questions used extensively in prior research
(Cutrona, 1986; Hodgins & Zuckerman, 1990; Reis, Senchak, & Solomon,
1985; Reis & Wheeler, 1991) as an indication of accuracy with the diary
procedure. The DAS was completed at the end of each monitoring day and
was used to estimate the participants difficulty and accuracy associated
with the ambulatory protocol. Participants rated the difficulty in remem-
bering to inflate the cuff or record the interaction (1 no difficulty, 7
very much difficult). They also estimated the percentage of interactions in
which they did not inflate the cuff and did not record the interaction in the
diary. Finally, participants rated the extent to which cuff inflation and diary
recordings interfered with the interaction (1 none, 7 a great deal).
Preliminary Analyses
We first examined participants ratings of the difficulty and
accuracy associated with the protocol as assessed by the end-of-
day DAS. Mixed-model analyses of variance were performed
using time (Day 1, Day 2, Day 3) as a repeated variable. Gender
was also examined as a between-participant variable. Results re-
vealed no significant gender main effects on any of these items.
Furthermore, no significant changes occurred over time on any of
these measures, and no Gender Time interactions approached
significance. Average levels of these ratings are summarized in
Table 1 and suggest that participants (a) did not view the protocol
as particularly difficult, (b) were relatively accurate in their rat-
ings, and (c) missed recording a relatively small proportion of the
potential interactions. These findings are relatively consistent with
prior studies that use event-contingent sampling of social interac-
tions (Reis & Wheeler, 1991). Nevertheless, because of the retro-
spective nature of these ratings, it is possible that participants may
have overestimated their adherence. In addition to reports of ac-
curacy, we also examined the number of BP readings that did not
have a corresponding diary. Our participants were missing an
average of 1.04 diary record(s) per day,
which is only slightly
higher than their self-reported accuracy assessment.
We next examined the number and nature of the social interac-
tions that occurred during the study. Participants had a mean
of 18.8 interactions during the 3-day recording period (M 6.3
interactions/day). As shown in Table 2, many of these interactions
occurred with friends (29.8%), immediate family (20.7%), or
coworkers (13.1%). Importantly, as evidenced by Table 2, we
appear to have a broad sampling of participants interactions with
their social network.
Do Different Categories of Relationships Predict
Characteristics of the Interaction?
Before examining if more structural aspects of relationships
influence ambulatory BP, we first examined the psychological
processes associated with these different structural categories. To
have an adequate number of readings in each relationship category,
we divided relationships into the higher order categories of famil-
ial relationships (Categories 1 to 7; see Table 2) versus nonfamilial
relationships (Categories 8 to 16), romantic relationships (Catego-
ries 6 to 9) versus nonromantic relationships (Categories 1 to 5
and 10 to 16), spousal relationships (Categories 6 and 7) versus
nonspousal relationships (Categories 1 to 5 and 8 to 16), and work
relationships (Categories 12, 13, and 15) versus nonwork relation-
ships (Categories 1 to 11 and 14, 16). In the present analysis,
Category 17 (other) presented problems as the nature of these
relationships was not known. As a result, these data were excluded
from the analyses.
We used Proc Mixed (SAS Institute; Littell, Milliken, Stroup, &
Wolfinger, 1996) to examine the diary ratings of positive affect,
negative affect, intimacy, self-disclosure, and influence (see
Schwartz & Stone, 1998). Proc Mixed uses a random regression
model to derive parameter estimates both within and across indi-
This is equivalent to 16% over the 3-day study.
It is possible that some BP readings without corresponding diary sheets
may have been due to erroneously hitting the start button or initiating a BP
reading out of curiosity in nonsocial situations. Therefore, it is unclear if
participants were overestimating their compliance.
Table 2
Frequency and Percentage of Interactions With Different
Category and
relationship type Frequency %
Mean no. of
per individual
1. Father 42 2.28 0.35
2. Mother 94 5.11 0.69
3. Sister 50 2.72 0.40
4. Brother 42 2.28 0.37
5. Relative other 114 6.20 0.75
6. Husband 49 2.66 0.38
7. Wife 104 5.66 0.74
8. Girlfriend 49 2.66 0.44
9. Boyfriend 57 3.10 0.41
10. Friend 548 29.80 4.50
11. Roommate 29 1.58 0.25
12. Coworker 241 13.10 2.14
13. Boss 68 3.70 0.62
14. Neighbor 7 0.38 0.03
15. Client or customer 20 1.09 0.13
16. Classmate 40 2.18 0.32
17. Other 285 15.50 2.42
Table 1
Mean Diary Rating Scale Items
Item MSD
Difficulty remembering to inflate cuff 1.96 1.04
Difficulty remembering to record diary 2.24 1.20
Accuracy in recorded diary 2.28 0.91
Interactions cuff not inflated (%) 10.42 14.03
Interactions not recorded in diary (%) 7.73 10.32
Cuff interfered with interaction 2.26 1.30
Diary interfered with interaction 1.94 1.17
viduals (Singer, 1998). In the present study, the intercept was
treated as a random variable (see Singer, 1998). By default, Proc
Mixed also treats the unexplained variation within individuals as a
random factor. The unstructured variancecovariance matrix was
specified in this model and allowed the intercept and variation
within individuals to be estimated from the data. Unless otherwise
noted, all other factors were treated as fixed variables to reduce the
complexity of the model.
In these analyses, gender and relation-
ship categorization were first centered at their grand mean before
inclusion into the model (Singer, 1998). Because of the overlap
inherent in these categorizations (e.g., spouses are included in
familial ties), we performed separate analyses for each relationship
In these analyses, we first examined the psychological correlates
of interacting with familial versus nonfamilial relationships. As
shown in Table 3, results revealed that interactions with familial
ties were associated with greater levels of positive affect, intimacy,
and self-disclosure. Of interest, familial interactions did not differ
from nonfamilial interactions in terms of negative affect (p .90)
or influence (p .50). Analyses involving spousal and romantic
relationships were consistent with the above findings as spousal
interactions were associated with higher positive affect, intimacy,
and self-disclosure. Again, no significant differences were found
for negative affect and influence.
Finally, analyses of interactions with work relationships re-
vealed that these were viewed more negatively than interactions
with nonwork relationships. Work relationships were associated
with lower positive affect, greater negative affect, lower intimacy,
and lower self-disclosure.
Do Different Categories of Relationships Predict
Ambulatory Cardiovascular Assessments?
The preceding analyses suggested that interactions with familial,
spousal, romantic, and nonwork relationships were viewed more
positively than their counterparts. In our first analyses of BP, we
examined the association between the structural categories of
relationships and ambulatory BP. However, before proceeding to
these analyses we examined the potential contribution of extrane-
ous factors such as posture that might need to be statistically
controlled (Schwartz, Warren, & Pickering, 1994). In these anal-
yses, time (first reading, second reading, etc.) was also included as
a random factor, whereas gender, body mass index, time passed
before cuff inflation, posture (lyingsitting, sittingstanding), ac-
tivity level, talking, location (homework, homeother), and tem-
perature (comfortabletoo cold, comfortabletoo hot) were treated
as fixed factors in the model. In addition, exercise and consump-
tion of nicotine, caffeine, alcohol, and a meal since the last reading
were treated as fixed factors and examined because of their po-
tential influence on ambulatory BP. All variables were first cen-
tered at their grand mean before inclusion into the model (Singer,
1998) and were analyzed using Proc Mixed.
