Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The impact of "non-evaluative" and "evaluative" friends on cardiovascular reactivity during a stressful… Kors, Deborah Joy 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1994-0524.pdf [ 987.49kB ]
JSON: 831-1.0087543.json
JSON-LD: 831-1.0087543-ld.json
RDF/XML (Pretty): 831-1.0087543-rdf.xml
RDF/JSON: 831-1.0087543-rdf.json
Turtle: 831-1.0087543-turtle.txt
N-Triples: 831-1.0087543-rdf-ntriples.txt
Original Record: 831-1.0087543-source.json
Full Text

Full Text

The Impact of “Non-Evaluative” and “Evaluative” Friends on CardiovascularReactivity During a Stressful Math TaskbyDeborah Joy KorsB.S., Emory University, 1984A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREWNTS FOR THE DEGREE OFMASTER OF ARTSinTHE FACULTY OF GRADUATE STUDIES(Department of Psychology)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIASeptember, 1994© Deborah Joy Kors, 1994Signature(s) removed to protect privacyIn presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)___________________Department of P hoThe University of British ColumbiaVancouver, CanadaDate (J( I 3 / 9 c( qDE-6 (2188)Signature(s) removed to protect privacyAbstractThis study was conducted to further examine the role of social support in moderatingcardiovascular reactivity to behavioral stress. Specifically, evaluation potential wasmanipulated in order to determine if absence of evaluation is a critical ingredient for thebeneficial effects of support on cardiovascular reactivity. Forty-eight female universitystudents performed a stressful math task while alone or in the presence of a close femalefriend. The friend condition was designed to be either high or low in evaluation potential.This was achieved by manipulating whether the friend could view the subject’s questions andanswers to the math task.Findings indicate that subjects who were assigned to the “non-evaluative” friendcondition showed a trend for reduced systolic blood pressure reactivity compared to subjectswho were alone during the task. Significant differences only emerged during the last minute ofthe task. Subjects in the “evaluative” friend condition did not show differences from the othergroups on any cardiovascular measure, thereby suggesting that a social support manipulationcan fail if evaluative situational elements are not carefully controlled for. The validity of theevaluation vs. non-evaluation distinction was supported by the observation that the friends’cardiovascular activity level during the task showed significant differences in systolic anddiastolic blood pressure. Specifically, the “non-evaluative” friends showed habituation onthese measures during the task, whereas the “evaluative” friends did not habituate, remainingat pre-task levels of response.ftTABLE OF CONTENTSPageAbstract iiTable of Contents iiiList of Tables vList ofFigures viAcknowledgement viiIntroduction 1Effects of social support on health 1What is social support and how does it work9 2Laboratory investigations of social support andpsychophysiological mechanisms 3Objectives of the present study 6Methods 8Subjects 8Measures 8Procedure 10Data analysis 13Results 15Demographics and psychological measures 15Cardiovascular measures 18Friends’ data 28Discussion 34Appendix 1 40ifiAppendix2.42Appendix 3 47References 50lvList of TablesPageTable 1. Demographics of subjects by condition. 16Table 2. Subjects’ responses to the ffiendship questionnaire by condition. 17Table 3. Post-task visual analog scales and math task performance 19by condition.Table 4. Pearson product correlation coefficients for visual analog 20scale! friendship dimensions and cardiovascular response to task.Table 5. Cardiovascular readings during baseline and task by condition. 21Table 6. Change scores for all cardiovascular measures by condition. 22Table 7. Post-task visual analog scales by friendship condition, friends’ ratings. 29Table 8. Friends’ cardiovascular readings during baseline and task by condition. 30Table 9. Friends’ change scores for all cardiovascular measures by conditon. 31VList of FiguresPageFigure 1. Systolic blood pressure response over time. 24Figure 2. Diastolic blood pressure response over time. 25Figure 3. Heart rate response over time. 26viAcknowledgementI wish to acknowledge a number of people who have played an important role during variousstages of this research. I am most grateful to my advisor, Wolfgang. He has been wonderfulto work with and has allowed me the freedom I needed to develop as an academician. I amalso indebted to Tracey, Carmen, Jeff, Joseph and Peter, as well as other members of my lab,for their ongoing assistance and feedback. My friend Karen, as well, has been a tremendoussource of support throughout the entire process. She has added humor to my life, duringtimes when it was most needed. Finally, I want to thank my mother and father for alwayshaving been there for me. Their love and encouragement throughout my life has instilled theconfidence in me to pursue my goals.vii1Effects of Social Support on HealthThe past 20 years have witnessed a burgeoning of research on the beneficial effects ofsocial relationships for health. It has been shown, via retrospective and prospective studies,that individuals who are more socially isolated and without adequate levels of support are atgreater risk for certain diseases and death. Several large scale epidemiological studies(Berkman & Syme, 1979; Blazer, 1982; House, Robbins, & Metzner, 1982; Kaplan,Salonen, Cohen, Brand, Syme, & Puska, 1988; Orth-Gomer & Johnson, 1987; Schoenbach,Kaplan, Fredman, & Kleinbaum, 1986) have shown a strong relationship between socialintegration and all-cause mortality. A number of these studies (Kaplan et al., 1988; OrthGomer & Johnson, 1987) have also demonstrated a link between social integration andmortality from cardiovascular disease. While these studies consistently suggest thatindividuals with minimal social contact have a higher risk of mortality, the magnitude of therisk has varied across studies, and the pathway has remained a source of speculation. Also themeasurement of social relationships has varied between studies.Unfortunately, the research linking support to coronary heart disease (CHD) morbidity(as well as the link between support and other diseases) does not appear to be as strong as themortality data. Several studies have shown that social integration was associated with theprevalence of CHD and angina pectoris (Joseph, 1980; Reed, McGee, Yano, & Feinleib,1983) but this relationship did not hold for CHD incidence (Reed et al., 1983). Another study(Seeman & Syme, 1987) found that one’s feeling of being loved was a significant predictor ofcoronary atherosclerosis. Social integration, however, was not related to atherosclerosisdevelopment. Moreover, the love and support of one’s wife was found to reduce the risk ofdeveloping angina pectoris, but this only remained true for men experiencing high levels ofanxiety (Medalie & Goldbourt, 1976). Finally, Haynes and Feinleib (1980) reported thatfemale clerical workers with non-supportive bosses were at increased risk for the developmentof coronary heart disease.2Other studies have examined the impact of