Results of this initial model revealed that time, prior nicotine
use, homework location, and homeother location were indepen-
dent predictors of higher ambulatory SBP, whereas lyingsitting
posture, sittingstanding posture, prior nicotine use, talking, and
homeother location independently predicted higher ambulatory
DBP. In analyses of HR, women had higher ambulatory HR than
men. Likewise, lyingsitting posture, sittingstanding posture, ac-
tivity level, prior nicotine, prior meal, and homeother location
predicted greater ambulatory HR. No other variable approached
In our main analyses, the covariates (p .10) from our first
analyses were included in the equation to provide a test of the
independent effects of relationship categories on ambulatory car-
diovascular function. Results of these analyses converged some-
To take a more conservative approach, we also performed our analyses
treating the major structural and qualitative relationship assessments as
random factors in the model. Results of these analyses were comparable
with those reported later.
Table 3
Primary Results of Interactions Between Structural and Functional Aspects of Social Relationships, Psychological Factors, and
Cardiovascular Functioning
Dependent variable
Structural aspects
Familial vs.
Romantic vs.
Spousal vs.
nonspousal Work vs. nonwork
b t df b t df b t df b t df
Positive affect 0.23** 2.80 1,169 0.21* 1.96 1,169 0.31* 2.33 1,169 0.53*** 5.87 1,169
Negative affect 0.01 0.11 1,169 0.04 0.39 1,169 0.15 1.06 1,169 0.33*** 3.48 1,169
Intimacy 0.80*** 7.70 1,169 1.04*** 7.98 1,169 0.97*** 5.72 1,169 1.27*** 11.36 1,169
Disclosure 0.29** 3.00 1,163 0.47*** 3.86 1,163 0.44** 2.77 1,163 0.58*** 5.40 1,163
Influence 0.05 0.63 1,168 0.09 0.93 1,168 0.14 1.16 1,168 0.10 1.19 1,168
Ambulatory cardiovascular
SBP 3.42* 2.30 1,141 2.07 1.16 1,141 3.48 1.61 1,386 3.14 1.44 1,141
DBP 2.46** 3.13 1,099 1.93* 1.99 1,099 3.02* 2.55 1,336 0.20 0.24 1,099
HR 0.84 0.97 1,110 1.01 0.94 1,110 1.25 0.93 1,342 0.14 0.15 1,110
Note. SBP systolic blood pressure; DBP diastolic blood pressure; HR heart rate.
* p .05. ** p .01. *** p .001.
what with the findings regarding the psychological correlates of
relationship categories (see Table 3). Consistent with Spitzer and
colleagues (Spitzer et al., 1992), familial interactions were asso-
ciated with lower levels of ambulatory SBP (for nonfamily, M
138.98; for family, M 135.84) and DBP (for nonfamily,
M 81.05; for family, M 78.79). In addition, DBP was lower
when interacting with romantic (for nonromantic, M 77.28; for
romantic, M 75.86) and spousal relationships (for nonspouse,
M 77.39; for spouse, M 75.79). Interactions with work versus
nonwork relationships did not predict ambulatory SBP, DBP, or
HR (ps .15).
Finally, we examined whether the psychological variables me-
diated the effects of these structural aspects of relationships on
ambulatory BP. We focused on familial versus nonfamilial rela-
tionship as it had the most consistent associations with ambulatory
BP. Analyses reveal that statistically controlling for these psycho-
logical factors did not alter the effects for familial versus nonfa-
milial relationships on ambulatory SBP (
⫽⫺4.03, p .007) or
⫽⫺3.03, p .0002). These analyses suggest that these
psychological factors do not account for the observed effects of
relationship type on ambulatory BP.
Does the Quality of the Relationship Predict
Characteristics of the Interaction?
The second major aim of this study was to investigate whether
the quality of the relationship (i.e., how positive and negative
participants typically felt toward the primary interaction partner)
predicted ambulatory cardiovascular function. As in the prior
analyses, we first examined more specific psychological processes
potentially associated with these qualitative measures. Gender,
normally perceived positivity, and normally perceived negativity
were treated as fixed factors and centered at their grand mean.
These variables, along with the normally perceived positivity and
negativity cross-product term, were entered simultaneously into
the random regression models predicting positive affect, negative
affect, intimacy, self-disclosure, and influence. These analyses
revealed several main effects that replicated prior research on
social relationships and psychological outcomes. Consistent with
prior research on supportive relationships, results revealed that
normally perceived positivity was associated with higher levels of
positive affect, lower levels of negative affect, greater ratings of
intimacy, and greater self-disclosure (see Table 3). In comparison,
perceived negativity toward the primary person with whom par-
ticipants interacted predicted less positive affect, higher negative
affect, and lower levels of self-disclosure. Surprisingly, normally
perceived negativity toward the person predicted greater ratings of
Of greater interest for the conceptual framework depicted in
Figure 1 were the statistical interactions between ratings of nor-
mally perceived positivity and negativity that emerged for ratings
of self-disclosure and negative affect. The form of these interac-
tions was examined by computing predicted values one standard
deviation above and below the mean for perceived positivity and
negativity. These predicted values revealed that self-disclosure
was highest when interacting with more supportive ties (high
positivity, low negativity) and lowest when interacting with indif-
ferent (low positivity, low negativity) and aversive (low positivity,
high negativity) ties. Interactions with ambivalent ties were asso-
ciated with a moderate level of self-disclosure. Predicted negative
affect score also revealed that negative affect was highest when
interacting with aversive ties, moderate when interacting with
ambivalent ties, and lowest when interacting with supportive and
indifferent ties.
Although it is difficult to explain this finding, we suspect that it may
be a statistical artifact as no other ratings paralleled this perceived nega-
tivity main effect.
Functional aspects
General positivity General negativity Positivity Negativity
b t df b t df b t df
0.36*** 12.63 1,378 0.21*** 5.63 1,378 0.04 1.71 1,378
0.09** 2.96 1,378 0.45*** 10.56 1,378 0.08*** 3.80 1,378
0.61*** 16.68 1,376 0.17*** 3.37 1,376 0.00 0.17 1,376
0.35*** 9.86 1,373 0.10* 2.05 1,373 0.08** 3.18 1,373
0.02 0.61 1,377 0.04 1.01 1,377 0.00 0.06 1,377
0.42 0.80 1,342 0.74 1.04 1,342 0.73* 2.03 1,342
0.05 0.18 1,289 0.22 0.58 1,289 0.44* 2.20 1,289
0.16 0.49 1,302 0.01 0.01 1,302 0.07 0.32 1,302
Does the Quality of the Relationship Predict Ambulatory
Cardiovascular Assessments?
In our main analyses, we again used Proc Mixed to investigate
if the quality of the relationships predicted ambulatory cardiovas-
cular assessments. In the random regression model, the covariates
(p .10) from our first analyses were again included in the
equation to provide a test of the independent effects of relationship
quality. These covariates were entered along with measures of
normally perceived positivity, normally perceived negativity, and
the cross-product term for positivity and negativity to test the
utility of the model depicted in Figure 1. All main effect variables
were again centered at the grand mean before inclusion in the
model (Singer, 1998).
No main effects of normally perceived positivity and negativity
emerged for ambulatory SBP, DBP, and HR. However, significant
statistical interactions between normally perceived positivity and
negativity emerged in predicting ambulatory SBP and ambulatory
DBP (ps .05). The form of these statistical interactions were
examined by plotting predicted values for SBP and DBP one
standard deviation below and above the mean for perceptions of
relationship positivity and negativity (Aiken & West, 1991). As
shown in Figure 2, individuals interacting with social ties about
which they normally felt relatively high levels of both positivity
and negativity (ambivalent) had elevated levels of ambulatory SBP
compared with the other relationship categories. Similar to the
pattern for SBP, interactions with ambivalent ties were associated
with the highest levels of ambulatory DBP. No significant statis-
tical interactions were found in analyses of HR.
Overall, these
data provided evidence for the potential detrimental influence of
interactions with ambivalent ties on ambulatory BP.
One potential alternative explanation for these findings on am-
bulatory BP is that the relationship ratings were biased by current
levels of affect during the interaction. To address this issue, we
examined the statistical interaction between rated positive and
negative affect during the actual interaction on ambulatory BP. If
characteristics of the interaction significantly biased the relation-
ship ratings, then the same patterns of association should be
evident using the mood ratings of the interactions. Inconsistent
with this possibility, results of these ancillary analyses revealed
that none of the statistical interactions between positive and neg-
ative affect approached significance (ps .25). In addition, sta-
tistically controlling for the main effects and cross-product term
for state positive and negative affect did not alter the significance
levels of the association between relationship quality and ambu-
latory BP. Therefore, it appears that state measure of affect did not
influence the ambulatory BP effects we found for relationship
In more conceptually interesting analyses, we examined whether
the psychological variables mediated the effects of social relation-
ships on ambulatory BP. Although the statistical interactions be-
tween positivity and negativity for self-disclosure and negative
affect did not evidence the same pattern as those for ambulatory
BP, nevertheless we examined these as potential mediators of our
observed results. The interaction of positive and negative relation-
ship quality remained significant for SBP (
0.73, p .05) and
0.44, p .05) when statistically controlling for these
factors, suggesting that these psychological processes did not
account for the effects of relationship quality on ambulatory BP.
What Is the Independent Contribution of Structural
Versus Qualitative Aspects of Relationships in Predicting
Ambulatory BP?
Results of this study suggest that both structural and qualitative
aspects of social relationships predict ambulatory BP. An impor-
tant question raised by these data is the independent contribution
of these factors. For instance, is the effect of the quality of the
relationship simply an artifact of its overlap with familial versus
nonfamilial interactions? Or more interesting, given data linking
both structural and qualitative aspects of relationships to health
outcomes, are these independent pathways by which relationships
may influence ambulatory BP? To address this question, we du-
plicated the analyses in the last section but included the familial
versus nonfamilial category into the random regression equation.
The familial category was used because it was the only significant
predictor of both ambulatory SBP and DBP. Results of these
analyses for SBP revealed that both the familial categorization
(p .03) and the normally perceived Positivity Negativity
interaction (p .02) remained significant. These data suggest that
the elevated ambulatory SBP seen when interacting with ambiva-
lent ties was independent of the familial categorization, and vice
versa. However, analyses of DBP revealed that only the familial
categorization main effect remained significant (p .001).
The primary aims of this study were to examine the prediction
of ambulatory cardiovascular function by (a) different structural
categories of relationships and (b) their quality. Results not only
replicated and extended prior research on the BP correlates of
structural social network classifications but also demonstrated the
potential importance of the conceptual framework depicted in
Figure 1. As noted earlier, prior research has examined positivity
and negativity in social relationships from a bipolar perspective.
This perspective would not have captured the synergistic effects of
positivity and negativity (ambivalence) for ambulatory SBP, an
We also conducted ancillary analyses in which we examined the
three-way interaction among gender, normally perceived positivity, and
normally perceived negativity. Results did not suggest that gender moder-
ated our main results (ps .13). However, the sample size in the present
study may have limited the power to test this three-way interaction.
Figure 2. Predicted systolic blood pressure (SBP) levels during social
interactions as a function of normally perceived relationship positivity and
negativity (one standard deviation above and below the mean).
effect that was independent of the structural categorization. These
results are particularly interesting in light of research emphasizing
the prognostic importance of SBP in predicting cardiovascular
disorders (Lloyd-Jones, Evans, Larson, ODonnell, & Levy, 1999).
Consistent with Spitzer et al. (1992), familial ties proved to be
a strong, independent predictor of ambulatory SBP and DBP.
Importantly, we were able to rule out some of the alternative
explanations suggested by Spitzer and colleagues, such as physical
activity and talking. However, an important unanswered question
relates to why familial relationships are associated with lower
levels of ambulatory cardiovascular assessments. The self-reported
ratings of positive affect, negative affect, self-disclosure, and in-
timacy differed by the relationship categorizations but did not
statistically mediate the effects on ambulatory BP. One possible
explanation for these findings is that increased familiarity associ-
ated with familial ties may have a calming effect on the cardio-
vascular system. If this familiarity is indeed shaped by years of
contact, then self-reported affect rating taken at the time of the
interaction may not tap into this more automatic process. However,
we only examined a limited set of psychological and behavioral
processes. Therefore, although our failure to demonstrate media-
tion may appear to challenge the idea that psychological factors
can have a direct impact on physiology, there may be other
psychological factors mediating this relationship or perhaps our
specific methodology was not sensitive enough. Future research
should expand on the set of psychological states that we have
The ambulatory BP effects of interacting with supportive ties
were weak in the present study. Of course, whether these data
generalize to other populations (e.g., older adults) and contexts
(e.g., times of high stress) remain questions for further research as
we have found that social support is a stronger predictor of BP in
older adults (e.g., Uchino, Holt-Lunstad, Uno, Betancourt, &
Garvey, 1999). These data, nonetheless, are consistent with a
meta-analysis that showed familial sources of support appeared to
be particularly important predictors of resting BP (Uchino et al.,
1996). The consistency of the effects reported for the familial
categorization highlight the potential importance of interventions
aimed at fostering familial social support.
We also found that interactions with ambivalent network mem-
bers were associated with the highest levels of ambulatory SBP,
independent of the familial categorization. We should emphasize
that these results were not found for measures of ambulatory DBP
or HR. These null results notwithstanding, these data combined
with our earlier laboratory research (Uchino et al., 2001) suggest
that ambivalent ties may be associated with negative effects on
cardiovascular outcomes. One reason for this effect may be due to
the increased interpersonal stress associated with ambivalent net-
work members. Individuals can cope with aversive ties by dis-
counting or withdrawing from such interactions. However, ambiv-
alent ties are more complex, less readily avoided or discounted,
and relatively less predictable, requiring heightened attention and
effort during social interactions, processes that may be associated
with potentiated cardiovascular responses (e.g., Smith, Ruiz, &
Uchino, 2000). In addition, the negativity in ambivalent relation-
ships may be more impactful because people care about these
relationships at some level. Although the analyses of psychological
processes again did not reveal a pattern that could explain the BP
effect, perhaps more specific measures of vigilance, controllabil-
ity, or interpersonal stress would provide stronger tests of these
Overall, the data on ambivalence and ambulatory BP may have
implications for the conceptualization and assessment of social
relationships in the health domain. Much of the prior research on
relationship quality and health has only assessed one dimension
(typically positivity or social support). Even in studies in which
both dimensions were assessed, researchers have typically exam-
ined the effects of one dimension by statistically controlling for the
other (e.g., Finch et al., 1989; Fiore et al., 1983). As outlined in
Figure 1, however, high negativity includes both social aversion
and social ambivalence, whereas high positivity includes both
social support and social ambivalence. Thus, examining either
positive or negative dimensions of social networks in isolation
(i.e., by assessing only a single dimension or statistically control-
ling for one dimension) may mask the different health effects of
these network types and obscure reliable associations between
social relationships and health-related outcomes. For instance, the
adverse effects attributed to negative aspects of relationships in
prior research may be specifically due to the influence of ambiv-
alent ties. Future research will be needed to examine the implica-
tions of our framework for the links between social relationships
and health outcomes.