social support after an initial myocardialinfarction (MI). One study (Berkman, Leo-Sumers, & Horwitz, 1992) found that emotionalsupport was independently related to the risk of death during the 6 months following an Ml, inan elderly sample. Another study (Case, Moss, Case, McDermott, & Eberly 1992) found thatfollowing an initial MI, living alone was an independent risk factor for the recurrence of amajor cardiac event.These studies are strongly suggestive of a link between social relationships and CHDmorbidity, but the morbidity studies have greatly differed with respect to study populationsand endpoints in the disease process that have been looked at. These differences, in part, mayhave contributed to a lack of consistency among the findings.What is Social Support and How Does it Work?The mechanism by which social support influences CHD morbidity is far from clear.Two mechanisms have been posited in the past to explain the general relationship betweensocial relationships and health. The main effects model hypothesizes that social support has adirect beneficial influence on health, regardless of level of stress. In contrast, the bufferinghypothesis suggests that there is an interaction between social support and stress. That is,social support buffers the individual from the adverse effect of stress, thereby preventing theonset of disease. The individual’s perception of support, rather than actual support received,appears to be most important for buffering to take place (Cassel, 1976; Cohen & McKay,1984; Cohen& Wills, 1985).With both models, support is hypothesized to work by influencing health behaviors ofthe individual or by impacting directly on biological processes (e.g., neuroendocrine,hemodynamic, and immune responses), thereby decreasing the likelihood of disease. Supportmay also indirectly affect biological functioning, through its influence on behavior (e.g., activecoping) (Cohen, 1988).Since the buffering hypothesis speculates that support is beneficial primarily when anindividual experiences stress, two time points have been specified for which support could3impact on the stress-disease causal chain. Support may prevent an individual from viewing asituation as stressful, thereby preventing the stress response (Cohen & McKay; House, 1981).In terms of Lazarus’s model, support would impact on the primary appraisal of the stressor,making it less likely that the event is perceived as stressful (Lazarus & Folkman, 1984). Theother possibility is that support may intervene after the situation has been appraised asstressful, but prior to the onset of disease. Specifically, support may facilitate positivebehaviors (e.g. coping, health behaviors), decrease affective response, or reduce physiologicalreactivity to the stressor (Cohen, 1988; Cohen & Wills, 1985; House, 1981). In this case,support would influence the secondary appraisal of the stressor, facilitating one’s perceivedability to cope with the stressful event (Lazarus & Folkman, 1984).Since the mechanism by which social support impacts upon disease development is notknown, it is also not clear at what point in the disease process (i.e. onset, progression,recovery) that support would have the most pronounced protective function. According tothe reactivity hypothesis, individuals who experience excessive cardiovascular responses tostress may be at greater risk for coronary heart disease (Manuck & Krantz, 1986).Presumably, then, social support might buffer the individual from the impact of thesestressors, thereby preventing or decreasing the physiological responses to the stressfulsituations. Thus, social support might prevent the individual from appraising the event asstressful or it might intervene after the stressful event, by accelerating recovery from thephysiological stress response (House, 1981). In the latter case, support would have a directpsychophysiologic influence, dampening the neuroendocrine activation so that the individual isless reactive to stress.Laboratory Investigations of Social Support and Psychophysiological MechanismsThere have been several studies looking at the influence of social support oncardiovascular reactivity in the laboratory, because the lab allows controlled manipulationsand facilitates tests of the mechanisms (Allen, Blascovich, Tomaka, & Kelsey, 1991; Edens,Larkin, & Abel, 1992; Kamarck, Manuck, & Jennings, 1990; Sheffield & Carroll, 1993).4Kamarck, Manuck, and Jennings (1990) had female subjects do a serial subtraction math taskeither in the presence of a friend or by themselves. The friends continuously touched thesubjects on the wrist throughout the baseline and task, while listening to headphones ofwhitenoise and filling out questionnaires of their own. Thus, the friend could not hear what thesubject said (or presumably see them) during the task. They found that subjects showeddecreased reactivity (systolic blood pressure and heart rate) in the presence of their friendswhen compared to subjects who did the math task alone. Thus, the friend was thought to be asource of social support, buffering the subject from the stressflulness of the task.Edens, Larkin, and Abel (1992) tried to replicate the Kamarck et al. (1990) study, withseveral additional conditions. Female subjects performed a serial subtraction task while alone,in the presence of a friend, or in the presence of a stranger. The friend or stranger conditionswere further subdivided according to whether they touched the subject on the wrist (as theydid in Kamarck et al.’s (1990) study) or were present but did not touch the subjects on thewrist. Thus, five conditions were included in this study: an alone condition, a friend/no-touchcondition, a friend /touch condition, a stranger/no-touch condition and a stranger/touchcondition. In all cases, the friend or strangers filled out questionnaires and listened toheadphones ofwhite noise, thereby preventing them from attending to what the subject wasdoing. Subjects assigned to both “Stranger” conditions showed significantly greatercardiovascular reactivity to the stressor than those assigned to the “Friend” conditions.Subjects in both “Touch” conditions (with either strangers or friends), showed greaterreactivity than those in the “No-Touch” conditions. This is in direct contradiction to theKamarck et a!. (1990) study which found that friends touching the subjects was associatedwith a decrease in reactivity. However, the “Friend-No-Touch” group showed decreases inreactivity (systolic blood pressure) when compared to the “Alone” group. These researchersconcluded that it was the presence of the friend, and not the physical touch, that constitutedsupport and led to the decrease in reactivity.5A third study (Allen et al., 1991) had female subjects perform a math task in theirhome either alone, in the presence of a close friend, or in the presence of their pet dogs. Inthis study, the friends were not wearing headphones and were not completing questionnairesduring the task. Allen et al. (1991) found that female subjects showed decreased reactivity(systolic blood pressure and pulse rate) in the presence of their pet dog, when compared todoing a math task alone. In contrast, these researchers found that