There are several limitations of the present study that should be
discussed. First, we used a cross-sectional design and a young,
healthy sample. Although ambulatory BP appears to be a strong
predictor of future cardiovascular disorders (Perloff et al., 1983;
Verdecchia et al., 1994), whether these BP differences in our
young sample would result in cardiovascular risk needs to be
determined. Longitudinal studies will be necessary to clarify the
health significance of short-term fluctuations in ambulatory BP as
seen during mood states (e.g., Kamarck et al., 1998) or, in our case,
during social interactions. The nature of the relationships sampled
in this study (e.g., family, friends, significant other, coworkers)
does suggest that such BP-altering interactions occur regularly,
which may increase its long-term implications for health.
We also do not know in any objective sense how compliant
participants were in sampling their social interactions. Although
the DAS suggests adequate compliance, future studies that care-
fully compare event-sampling with interval-contingent sampling
will be needed. In addition, we are unsure to what extent the social
situation influenced the participants compliance or the extent to
which having to monitor an interaction influenced the nature of the
It should also be noted that we did not explicitly
compare BP readings during social interactions with alone condi-
tions. Prior research does suggest that structural measures of social
interactions (e.g., family, spouses) predict lower levels of ambu-
latory BP compared with being alone (Gump, Polk, Kamarck, &
Shiffman, 2001; Spitzer et al., 1991). Nevertheless, the present
study can only draw conclusions about the relative differences in
ambulatory BP during different types of social interactions.
For instance, noncompliance may not be random (e.g., participants
may be less likely to monitor negative interactions than positive interac-
tions). Likewise, it is possible that because the participants knew they must
monitor the interaction, they may have been less likely to get into a heated
argument than if they were not wearing the monitor.
Compared with laboratory paradigms in which interactions with
specific relationships are more tightly controlled (e.g., Gerin,
Pieper, Levy, & Pickering, 1992; Kamarck, Manuck, & Jennings,
1990; Lepore, 1992), the nature of the interactions were also less
specified in the present study. However, in this ambulatory para-
digm with event-contingent sampling, this loss of specificity and
experimental control was offset by increased generalizability be-
cause we were able to sample a wide variety of relationships and
interactions. Furthermore, the ability to measure cardiovascular
function in the everyday life of participants may better capture
their cardiovascular risk (e.g., Perloff et al., 1983). Hence, in
constructing a social psychophysiology of cardiovascular risk and
response (Smith & Gerin, 1998), laboratory and ambulatory ap-
proaches provide complementary methodologies.
Despite possible limitations, there are a number of strengths of
this study that are worth repeating. As noted, the effects of both the
structural and qualitative aspects of ones social relationships on
cardiovascular functioning were examined in a naturalistic context.
This study is also unique in that it extended prior research on social
relationships and health by using event-contingent ambulatory
monitoring over multiple recording days to increase the reliability
and generalizability of these findings. With rare exceptions, most
prior research has used shorter (24 hr) interval-contingent sam-
pling approaches to ambulatory BP monitoring. This approach
may be influenced by atypical days, especially among college
students, because each day may have different structure (e.g.,
particular classes, work, social events, weekend dates) that affects
social interactions (Reis & Wheeler, 1991). Although event-
contingent sampling may be more prone to noncompliance than
interval-contingent sampling, it can be a useful complement to
more traditional methods because it can provide a more controlled
assessment of the event of interest. Finally, the conceptual model
of relationships underlying this research helped to derive more
specific predictions about associations between the quality of the
relationship and physiological outcomes. Further research on this
framework may prove helpful in developing a more complete
understanding of the impact of relationships on health and guide
the design of appropriate interventions to promote positive health
Aiken, L. D., & West, S. G. (1991). Multiple regression: Testing and
interpreting interactions. Thousand Oaks, CA: Sage.
Berkman, L. F. (1995). The role of social relations in health promotion.
Psychosomatic Medicine, 57, 245254.
Berkman, L. F., Leo-Summers, L., & Horwitz, R. I. (1992). Emotional
support and survival after myocardial infarction. Annals of Internal
Medicine, 117, 10031009.
Cacioppo, J. T., & Berntson, G. G. (1994). Relationship between attitudes
and evaluative space: A critical review, with emphasis on the separa-
bility of positive and negative substrates. Psychological Bulletin, 115,
Cacioppo, J. T., Malarkey, W., Kiecolt-Glaser, J. K., Uchino, B. N.,
Sgoutas-Emch, S. A., Sheridan, J. F., et al. (1995). Heterogeneity in
neuroendocrine and immune responses to brief psychological stressors
as a function of autonomic cardiac activation. Psychosomatic Medi-
cine, 57, 154164.
Cohen, S. (1988). Psychosocial models of the role of social support in the
etiology of physical disease. Health Psychology, 7, 269297.
Coyne, J. C., & DeLongis, A. (1986). Going beyond social support: The
role of social relationships in adaptation. Journal of Consulting and
Clinical Psychology, 54, 454460.
Cutrona, C. E. (1986). Objective determinants of perceived social support.
Journal of Personality and Social Psychology, 50, 349355.
Dunkel-Schetter, C., & Bennett, T. L. (1990). Differentiating the cognitive
and behavioral aspects of social support. In B. R. Sarason, I. G. Sarason,
& G. R. Pierce (Eds.), Social support: An interactional view (pp.
267296). New York: Wiley.
Finch, J. F., Okun, M. A., Barrera, M., Zautra, A. J., & Reich, J. W. (1989).
Positive and negative social ties among older adults: Measurement
models and the prediction of psychological distress and well-being.
American Journal of Community Psychology, 17, 585605.
Fiore, J., Becker, J., & Coppel, D. B. (1983). Social network interactions:
A buffer or a stress? American Journal of Community Psychology, 11,
Gellman, M., Spitzer, S., Ironson, G., Llabre, M., Saab, P., DeCarlo Pasin,
R., et al. (1990). Posture, place, and mood effects on ambulatory blood
pressure. Psychophysiology, 27, 544552.
Gerin, W., Pieper, C., Levy, R., & Pickering, T. G. (1992). Social support
in social interaction: A moderator of cardiovascular reactivity. Psycho-
somatic Medicine, 54, 324336.
Gump, B. B., Polk, D. E., Kamarck, T. W., & Shiffman, S. M. (2001).