the presence of the friendled to an increase in systolic blood pressure and heart rate reactivity. These results seem toconflict with the results of the Kamarck et al. (1990) and Edens et al. (1992) studies. Allenet al. (1991) speculated that the friends in their study were perceived as evaluative, since theywere able to look at the subjects performing the tasks and could obviously hear the subject’sresponses. A perception of being evaluated while performing a mentally challenging task,then, would understandably lead to increases in reactivity. In contrast, the pets were positedto be an inherently non-evaluative source of support, thereby leading to a decrease inreactivity. Since the friends in the Kamarck (1990) and Edens (1992) study were not able tohear or see what the subject was doing during the task, one could assume that they too wereperceived as non-evaluative (“pet-like”), also leading to decreases in reactivity.More recently, Sheffield and Carroll (1993) compared the cardiovascular responses ofand female subjects during a serial subtraction task performed while alone, in thepresence of a friend, or in the presence of a stranger. The authors reported no significantdifferences in reactivity across these conditions. Hence, they did not find attenuated reactivityeven in the presence of a friend, and it did not matter whether the additional person in theroom was well known to the subject or not.In light of the Allen et al. (1991) study, however, these results would be predicted,since the subject may have perceived both the stranger and the friend as evaluative. That is, inthis study (Sheffield & Carroll, 1993), neither strangers nor friends wore headphones or filledout questionnaires. Thus, both the confederates and friends could fully attend to what thesubject was doing and hence be perceived by the subject as evaluative. Consequently, the6subjects would be expected to show similar increases in reactivity as did the subjects in Allen’s(1991) study.Objectives of the Present StudyThe present study was designed to resolve the above discrepancies. We hypothesizedthat performing a stressful math task in the presence of a friend who is perceived as non-evaluative is essential for the subject to feel supported. Such a condition would be expectedto buffer the subject from the stressftilness of the math task and should lead to a decrease inreactivity, compared to performing the task alone. In contrast, it was hypothesized that thesame task performed with a friend who is perceived to be evaluative would lead to an increasein cardiovascular reactivity, when compared to an alone condition, for failure to provide a truesupport condition.In the present study, the evaluation potential of the friend was manipulated by allowingthe friend to be more or less “tuned in” to the subject’s task performance by seeing or notseeing the subject’s questions and answers. Thus, the “non-evaluative” condition was one inwhich the friend was in the room with the subject during the task, but was not able to see thequestions that the subject received nor the answers. Also, in the non-evaluative condition, thefriend was occupied herself. Since the impact of touch on cardiovascular reactivity appears tobe quite controversial (with Kamarck et al. (1990) and Edens et al. (1992) showing discrepantresults in response to touch), touch was not used in the present study. In contrast, the“evaluative” condition was manipulated by having the subject exposed to exactly the sametask, with the friend able to see the questions as well as the answers that the subjects wrote,and with the friend appearing attentive to the subject’s performance.Since no previous studies have looked at the physiological reactivity of the“supportive” friends, the present study also monitored the friends’ cardiovascular response. Itwas predicted that those friends assigned to an evaluative condition would be morephysiologically aroused than those assigned to a non-evaluative support condition.Finally, while previous studies (with the exception of the Sheffield et al. study) haveused primarily female subjects, recent evidence suggests that there may be important genderdifferences in social support (Linden et al., 1993). Consequently, it seems important toinvestigate the impact of social support on cardiovascular reactivity in women and men.Although this study initially used female subjects (because it was aimed at resolving thediscrepancy between two studies that used only female subjects), a replication of this studywith males is planned after the completion of this study.78MethodsSubjectsForty-eight female university students were recruited from undergraduate psychologycourses. All subjects were also asked to recruit a close female friend who could come withthe subject to the laboratory session. Same-sex friends were requested in order to decreasepotential ambiguity present in opposite-sex interactions. In return for their participation,subjects were offered course credit and friends were offered a small gift (value under 5$).Inclusion and exclusion criteria. Prior to scheduling of the laboratory session, subjectswere asked whether they had any diseases or chronic illnesses, as well as whether theyregularly used any medications. Subjects were not eligible for participation if they had beendiagnosed with essential hypertension or diabetes. Additionally, individuals who were takingmedications with cardiovascular effects were not included in the study.MeasuresMeasures of cardiovascular reactivity. An automated blood pressure monitor(Dinamap 845, Adult/Pediatric Vital Signs Monitor: Critikon Corp, Tampa, Fl.) was used tomeasure systolic blood pressure (SBP), diastolic blood pressure (DBP), and heart rate (FIR).Standard occluding cuffs were attached to both the subject’s and the friend’s non-dominantarms. Cardiovascular measures were taken during baseline, task, and recovery phases by anexperimenter located in an adjoining room. Specifically, cardiovascular measures taken atminutes 9 and 13.5 of a 15 minute adaptation period served as baseline readings for allgroups. During the 5 minute math task, cardiovascular responses were recorded at minutes 1,2.5, and 4.5, respectively. Recovery measures were taken at minutes 1 and 5 of a 5 minuterecovery period. (See Appendix 1 for data collection sheet.)Psychological measures. Subjects and friends completed several questionnaires duringthe baseline and recovery periods. These included: a student version of the InterpersonalSupport Evaluation List (ISEL) (Cohen, Mermeistein, Kamarck, & Hoberman, 1985), anabbreviated version of the Fear ofNegative Evaluation Questionnaire (FNE) (Leary, 1983),9the Self-esteem scale (Rosenberg, 1965), the Internal/External Locus of Control (Rotter,1966), and the Balanced Inventory ofDesirable Responding (Paulhus, 1984).The ISEL (Cohen et al., 1985) is a 48-item scale that assesses perceived availability ofsupport throughout the individuals support network. The scale is subdivided into four 12-item subscales which measure appraisal, belonging, tangible, and self-esteem support, inaddition to providing an overall support score. Test-retest reliability (for the entire scale) isapproximately .87, and internal consistency ranges from .77 to .86. As a measure of constructvalidity, the ISEL shows a correlation of .46 with the ISSB (Inventory of Socially SupportiveBehaviors). In addition, correlations of-.52 to -.64 have been reported between the ISEL andmeasures of social anxiety, thereby providing evidence of its discriminant validity. The ISELhas also been found to be free of social desirability bias.The Brief FNE (Leary, 1983) is a 12-item scale measuring the fear of being evaluatednegatively by others and specifically indicates a fear of losing social approval. This scaleshows a test-retest correlation of .75 (over a four week period), and its internal consistency is.90. The Brief FNE is also highly correlated with the full length FNE.The Self-Esteem Scale (Rosenberg, 1965) is a 10 item scale that measures globalfeelings of self-worth and self-acceptance. Internal consistency and test-retest reliability areapproximately .77 and .85, respectively. As evidence for its convergent validity, a correlationof .72 has been reported between this scale and the Lerner Self-Esteem Scale.The Internal-External Locus of Control (Rotter, 1966) is a scale that measures thedegree to which an individual believes that specific outcomes are determined by personalcharacteristics or by external forces. This 23-item scale consists of pairs of statements (oneinternal and one external statement) for each question, with one point given for each externalstatement selected. Internal consistency and test-retest reliability are found to be .70 and .72,respectively.The BIDR (Paulhus, 1984) is a 40-item scale which forms two separate subscales. TheSD subscale reflects exaggerated claims regarding one’s cognitions and private behaviors,10while the OD subscale (or impression management scale) reflects exaggerated claims withrespect to overt behaviors. Internal consistency of these scales has been reported to be .68 to.80 for the SD scale, and .75 to .86 for the OD subscale. Test-retest correlations have beendemonstrated to be .69 and .65 for the SD and OD scales, respectively. As a measure ofvalidity, the BIDR has been found to show a .71 correlation with the Marlowe-Crowne scale(Paulhus, 1988) and .80 with the Multidimensional Social Desirability Inventory (Jacobson,Kellogg, Cauce, and Slavin, 1977). In addition, factor analyses have shown the SD and ODscales to form separate factors.Questionnaire completion was divided between adaptation and recovery periods inorder to keep subjects filly occupied during these periods. Questionnaire data were primarilyused to ensure that there were no preexisting differences between groups. Immediatelyfollowing each math task, subjects completed 10 cm. visual analog scales. These scalesmeasured their degree of stress, degree of confidence (of their abilities during the task),degree of support and evaluation in the presence of their ffiend, as well as how competitivethey felt with their friend. The friends were provided with the same visual analog scales, andwere asked to rate how supportive, evaluative, and competitive that they felt toward thesubject during the task, as well as how stressed they felt during the task. These scales wereincluded as both a manipulation check for the friend condition, as well as a way of determiningexperiential differences in subjects and friends across the two friend conditions. Finally, bothsubjects and friends completed a questionnaire asking them to rate the closeness of theirfriendship.(See Appendix 2 for visual analog scales and the friendship questionnaire.)ProcedureMental arithmetic tasks. The present study employed a 5-minute math task which wasdelivered by video tape. The task consisted of 30 equations, with one question deliveredevery 10 seconds. The equations in the task were either addition, subtraction, multiplication,or division. Subjects viewed the questions on a television screen and wrote their answers on awhite board, which was numbered 1-3 0. Subjects were instructed to work as quickly as11possible and were told that they would be evaluated for both speed and accuracy. Subjectswere also told to proceed to the next question, if they were not able to answer the previousquestion within the allotted 10 seconds. Transcribing the question to the white board was notpermitted, and subjects were required to perform all calculations mentally.The task used in this study has been shown to reliably trigger increases ofapproximately 10 mm/Hg in systolic and diastolic blood pressure, as well as an increase ofapproximately 10 bpm in heart rate (Linden, 1991). Women have been shown to be asresponsive as men to this type of task, and habituation of response does not occur during thetask period,Setting. The experiment took place in a sound-proofed room, measuringapproximately 3 m. X 4 m. Subjects were seated upright on a hard chair and faced a whiteboard which was positioned near their dominant hand. A television monitor, which displayedthe task, was located within 1.5 meters of the subject and was angled toward the side oppositethe white board. When present, the friend (whether “evaluative” or “non-evaluative”) wasseated to the side of the subject, approximately 1.5 meters away and at about a 90 degreeangle from the subject. The “evaluative” or “non-evaluative” friend was seated at a distanceand angle vis-a-vis the subject that remained the same in both conditions.In the non-evaluative condition, the white board was angled so that the friend was notable to see the subject’s answers to the math task. In addition, a shield was placed on the sideof the t.v. nearest the friend, so that the friend could not see the math questions given. Thefriend was seated in an “easy chair” to allow for a comfortable body position during the task.In order to keep the friend occupied during the task, picture books were placed near thefriend’s chair for the friend to look through during the task.In the evaluative condition, the white board was rotated so that the friend could clearlysee the subject’s answers. In addition, the shield was removed from the t.v., so that the friendcould see the questions that the subject responded to. The friend was seated in an upright12chair and no picture books were provided in order to encourage the friend to be attentive tothe subject’s task.Design. This study is a between design, with subjects performing the math task eitheralone or with an “evaluative” friend or with a “non-evaluative” friend. Subjects wererandomly assigned to one of these 3 conditions with 16 subjects per condition. Subjectsassigned to the alone condition were still asked to bring a friend with them, although theirfriend did not actually participate in the study, and, instead waited for the subject in a waitingarea down the hail.Specific procedures. Prior to participating, subjects were asked to refrain fromsmoking, drinking alcohol, exercising, or ingesting caffeine for at least 2 hours prior to theirsession in the laboratory. Friends were also asked to comply with these instructions, sincetheir cardiovascular responses were monitored as well (with the exception of the friends thatcame to the laboratory with subjects assigned to the alone condition). All