Partner interactions are associated with reduced blood pressure in the
natural environment: Ambulatory monitoring evidence from a healthy,
multiethnic adult sample. Psychosomatic Medicine, 63, 423433.
Guyll, M., & Contrada, R. J. (1998). Trait hostility and ambulatory car-
diovascular activity: Responses to social interaction. Health Psychol-
ogy, 17, 3039.
Hodgins, H. S., & Zuckerman, M. (1990). The effect of nonverbal sensi-
tivity on social interaction. Journal of Nonverbal Behavior, 14, 155170.
House, J. S., Landis, K. R., & Umberson, D. (1988). Social relationships
and health. Science, 241, 540545.
Kamarck, T. W., Manuck, S. B., & Jennings, R. J. (1990). Social support
reduces cardiovascular reactivity to psychological challenge: A labora-
tory model. Psychosomatic Medicine, 52, 4258.
Kamarck, T. W., Shiffman, S. M., Smithline, L., Goodie, J. L., Paty, J. A.,
Gnys, M., & Yi-Kuan, J. J. (1998). Effects of task strain, social conflict,
and emotional activation on ambulatory cardiovascular activity: Daily
life consequences of recurring stress in a multiethnic adult sample.
Health Psychology, 17, 1729.
Kiecolt-Glaser, J. K., Dyer, C. S., & Shuttleworth, E. C. (1988). Upsetting
social interactions and distress among Alzheimers disease family care-
givers: A replication and extension. American Journal of Community
Psychology, 116, 825837.
Kiecolt-Glaser, J. K., & Newton, T. R. (2001). Marriage and health: His
and hers. Psychological Bulletin, 127, 472502.
Lepore, S. J. (1992). Social conflict, social support, and psychological
distress: Evidence of cross-domain buffering effects. Journal of Person-
ality and Social Psychology, 63, 857867.
Light, K. C., Obrist, P. A., Cubeddu, L. X. (1988). Evaluation of a new
ambulatory blood pressure monitor (Accutracker 102): Laboratory com-
parisons with direct arterial pressure, stethoscopic ausculatory pressure,
and readings from a similar monitor (Spacelabs Model 5200). Psycho-
physiology, 25, 107116.
Little, R. C., Milliken, G. A., Stroup, W. W., & Wolfinger, R. D. (1996).
SAS system for mixed models. Cary, NC: SAS Institute.
Lloyd-Jones, D. M., Evans, J. C., Larson, M. G., ODonnell, C. J., & Levy,
D. (1999). Differential impact of systolic and diastolic blood pressure
level on JNC-VI staging. Joint national committee on the prevention,
detection, evaluation, and treatment of high blood pressure. Hyperten-
sion, 34, 381385.
MacMahon, S., Peto, R., Cutler, J., Collins, R., Sorlie, P., Neaton, J., et al.
(1990). Blood pressure, stroke, and coronary heart disease: Part 1.
Prolonged differences in blood pressure: Prospective observational stud-
ies corrected for the regression dilution bias. Lancet, 335, 765774.
Major, B., Zubek, J. M., Cooper, M. L., Cozzarelli, C., & Richards, C.
(1997). Mixed messages: Implications of social conflict and social
support within close relationships for adjustment to a stressful life event.
Journal of Personality and Social Psychology, 72, 13491363.
Marler, M. R., Jacob, R. G., Lehoczky, J. P., & Shapiro, A. P. (1988). The
statistical analysis of treatment effects in 24-hour ambulatory blood
pressure recordings. Statistics in Medicine, 7, 697716.
Orth-Gomer, K., Rosengren, A., & Wilhelmsen, L. (1993). Lack of social
support and incidence of coronary heart disease in middle-aged Swedish
men. Psychosomatic Medicine, 55, 3743.
Perloff, D., Sokolow, M., & Cowan, R. (1983). The prognostic value of
ambulatory blood pressure. Journal of the American Medical Associa-
tion, 249, 27932798.
Pickering, T., Harchfield, G., Devereux, R., & Laragh, J. (1983). What is
the role of ambulatory blood pressure monitoring in the management of
hypertensive patients? Hypertension, 7, 171177.
Porter, L. S., Stone, A. A., & Schwartz, J. E. (1999). Anger expression and
ambulatory blood pressure: A comparison of state and trait measures.
Psychosomatic Medicine, 61, 454463.
Prisant, L. M., Carr, A. A., Wilson, B., & Converse, S. (1990). Ambulatory
blood pressure monitoring and echocardiographic left ventricular wall
thickness and mass. American Journal of Hypertension, 3, 8189.
Raikkonen, K., Matthews, K. A., Flory, J. D., Owens, J. F., & Gump, B. B.
(1999). Effects of optimism, pessimism, and trait anxiety on ambulatory
blood pressure and mood during everyday life. Journal of Personality
and Social Psychology, 76, 104113.
Reis, H. T., Senchak, M., & Solomon, B. (1985). Sex differences in the
intimacy of social interaction: Further examination of potential expla-
nations. Journal of Personality and Social Psychology, 48, 12041217.
Reis, H. T., & Wheeler, L. (1991). Studying social interaction with the
Rochester interaction record. In L. Berkowitz (Ed.), Advances in exper-
imental social psychology (pp. 269318). New York: Academic Press.
Rook, K. S., & Pietromonaco, P. (1987). Close relationships: Ties that heal
or ties that bind. Advances in Personal Relationships, 1, 135.
Ruehlman, L. S., & Karoly, P. (1991). With a little flak from my friends:
Development and preliminary validation of the Test of Negative Social
Exchange (TENSE). Psychological Assessment, 3, 97104.
Sarason, B. R., Pierce, G. R., & Sarason, I. G. (1991). Social support: The
sense of acceptance and the role of relationships. In Wiley series on
personality processes: Social support. An interactional view (pp. 97
128). New York: Wiley.
Schwartz, J. E., & Stone, A. A. (1998). Strategies for analyzing ecological
momentary assessment data. Health Psychology, 17, 616.
Schwartz, J. E., Warren, K., & Pickering, T. G. (1994). Mood, location and
physical position as predictors of ambulatory blood pressure and heart
rate: Application of a multi-level random effects model. Annals of
Behavioral Medicine, 16, 210220.
Singer, J. D. (1998). Using SAS PROC MIXED to fit multilevel models,
hierarchical models, and individual growth models. Journal of Educa-
tional and Behavioral Statistics, 23, 323355.
Smith, T. W., & Gerin, W. (1998). The social psychophysiology of
cardiovascular response: An introduction to the special issue. Annals of
Behavioral Medicine, 20, 243246.
Smith, T. W., Ruiz, J. M., & Uchino, B. N. (2000). Vigilance, active
coping, and cardiovascular reactivity during social interaction in young
men. Health Psychology, 19, 282392.
Sokolow, M., Werdeger, D., & Hinman, A. (1966). Relationship between
level of blood pressure measured casually and by portable records and
severity of complications in essential hypertension. Circulation, 34,
Spitzer, S. B., Llabre, M. M., Ironson, G. H., Gellman, M. D., & Schnei-
derman, N. (1992). The influence of social situations on ambulatory
blood pressure. Psychosomatic Medicine, 54, 7986.