subjects werecontacted by telephone to remind them of these instructions on the evening prior to their studyvisit. Upon their arrival at the laboratory, after study procedures were explained, both thesubject and friend signed consent forms. Thereafter, blood pressure cuffs were attached.Subjects, in all conditions, had a 15 minute pre-task adaptation period. Only in thealone condition was the entire adaptation without the friend present. Otherwise, in both theevaluative and non-evaluative conditions, after 5 minutes of alone adaptation, the friendentered the room and was attached to their blood pressure monitor. Adaptation samplingoccurred during the remaining 10 minutes for all groups. The friend was included in theadaptation period to ensure that resting cardiovascular measures were not affected by thesocial presence of the friends, in the absence of a task challenge. Additionally, since thefriend’s cardiovascular responses were monitored during the task, it was necessary to have abaseline period for the friend, as well. During the remaining time that the friend was in theroom, her reactivity was monitored at the same time points as that of the subject.13The 5 minute math task was then performed in either the alone, evaluative, or non-evaluative conditions, according to randomization. Following the task was a 5 minuterecovery period, after which subjects and friends were debriefed separately.All subjects (whether assigned to the evaluative or non-evaluative condition) weregiven the same instructions, namely that the role of their friend was to provide nonverbalsupport for them during the task. Instructions to the friends were the same in both conditions,with the only difference being that “non-evaluative” friends were informed that they would notbe able to see the subject’s questions and answers, while “evaluative” friends were told thatthey would be able to see the subject’s questions and answers. (The “evaluative” friends wereasked, however, to not give any feedback to the subjects, either verbally or by nodding theirheads).Specifically, both sets of friends were instructed to be nonverbally supportive of thesubjects. They were provided with some concrete suggestions of “support”, but wereencouraged to do what seemed natural in order to “support” their friend. Friends assigned toboth conditions were told that they should not talk or actively interact with the subject duringthe baseline or task periods and should not stare continuously at the subject. Additionally,during the task, “non-evaluative” friends were asked to browse through several picture books,in order to decrease their attentiveness towards the subject during the task period. (SeeAppendix 3 for text of experiment instructions.)Data AnalysisRepeated measures analyses of covariance were conducted for systolic blood pressure(SBP), diastolic blood pressure (DBP) and heart rate (HR) measures to see if task responsesdiffered for subjects performing the task either alone, with an “evaluative” friend present, orwith a “non-evaluative” friend present. In addition, visual inspection of the data was used todetermine if time trends should be examined.MANOVAs were conducted on the self-report measures and on task performanceindices to compare differences between the three conditions. Additionally, Pearson product-14moment correlations were calculated for all state measures (visual analog scales) with SBP,DBP, and HR change scores to determine if a relationship existed between affect induced bythe task and cardiovascular reactivity. Each item of the friendship questionnaire was alsocorrelated with these cardiovascular measures, in order to see if there were any friendshipdimensions that were associated with cardiovascular response.Repeated measures analyses of covariance and MANOVAs were performed on thefriends’ cardiovascular and self-report data, respectively. These analyses were conducted inorder to determine if the two friend conditions showed differences in SBP, DBP, and HR, aswell as differences in the self-report measures.All multivariate analysis of variance (MANOVA) procedures were preceded by testsof normality, as well as tests of homogeneity of covariance via the Box test. Repeatedmeasures design analysis (i.e. repeated measures ANOVAs and repeated measuresANCOVAs) employed the Mauchly test of sphericity and used the Greenhouse-Geisserepsilon as a correction factor, when necessary. In addition, homogeneity of regression wastested prior to all ANCOVA analysis. All statistical inferences were drawn at the .05 alphalevel.15ResultsDemographics and Psychological MeasuresTable 1 presents the means and standard deviations of each condition on severaldemographic variables. The average age of subjects participating in this study was 20.4 yearsand was not significantly different among the three conditions ( (2, 43) =58, p=.57). Ascan be seen in Table 1, the three conditions also had similar ethnic composition, with amajority of Asian subjects, and with Caucasian women comprising a much smaller percentageof each condition.A multivariate analysis ofvariance (MANOVA) of subjects’ responses to thefriendship questionnaire indicated that there were no significant differences between theconditions in terms of perceived closeness to their friend, as well as on the other friendshipvariables (F (10, 82) .62, p = .79). (See Table 2).Moreover, there were no significant differences among the conditions on any of thetrait psychological questionnaires. In addition, the subjects’ perceptions of their stress,performance and confidence during the math task, as measured by post-task visual analogscales, showed no differences between the three conditions (multivariate (6, 86) = .55, p =.77). Performance on the math task, as evidenced by the number of correct responses, alsodid not show any differences among the three conditions, i (2, 45) = 1.57, p = .22. (SeeTable 3).In order to determine if the manipulation of perceived evaluation was successful, thenon-evaluative and evaluative conditions were compared, via ANOVAs, in their perception ofsupport and of evaluation. While these two conditions were not significantly different in theirperception of support ( (1, 29) = .02, p=. 89), there was a trend for evaluative subjects tofeel more evaluated than the non-evaluative condition ( (1, 29) = 3.32, p = .079).Additionally, subjects assigned to the evaluative condition tended to perceive their friends asmore competitive than those assigned to the non-evaluative condition (E (1,29) = 4.11,p=.O52), and there was a trend for evaluative subjects to feel less comfortable in the presence16Table 1. Demographics of Subjects by ConditionCondition Age Ethnic originMean (SD)12 AsianAlone 20.6 (5.1) 2 Caucasian1 Afro-American1 Aboriginal10 AsianEvaluative 21.4 (7.1) 2 Caucasian4 unidentified10 AsianNon-evaluative 19.4 (2.2) 5 Caucasian1 Aboriginal17Table 2. Subjects’ Responses to the Friendship Questionnaire by ConditionFriendship dimension Alone Evaluative Non-evaluativeMean (SD) Mean (SD) Mean (SD)Length of friendship(yrs.) 4.6(5.3) 5.0(5.6) 4.9(5.1)Length of time feelingclose to friend (yrs.) 3.5 (4.2) 2.9 (3.2) 3.9 (5.3)Frequency of contactwith friend 7.9(1.3) 7.0(2.0) 7.2(2.0)(0not very often;lOvery often)Perceived closeness tofriend 8.0(1.3) 7.0(1.7) 7.6(1.6)(O=not close at all;lOvery close)Perceived comfortwithfriend 8.8(.98) 8.1(1.5) 8.6(1.0)(O=not comfortable atall;1O=very comfortable)18of their friends, than their non-evaluative counterparts (E (1, 29) = 2.87 p=. 10). (See Table3.)Correlations between self-report data and cardiovascular reactivity. Table 4 shows thecorrelations between the visual analog data and cardiovascular change scores, as well asbetween the friendship dimensions and the change scores. While most correlations were weakand non-significant, greater confidence during the math task was associated with decreasedheart rate reactivity (r = -.358). Additionally, perceived closeness to the friend showed amoderate negative correlation with SBP change scores (t = -.43 4). Similarly, the length of thefriendship and the length of time the subject had felt close to her friend were also negativelycorrelated with SBP reactivity (r = -.498 and -.431, respectively). (The length of thefriendship and length of time that the subject felt close to the friend were highly correlated at.868.)Cardiovascular MeasuresThe baseline and math task group means for SBP, DBP, and HR are displayed in Table5. The table provides the 2 baseline readings prior to the task, as well as the three math taskreadings taken during the 5 minute math task for all cardiovascular parameters. In addition,SBP, DBP, and HR change scores are provided in Table 6.Baseline cardiovascular measures. ANOVAs were performed separately on eachbaseline reading, as well as on the mean of the two baselines, to see if there were anydifferences between the 3 conditions. All analyses suggested equivalence on eachcardiovascular baseline measure. Consequently, the presence of a friend (either evaluative ornon-evaluative) did not appear to affect resting blood pressure or heart rate values, whencompared to those subjects who were alone during the baseline period.Cardiovascular measures associated with task. In order to compare the conditions oneach cardiovascular measure, repeated measures ANCOVAs were conducted for SBP, DBP,and FIR, with the second baseline reading serving as the covariate. (The second19Table 3. Post-Task Visual Analog Scales and Math Task Performance by ConditionScale Alone Evaluative Non-evaluativeMean (SD) Mean (SD) Mean (SD)Evaluation 4.3(2.1) 3.0(2.1)Support 4.9(2.8) 5.1(2.3)Competition 4.2 (2.4) 2.6 (2.0)with friendComfort with 5.9 (2.3) 7.2 (1.8)friend presentMath task 23.8 (5.0) 20.7 (5.1) 22.2 (4.6)performanceNote: All visual analogue scales are 10 point scales, with 0 being the lowest and 10 being thehighest endpoint. Math task performance reflects the total number of correct responses froma total of 30 questions. The first four variables did not apply to the alone condition, hence thedashes in the table.20Table 4. Pearson Product Correlation Coefficients for Visual Analog Scale! FriendshipDimensions and Cardiovascular Response to the Task (for Subjects)Scale SBP DBP HRchange scores change scores change scoresDegree of stress .036 -.219 .191Confidence -.093 .053 .358*Performance .076 .058 -.299Support -.092 -.191 -.157Evaluation -.037 -.330 .069-.006 .035 .285Competition withfriend-.153 .090 -.384Comfort withfriend present.498*-.044 -.185Length offriendship.431* .005 -.177Length of timefeeling close tofriend-.080 -.064 .082Frequency ofcontact with friend_434*-.180 -.328Perceived closenessto friend-.217 -.079 -.256Perceived comfortwith friendNote: Correlations were computed for subjects assigned to a friend condition. However,correlations of stress, performance, and confidence included all 3 conditions.* l2 .0121Table 5. Cardiovascular Readings During Baseline and Task by ConditionSBP BL1 BL2 MT1 MT2 MT3(mmllg) Mean Mean Mean Mean Mean(SD) (SD) (SD) (SD) (SD)113.7 109.5 125.8 120.9 122.8Alone (106) (9.6) (11.0) (9.4) (8.8)110.0 107.3 118.9 117.4 116.4Evaluative (11.6) (11.8) (14.0) (12.1) (12.3)111.1 110.3 121.4 117.0 114.9Non- (10.1) (9.3) (11.9) (12.1) (9.8)evaluativeDBP(mmHg)70.3 69.9 77.7 74.9 75.8Alone (4.6) (6.9) (10.8) (9.2) (9.1)68.8 71.3 76.8 75.7 75.3Evaluative (8.2) (9.6) (8.0) (8.8) (7.7)70.8 67.7 78.7 75.6 73.5Non- (5.5) (7.6) (9.3) (8.2) (8.6)evaluativeHR(bpm)74.2 73.3 88.8 82.2 83.6Alone (11.9) (11.7) (15.6) (14.1) (13.0)75.1 75.0 90.2 83.2 85.0Evaluative (10.7) (8.0) (12.2) (11.0) (10.8)76.9 75.9 88.1 83.8 83,1Non- (8.6) (8.9) (12.1) (12.4) (10.4)evaluativeKey: BL1 = baseline 1; BL2 = baseline2; MT1 = first task reading (minute 1); MT2 =second task reading (minute 2.5); MT3 = third task reading (minute 4.5)Table 6. Change Scores for all Cardiovascular Measures by ConditionGroup SBP change score DBP change score HR change scoreMean (SD) Mean (SD) Mean (SD)Alone 13.6(7.0) 6.2(7.0) 11.6(8.3)Evaluative 10.3 (9.0) 4.6 (7.3) 11.1 (8.6)Non-evaluative 7.5 (8.3) 8.3 (6.1) 9.0 (5.9)2223baseline was used because it was found to be more stable than the first baseline value.) Therepeated measures analysis for SBP revealed a trend for condition (E (2, 44) = 2.64, p = .083),and there was a main effect for time (F (2, 44) = 5.46, p = .008). No interaction betweencondition and time period emerged. Follow-up orthogonal contrasts found that the alonecondition was significantly different from the non-evaluative subjects ( (1, 44) 5.04, p.03). No other group differences could be detected. When the repeated measures ANCOVAwas conducted for DBP, no significant main effect for group was found (E(2,44) = .54, p =.59). Similarly, the repeated measures ANCOVA conducted on HR revealed no significantdifferences between the conditions (E(2,44) = .41, p = .66). To test the stability of theseresults, all analyses were repeated with change scores and revealed the same pattern.In preparation for a possible test of time trends, the means of each baseline and taskreading were plotted for each condition. Figures 1, 2, and 3 display these trends for SBP,DBP, and HR, respectively.24Figure 1. SBP response over time.130 —125120 -/ /- AloneEvaluative110 9J• Non-evaluativeC,)105 —100 195 I IBL1 BL2 MT1 MT2 MT325• AloneEvaluative• Non-evaluativeFigure 2. DBP response over time.807876742 7227o68666462BL1 BL2 MT1 MT2 MT3‘1HR(bpm)-.)C*)(110)-.10)CDCoo000000000w w -1IH C.)27From Figure 1, it appears that the non-evaluative condition demonstrated a decrease inSBP from minute 1 to minute 4.5 of the task period, whereas subjects in the other conditionsdid not show such habituation. Figure 2 suggests a similar trend for DBP, although theconditions are clustered much closer together at minute 4.5 than they are for SBP. Visualinspection of Figure 3 did not suggest any differences at all between the conditions on heartrate change over time.To more fully examine potential differences between the conditions at each time point,ANCOVAs were conducted at each task reading, covarying out the second baseline reading.For SBP, there were no significant differences between conditions at the first or second taskreading, but a significant difference emerged at the third task reading (E (2, 44) = 3.74, =.032). Post hoc simple effects tests indicated that the non-evaluative condition had lower SBPreadings at the third