Stone, A. A., & Shiffman, S. (1994). Ecological momentary assessment
(EMA) in behavioral medicine. Annals of Behavioral Medicine, 16,
Stroebe, W., & Stroebe, M. (1996). The social psychology of social
support. In E. T. Higgins & A. W. Kruglanski (Eds.), Social psychology:
Handbook of basic principles (pp. 597621). New York: Guilford Press.
Uchino, B. N., Cacioppo, J. T., & Kiecolt-Glaser, J. K. (1996). The
relationship between social support and physiological processes: A
review with emphasis on underlying mechanisms and implications for
health. Psychological Bulletin, 119, 488531.
Uchino, B. N., Holt-Lunstad, J., Uno, D., Betancourt, R., & Garvey, T. S.
(1999). Social support and age-related differences in cardiovascular
function: An examination of potential mediators. Annals of Behavioral
Medicine, 21, 135142.
Uchino, B. N., Holt-Lunstad, J., Uno, D., & Flinders, J. B. (2001). Het-
erogeneity in the social networks of young and older adults: Prediction
of mental health and cardiovascular reactivity during acute stress. Jour-
nal of Behavioral Medicine, 24, 361382.
Verdecchia, P., Porcellati, C., Schillaci, C., Borgioni, C., Ciucci, A.,
Battistelli, M., et al. (1994). Ambulatory blood pressure: An independent
predictor of prognosis in essential hypertension. Hypertension, 24, 793
White, W. B., Lund-Johansen, P., & Omvik, P. (1990). Assessment of four
ambulatory blood pressure monitors and measurements by clinicians
versus intraarterial blood pressure at rest and during exercise. American
Journal of Cardiology, 65, 6066.
... The above theoretical predictions are confirmed by a growing body of evidence, based on both experimental and observational studies, showing that ambivalent social ties in personal relationships (e.g., partners, friends, and family) are negatively linked to psychological and physical health-related outcomes. These include greater psychological distress and reduced mental health, as well as higher blood pressure, cardiovascular responses, shorter telomere length, and higher levels of inflammatory markers 2,[5][6][7][8][9][10][11][12][13][14] . Combining the theoretical considerations and empirical findings, it seems justified to assume that ambivalence in social relationships leads to negative health outcomes through persistently elevated distress levels. ...
Full-text available
Ambivalence in social interactions has been linked to health-related outcomes in private relationships and recent research has started to expand this evidence to ambivalent leadership at the workplace by showing that ambivalent supervisor-employee relationships are related to higher stress levels in employees. However, the mental health consequences of ambivalent leadership have not been examined yet. Using a multilevel approach, this study estimated associations of ambivalent leadership with mental health indicators (depression, anxiety, vital exhaustion, fatigue) in 993 employees from 27 work groups. A total effect of ambivalent leadership was found for all four mental health measures, as well as within-group and between-group effects. The consistent relationships of ambivalent leadership with higher symptoms of mental ill-health at the individual- (i.e., within-group) and the group-level (i.e., between-group) support the existence of an un-confounded association, as well as group effects of collective ambivalence.
... Given that social support and relationships have robust effects on health outcomes and morbidity across many different diseases (Holt-Lunstad et al., 2010), researchers have been interested in examining whether co-occurring positivity and negativity in close relationships affect health. Evidence suggests that close relationships characterized by both positivity and negativity are associated with increased stress (Holt-Lunstad et al., 2007), increased cardiac reactivity (Holt-Lunstad et al., 2007;Reblin et al., 2010), and higher blood pressure in daily life (Holt-Lunstad et al., 2003), all of which are linked to poor long-term health (Benjamin et al., 2017). It appears that, despite the positive emotions these relationships elicit, experiencing both positivity and negativity in the context of a close relationship partner is associated with worse outcomes. ...
Full-text available
Research on mixed emotions is flourishing but fractured. Several psychological subfields are working in parallel and separately from other disciplines also studying mixed emotions, which has led to a disorganized literature. In this article, we provide an overview of the literature on mixed emotions and discuss factors contributing to the lack of integration within and between fields. We present an organizing framework for the literature of mixed emotions on the basis of two distinct goals: solving the bipolar-bivariate debate and understanding the subjective experience of mixed emotions. We also present a personalized perspective that can be used when studying the subjective experience of mixed emotions. We emphasize the importance of assessing both state and trait emotions (e.g., momentary emotions, general levels of affect) alongside state and trait context (e.g., physical location, culture). We discuss three methodological approaches that we believe will be valuable in building a new mixed-emotions literature-inductive research methods, idiographic models of emotional experiences, and empirical assessment of emotion-eliciting contexts. We include recommendations throughout on applying these methods to research on mixed emotions, and we conclude with avenues for future interdisciplinary research. We hope that this perspective will foster research that results in the organized accumulation of knowledge about mixed emotions.
... Supervisors relying on a combination of CLB and DLB send conflicting messages to employees about their worth, the quality of their contribution, and the quality of their relationships with them. Multiple studies have shown that relationships seen as both aversive and supportive tend to be more stressful than consistently aversive ones (Herr et al., 2018;Holt-Lunstad, Uchino, Smith, Olson-Cerny, & Nealey-Moore, 2003;Uchino, Holt-Lunstad, Uno, & Flinders, 2001). In addition, mixed messages can thwart employees' need to experience a coherent sense of self, leading to feelings of self-uncertainty (De Cremer, 2003;Lind & van den Bos, 2002;Nahum-Shani et al., 2014;Swann, Rentfrow, & Guinn, 2003), that could further impede thriving and empowered behaviors. ...
Full-text available
This study investigates the within-domain exacerbation phenomenon in relation to employees’ perception of their supervisors’ leadership behaviors. This phenomenon proposes that exposure to supervisors relying on a combination of destructive leadership behaviors (DLB; operationalized as petty tyranny) and constructive leadership behaviors (CLB; operationalized as transformational leadership) should have more negative consequences on followers’ levels of thriving and behavioral empowerment than exposure to supervisors relying more exclusively on DLB or CLB. This phenomenon was tested using a person-centered mixture regression approach with a sample of 2104 Canadian employees from a police organization. Three profiles of employees were identified, representing those exposed to moderately transformational (mostly CLB), destructive (mostly DLB), and inconsistent (CLB and DLB) supervisors. Members of the inconsistent profile displayed the lowest levels of thriving and behavioral empowerment, followed by members of the destructive profile, and finally by members of the moderately transformational profile. Results also suggest that the inability to determine if a supervisor is more destructive or constructive might explain the within-domain exacerbation phenomenon. Indeed, in the inconsistent profile, leadership clarification seemed beneficial for employees. Increases in DLB resulted in a matching increase in empowered behaviors centered on the group and organization, while increases in CLB resulted in increases in thriving and empowered behaviors centered on individual performance.
... The researcher then left the room with an audio-recorder running on the table to allow the dyad to discuss their goals. After 8 minutes had passed, the researcher came back into the room and the dyads completed a post-discussion rating questionnaire adapted from previous research (Holt-Lunstad et al., 2003). This questionnaire consisted of four items, including how stressful the conversation was, how similar it was to normal conversation, if any new information was discussed, and how much they felt they benefited from the discussion. ...
Primary brain cancer is a diagnosis that can have drastic health impacts on patient and caregiver alike. In high-stress situations, dyadic coping can improve psychosocial and health outcomes and communication about personal life goals maybe one way to facilitate this coping. In this study, we describe the feasibility and accessibility of a one-time, self-directed goal discussion pilot intervention for neuro-oncology patients and their primary caregivers. Ten dyads were taken to a private room to complete a pre-discussion questionnaire, a worksheet to elicit personal goals, complete an 8-min discussion of goals, a post-discussion questionnaire, and provided open-ended feedback about the process. Post-discussion, dyads reported that the intervention was not stressful. In open-ended feedback, dyads overwhelmingly reported that the intervention was a positive experience, providing a safe, calm environment to have difficult conversations. This intervention provides a positive framework for improving communication and discussion of goals between patient–caregiver dyads.
For over two decades, the minority stress model has guided research on the health of sexually-diverse individuals (those who are not exclusively heterosexual) and gender-diverse individuals (those whose gender identity/expression differs from their birth-assigned sex/gender). According to this model, the cumulative stress caused by stigma and social marginalization fosters stress-related health problems. Yet studies linking minority stress to physical health outcomes have yielded mixed results, suggesting that something is missing from our understanding of stigma and health. Social safety may be the missing piece. Social safety refers to reliable social connection, inclusion, and protection, which are core human needs that are imperiled by stigma. The absence of social safety is just as health-consequential for stigmatized individuals as the presence of minority stress, because the chronic threat-vigilance fostered by insufficient safety has negative long-term effects on cognitive, emotional, and immunological functioning, even when exposure to minority stress is low. We argue that insufficient social safety is a primary cause of stigma-related health disparities and a key target for intervention.
Introduction Stress from negative life events may be an important risk factor for chronic cardiometabolic conditions, which are increasingly prevalent among young adults. Support from personal networks is known to buffer stress from negative life events. Yet, evidence for these relationships among both young and older adults remains unclear. Methods Longitudinal data came from the University of California, Berkeley Social Networks Study (2015–2018), which followed young (aged 21–30 years) and late middle-aged (aged 50–70 years) adults over 4 years. Weighted hybrid fixed and random effects models (completed in 2020) were used to examine the causal relationships among 4 negative life events, distinct forms of network support (e.g., social companionship, emergency help), and self-reported chronic cardiometabolic disease outcomes (i.e., hypertension, diabetes, or a heart condition). Results Among young adults, both the death of a close tie (average marginal effect=0.10, p<0.001) and financial difficulties (average marginal effect=0.07, p<0.05) were associated with a higher probability of chronic cardiometabolic outcomes. Higher numbers of confidants (average marginal effect= −0.03, p<0.01) and practical helpers (average marginal effect= −0.02, p<0.01) were associated with a lower probability of chronic cardiometabolic outcomes, whereas higher numbers of social companions were associated with a higher probability of having chronic cardiometabolic outcomes among young adults (average marginal effect=0.02, p<0.01). Conclusions Negative life events may be important risk factors for chronic cardiometabolic disease outcomes, particularly among young adults. Although there is no evidence of network support mediating the effects of negative life events, increases in network support were directly associated with chronic cardiometabolic outcomes.
Intensity in adolescent romantic relationships was examined as a long-term predictor of higher adult blood pressure in a community sample followed from age 17 to 31 years. Romantic intensity in adolescence – measured via the amount of time spent alone with a partner and the duration of the relationship – was predicted by parents’ psychologically controlling behavior and was in turn found to predict higher resting adult systolic and diastolic blood pressure even after accounting for relevant covariates. The prediction to adult blood pressure was partially mediated via conflict in nonromantic adult friendships and intensity in adult romantic relationships. Even after accounting for these mediators, however, a direct path from adolescent romantic intensity to higher adult blood pressure remained. Neither family income in adolescence nor trait measures of personality assessed in adulthood accounted for these findings. The results of this study are interpreted both as providing further support for the view that adolescent social relationship qualities have substantial long-term implications for adult health, as well as suggesting a potential physiological mechanism by which adolescent relationships may be linked to adult health outcomes.
Purpose To examine whether social network characteristics of US-and foreign-born individuals are related to hypertension, diabetes and obesity prevalence. Design Cross-sectional. Setting Six San Francisco Bay Area counties. Participants N = 1153 cohorts of young and older adults (21-30 and 50-70 years). Measures Network structure and support measures were calculated using name elicitation and interpreter questions common in egocentric surveys. Hypertension and diabetes were self-reported, and overweight/obesity was determined using body mass index calculations. Foreign-birth status was based on country of birth. Analysis Adjusted and unadjusted logistic regression models were used to examine associations between network characteristics and hypertension, diabetes and overweight/obesity. These relationships were tested for moderation by foreign-birth status, age and gender. Results Higher percentages of family members (AOR = 4.16, CI: 1.61-10.76) and same-sex individuals (AOR = 3.41, CI: 1.25-9.35) in the composition of respondents’ networks were associated with overweight/obesity. Higher composition of family members (AOR = 3.54, CI: 1.09-11.48) was associated with hypertension. Respondents whose networks composed of higher numbers of advice individuals (AOR = 0.88, CI: 0.77-0.99), female respondents (AOR = 0.52, CI: 0.35-0.77) and foreign-born respondents (AOR = 0.54, CI: 0.32-0.92) were less likely to report overweight/obesity. Diabetes was associated with higher composition of individuals living within 5-minutes to respondents (AOR = 5.13, CI: 1.04-25.21). Conclusion Family and network support members such as advice individuals could be potential targets for chronic disease prevention, particularly among older adults and immigrants.
Background: Social network characteristics are associated with health outcomes in later life, including mortality. Moreover, there are well-established mortality disparities across race and ethnicity. Although previous studies have documented these associations separately, limited research considers the two in tandem. The present study addressed how the associations between social network characteristics and mortality differ across race and ethnicity in later life. Methods: Data were from the National Social Life, Health, and Aging Project. At baseline, 3005 respondents were interviewed with regards to their health and social networks. Five years later, 430 respondents had died. Logistic regression models were used to estimate the odds of all-cause mortality over the study period. Results: Network size and kin composition were negatively associated with mortality, whereas density was positively associated with mortality. There was a stronger negative association between the kin composition and mortality for Hispanic respondents compared with white and black respondents. Conclusion: The present study contributes to the large literature documenting the link between social networks and health by highlighting the importance of analyzing networks through a sociocultural lens.
Full-text available
Objective The current study was conducted in a naturalistic treatment setting to examine whether and how perceptions about social engagement, trauma coping self‐efficacy, and posttraumatic stress symptoms (PTS) influence one another across 6 months of psychotherapy for trauma survivors. Method The sample included 183 clients who reported exposure to traumatic events and significant PTS (PCL‐5 ≥ 33). Participants (M age = 37.8, 53.6% female) completed surveys at intake, 3 months, and 6 months into treatment. A cross‐lagged panel analysis was used to test the relationships among perceived social engagement, coping self‐efficacy, and PTS across three assessment points. Results PTS at 3‐months was a mediator in the relationship between intake perceived social engagement and 6‐month coping self‐efficacy and between intake perceived social engagement and 6‐month perceived social engagement. Conclusions PTS several months into treatment may serve as a mechanism between intake perceived social engagement and functional outcomes such as coping self‐efficacy.