task reading (mm. 4.5) than the alone condition. The evaluativecondition’s scores fell in between and were not significantly different than the other twoconditions.When ANCOVAs were conducted for DBP at each task reading, there were nosignificant differences between the groups at any time point. Not surprisingly, no differencesemerged with respect to HR at any of the task readings.Taken together, these analyses indicate that the non-evaluative condition tended todisplay decreased SBP reactivity to the math task, compared to subjects performing the mathtask alone. This difference only emerged at minute 4.5 of the 5 minute task and was notpresent for either DBP or HR. The evaluative condition could not be distinguished from theother conditions on any cardiovascular measure.For the groups that were paired with friends, the visual analog scale data indicate thatthe evaluative subjects tended to feel more evaluated during the task than the non-evaluativesubjects (although this difference did not reach the .05 significance level), while there were nodifferences in perceived support. The evaluative group also felt more competitive towardstheir friends than the non-evaluative group.28Finally, the correlational data indicate that closeness to the friend showed a moderatenegative correlation with SBP reactivity.Friend’s DataDemographics and Psychological QuestionnairesThe two friendship groups (evaluative and non-evaluative) were not found to besignificantly different on any demographic variables. The two conditions also did not differ interms of each friend’s perceived closeness to the subject.Interestingly, univariate ANOVAs conducted on the post-task visual analog scalesrevealed that the “evaluative” friends perceived themselves as significantly more evaluative, incomparison to the “non-evaluative” friends during the math task (E (1, 27) = 6.30, p .0 18),but there were no differences between the 2 conditions in terms of perception of providedsupport ( (1, 27) = .707, p = .41). Additionally, the “evaluative” friends felt significantlymore stressed (F (1, 27) = 19.33, p .001) and more competitive towards the respectivesubjects (E (1, 27) = 10.51, p .003) during the math task than the “non-evaluative” friends.(See Table 7.)Friends’ Cardiovascular MeasuresBaseline and task means of the friends’ SBP, DBP, and HR are displayed in Table 8.As with the subjects, each table provides 2 baseline and 3 task readings for the two conditions.Cardiovascular change scores for each measure are displayed in Table 9.Friends’ baseline cardiovascular measures. ANOVAs were conducted separately foreach baseline reading, as well as for the mean of the baseline period. While the first baselinemeasure and the mean of the two baselines showed no differences between the groups, thegroups were marginally different at the second baseline reading (p = .054). As a result, themean of the two baseline readings was used for all further analyses.Friends’ cardiovascular measures associated with task. In order to compare thecardiovascular response of the two groups of friends during the task, repeated measures29Table 7. Post-Task Visual Analog Scales by Friendship Condition, Friends’ RatingsScale Evaluative group Non-evaluative groupMean (SD) Mean (SD)Evaluation 5.8(2.1) 3.5(2.6)Support 4.9 (1.5) 5.6 (2.8)Competition 5.5 (2.1) 2.6 (2.7)Degree of stress 5.2(2.4) 2.0(1.5)Note: All visual analogue scales are 10 point scales, with 0 being the lowest and 10 being thehighest endpoint.30Table 8. Friends’ Cardiovascular Readings During Baseline and Task by ConditionSBP(mmHg)BL1Mean(SD)114.6(10.4)BL2Mean(SD)115.9(9.9)MT1Mean(SD)115.1(7.8)MT2Mean(SD)114.7(7.9)MT3Mean(SD)112.5(8.7)EvaluativeNon- 115.3 109.1 108.9 106.5 104.1evaluative (7.6) (8.2) (9.8) (8.0) (6.9)DBP(mmHg)70.5 70.9 72.8 70.9 70.2Evaluative (12.9) (9.7) (10.6) (10.1) (9.8)Non- 68.2 68.3 66.9 62.9 62.2evaluative (9.0) (9.4) (6.7) (7.1) (7.4)HR(bpm)69.1 74 70.2 70.8 71.4Evaluative (7.4) (9.5) (9.4) (8.9) (7.5)Non- 69.6 68.6 67.3 66.2 67.3evaluative (9.9) (11.1) (9.6) (9.9) (10.7)Table 9. Friends’ Change Scores for all Cardiovascular Measures by ConditionGroup SBP change score DBP change score HR change scoreMean (SD) Mean (SD) Mean (SD)Evaluative friends -1.1 (4.9) .63 (4.9) -.75 (3.8)Non-evaluative (4.6) -4.3 (3.4) -2.1 (3.5)friends3132ANCOVAs were conducted for all cardiovascular measures, with the mean of the 2 baselinesserving as the covariate. For SBP, a strong main effect for group ( (1, 26) = 9.78, p = .004)and for time ( (2, 54) 7.36, p = .001) was found, and there was no interaction. Similarlyfor DBP, strong differences emerged between the conditions (E (1, 26) = 16.69, p .00 1)and for the time factor ( (2, 54) = 9.85, p .00 1). There was no interaction. Finally, nodifferences were found between the groups for HR (. (1, 26) = 1.46, p = .24). Additionally,there were no differences between HR task readings nor was there an interaction betweengroup and time. These results support the predictions of this study, in finding that the“evaluative” friends show less of a decrement in SBP and DBP than the “non-evaluative”friends.Finally, ANCOVAs for each task reading were conducted to more definitively look atdifferences between the conditions at each of these time periods. As with all previousanalyses, the mean of the 2 baseline readings served as the covariate. For SBP, ANCOVAsdemonstrated significant differences at the second task reading (minute 2.5), F (1, 26) = 8.75,p .007 and at the third task reading (minute 4.5), (1, 26) = 12.92, p = .001. Nodifferences between the groups were found at the initial task reading. For DBP, ANCOVAsdemonstrated significant differences at all task readings. (Specifically, for task 1, (1, 26) =4.90, p= .036; fortask2,E(1, 26)= 11.86, p= .002; fortask3,E(1, 26)= 19.63, p.00 1) Not surprisingly, for fiR, ANCOVAs revealed no differences between the groups atevery task reading.The preceding analyses indicate that the “non-evaluative” friends displayed SBP andDBP habituation during the math task, in contrast to the “evaluative” friends who remained attheir pre-task baseline levels on these parameters. These differences were evident at all timepoints of the task for DBP. However, for SBP, these differences did not emerge until thesecond task reading (minute 2.5). No differences between the groups were found for heartrate.33Additionally, manipulation checks revealed that the evaluative group perceivedthemselves to be more evaluative, more competitive, and more stressed than the nonevaluative group. However, the two groups did not differ in perception of provided support.34DiscussionThe present study attempted to clarify why the presence of a friend may sometimesserve as social support and reduce arousal, but also may be perceived as evaluative andincrease arousal in other circumstances. Specifically, we tested the hypothesis that a non-evaluative condition would reduce cardiovascular reactivity relative to subjects performing astressful math task alone. An evaluative condition was predicted to augment reactivity to thetask. In support of these predictions, the creation of a non-evaluative scenario was found topartially attenuate the systolic blood pressure