Full-text available
Evaluative processes refer to the operations by which organisms discriminate threatening from nurturant environments. Low activation of positive and negative evaluative processes by a stimulus reflects neutrality, whereas high activation of such processes reflects maximal conflict. Attitudes, an important class of manifestations of evaluative processes, have traditionally been conceptualized as falling along a bipolar dimension, and the positive and negative evaluative processes underlying attitudes have been conceptualized as being reciprocally activated, making the bipolar rating scale the measure of choice. Research is reviewed suggesting that this bipolar dimension is insufficient to portray comprehensively positive and negative evaluative processes and that the question is not whether such processes are reciprocally activated but under what conditions they are reciprocally, nonreciprocally, or independently activated. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
In a highly select group of stable hypertensive patients, we have assessed the strength of association between various blood pressure measurements (24 h average automated ambulatory blood pressure, 4 h automated ambulatory morning average blood pressure, multiple office visit average blood pressure, and a single office visit average blood pressure) and various echocardiographic indices of hypertensive cardiac target organ damage (left atrial diameter, left ventricular end diastolic diameter, posterior wall thickness, combined wall thickness, relative wall thickness, left ventricular mass and mass index, and combined wall thickness/left ventricular diastolic diameter ratio). These data demonstrated that a single 24 h average diastolic blood pressure by automatic noninvasive ambulatory monitoring was a significantly better predictor of echocardiographic posterior wall thickness, combined wall thickness or relative wall thickness than the multiple office or single office average diastolic blood pressure. Also there were highly significant correlations between both 24 h average systolic and diastolic blood pressure and these echocardiographic parameters (in descending order of correlation coefficient): combined wall thickness, posterior wall thickness, combined wall thickness/left ventricular diastolic diameter, left ventricular mass index, relative wall thickness, and left ventricular mass. Left ventricular end diastolic dimension did not linearly correlate with any systolic or diastolic blood pressure measurement. Left atrial dimension demonstrated only a significant association with 24 h average diastolic blood pressure. Single office average blood pressure did not linearly correlate with any echocardiographic parameter.(ABSTRACT TRUNCATED AT 250 WORDS)
Conference Paper
This study of 72 undergraduate men examined the effects of two determinants of cardiovascular response-active coping and vigilance-on blood pressure and heart rate responses to social stressors. Observation of a future debate partner (i.e., vigilance) evoked larger increases in blood pressure than did observation of a less relevant person, apparently through the combination of increases in cardiac output and vascular resistance. Preparation and enactment of efforts to exert social influence (i.e., active coping) evoked heightened blood pressure and heart rate responses through increased cardiac contractility and output. Thus, both vigilance and active coping in social contexts increased cardiovascular reactivity, but apparently through different psychophysiological processes.
SAS PROC MIXED is a flexible program suitable for fitting multilevel models, hierarchical linear models, and individual growth models. Its position as an integrated program within the SAS statistical package makes it an ideal choice for empirical researchers and applied statisticians seeking to do data reduction, management, and analysis within a single statistical package. Because the program was developed from the perspective of a "mixed" statistical model with both random and fixed effects, its syntax and programming logic may appear unfamiliar to users in education and the social and behavioral sciences who tend to express these models as multilevel or hierarchical models. The purpose of this paper is to help users familiar with fitting multilevel models using other statistical packages (e.g., HLM, MLwiN, MIXREG) add SAS PROC MIXED to their array of analytic options. The paper is written as a step-by-step tutorial that shows how to fit the two most common multilevel models: (a) school effects models, designed for data on individuals nested within naturally occurring hierarchies (e.g., students within classes); and (b) individual growth models, designed for exploring longitudinal data (on individuals) over time. The conclusion discusses how these ideas can be extended straighforwardly to the case of three level models. An appendix presents general strategies for working with multilevel data in SAS and for creating data sets at several levels.
SAS PROC MIXED is a flexible program suitable for fitting multilevel models, hierarchical linear models, and individual growth models. Its position as an integrated program within the SAS statistical package makes it an ideal choice for empirical researchers and applied statisticians seeking to do data reduction, management, and analysis within a single statistical package. Because the program was developed from the perspective of a "mixed" statistical model with both random and fixed effects, its syntax and programming logic may appear unfamiliar to users in education and the social and behavioral sciences who tend to express these models as multilevel or hierarchical models. The purpose of this paper is to help users familiar with fitting multilevel models using other statistical packages (e.g., HLM, MLwiN, MIXREG) add SAS PROC MIXED to their array of analytic options. The paper is written as a step-by-step tutorial that shows how to fit the two most common multilevel models: (a) school effects models, designed for data on individuals nested within naturally occurring hierarchies (e.g., students within classes); and (b) individual growth models, designed for exploring longitudinal data (on individuals) over time. The conclusion discusses how these ideas can be extended straighforwardly to the case of three level models. An appendix presents general strategies for working with multilevel data in SAS and for creating data sets at several levels.
This study investigated 3 broad classes of individual-differences variables (job-search motives, competencies, and constraints) as predictors of job-search intensity among 292 unemployed job seekers. Also assessed was the relationship between job-search intensity and reemployment success in a longitudinal context. Results show significant relationships between the predictors employment commitment, financial hardship, job-search self-efficacy, and motivation control and the outcome job-search intensity. Support was not found for a relationship between perceived job-search constraints and job-search intensity. Motivation control was highlighted as the only lagged predictor of job-search intensity over time for those who were continuously unemployed. Job-search intensity predicted Time 2 reemployment status for the sample as a whole, but not reemployment quality for those who found jobs over the study's duration. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
examine what is known about the role of early experience in shaping people's perceptions of themselves and of what they can expect from relationships / show how the effects of both perceived support and supportive behavior can be better understood by studying them in the context of the ongoing relationships in which they occur (PsycINFO Database Record (c) 2012 APA, all rights reserved)
discusses concepts and measurement of social support and raises issues such as whether social support should be considered a unitary concept or several related concepts, or how differences in conceptualization are reflected in the ways social support has been measured / relates the functions of social support to the general functions groups serve for the individual / these group functions are discussed from the perspectives of major small group theories: attachment theory, social comparison theory, social influence theory, and exchange or interdependence theory / considers why group members support each other and interprets their motives in initiating and maintaining supportive interactions in terms of interdependence and attribution theory [review] the health outcomes of social support . . . and [present] a selective overview of studies that relate social support to mental and physical health as well as mortality / focuses on the functional relationship of social support to health in terms of main effect and buffering models / suggest different mechanisms to account for the 2 types of relationships (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Two studies, with 58 female and 49 male undergraduates, evaluated potential explanations of the finding that males' same-sex interaction is less intimate than that of females. These explanations concerned differing criteria for intimacy, labeling differences, selectivity in the occasions or partners for intimacy, the question of capability vs preference, and gender-cued stereotypic judgments. In a replication of the essential datum, diarylike reports of naturalistic interaction indicated that males' same-sex interaction was substantially less intimate than that of females. Subsequently, Ss were asked to judge standard stimuli and to have an intimate conversation in a laboratory setting. Analyses revealed that the sex difference could not be attributed to differing criteria, labeling, selectivity, or gender-cued judgments. Further analyses indicated that preference played more of a role in the sex difference than did capability, because situational manipulations eliminated the sex difference. (25 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)