response to a math task. On the whole, findingswere weak, and significant effects did not emerge until the last reading of the task.The evaluative condition did not lead to greater increases in blood pressure or heartrate than that occurring in the other conditions. In fact, the SBP readings of the evaluativecondition fell in between the alone and non-evaluative conditions and could not be readilydistinguished from them. These findings suggest that reductions in SBP reactivity attributed tothe presence of a supportive friend are eliminated when even a subtle evaluative component isadded.The friends’ cardiovascular data provide an appealing addition to the literatureexamining social support and cardiovascular reactivity. During the math task, the two groupsof friends differed on SBP and DBP measures, in that cardiovascular habituation was apparentonly in the “non-evaluative” friends. No differences between the groups were found for heartrate. The SBP differences occurred at the second and third task reading, while groupdifferences in DBP occurred at all task readings. Specifically, the “non-evaluative” friendsshowed increasing habituation in SBP and DBP over the duration of the task, while the“evaluative” friends remained at about the baseline level.The results of the present study are in agreement with the findings of Kamarck et al. ‘s(1990) and Eden et al.’s (1992) studies, which each found reductions in cardiovascularresponse to a stressful math task in the presence of a non-evaluative friend. Because thepresent study did not employ the use of touch in the non-evaluative condition, our study most35closely parallels the Eden et al. (1992) study, which provided a no-touch/non-evaluativesupport condition among their various groupings. The fact that differences between our aloneand non-evaluative conditions were not found for heart rate is also consistent with the Eden etal.’s findings. It is possible that Kamarck Ct al. (1990) found differences on this measure dueto the use of touch in their non-evaluative condition. That is, it has been suggested that touchmay reduce resting heart rate (Drescher, Gantt, & Whitehead, 1980).The present results differ from those of Allen Ct al. (1991) which found heightenedreactivity in subjects assigned to an evaluative condition, when compared to an alone or non-evaluative condition. Several differences between the studies, however, are worth noting.The Allen et al. (1991) study tested subjects in their home and used an older sample ofcommunity women (approximate age was 39 years) who were predominantly Caucasian. Incontrast, our study completed testing in a laboratory and used a college-aged sample; agreater proportion of subjects were Asian. It is possible that the meaning of support andevaluation may vary across different ages and cultures, and that it could vary as a function ofstudy location. Thus, it is not clear if these potential differences could have influenced theresults of the present study.Additionally, the present study may not have had as strong a manipulation ofevaluation potential as the Allen et al. (1991) design. In their study, the friends werepositioned well within the subjects’ view during the task, and the friends were described asusually leaning forward towards the subject. In the present study, the friend was not in thesubject’s direct field of vision. As a result, the subjects may have been less aware of what theirfriends were doing, causing the subjects to feel less evaluated. This weaker perception ofevaluation could have translated into reduced reactivity on the part of this group.Given that the observed differences between subject conditions were small, in thepresent study, there is always the possibility that the manipulation did not work. To counterthis, the visual analog scales provide some evidence that the manipulation of evaluationpotential was indeed effective. Although the subjects in both friend conditions did not differ36in their perception of support, the evaluative group perceived their friends as being somewhatmore evaluative than the non-evaluative subjects. This group also felt more competitivetowards their friends during the math task. The friends’ visual analog data clearly paralleledthat of the subjects, with “evaluative” friends perceiving themselves as more evaluative of thesubjects, as well as more competitive with them, when compared to the “non-evaluative”friends. The two groups of friends did not differ in their perception of how supportive theyhad been during the task, which also is consistent with the self-report data of the subjects.Thus, it seems plausible that the two groups of friends were indeed sending out different cuesto the subjects in their respective groups. Additionally, the differential cardiovascularhabituation of the friends in the two support conditions provides evidence for the effectivenessof the evaluation manipulation.It is not surprising that perception of support was not different for the two friendconditions, since both conditions were set up as “support” conditions. Subjects in theseconditions were told that their friend’s primary role was to support them during the task, andboth sets of friends were also provided with (the same) detailed instructions describing howthey could be supportive.It is interesting to note that the “evaluative” friends felt more stressed than the “non-evaluative” friends. One can speculate that they might have felt a subtle pressure to engage inthe task themselves. It is also feasible that feeling evaluative (because of the nature of themanipulation) when one is expected to be supportive may cause a certain amount of cognitivedissonance, thereby leading to feelings of stress.In terms of major theories about the mechanism of social support, the present studyprovided some evidence for a buffering effect of social support on cardiovascular reactivity.Since the presence of a friend did not influence resting blood pressure and heart rate values,there was no main effect of social support. Rather, differences only emerged during the mathtask, and at a specific time point of the task.37Although the stress-buffering model speculates that support may play a role in alteringone’s appraisal of the stressor or by intervening after the event has been appraised as stressful(Cohen & McKay, 1984; House, 1981), the present study does not clearly provide an answeras to where in the model the support actually has its effects. Because the non-evaluative andalone conditions did not differ in their perception of stress during the math task, it seemsunlikely that support is affecting one’s conscious appraisal of the stress. In terms ofLazarus’scoping model (Lazarus & Folkman, 1984), there were no differences in the primary appraisalof the threat. However, it is also conceivable that stress appraisal is not a conscious processand that people are not aware of changes in feelings of stress. Support could work byincreasing feelings of esteem and control, as well as one’s ability to cope with the stressor.Future studies need to look more closely at these variables, in an effort to fully understand themechanism (pathways) of social support as a buffer for cardiovascular reactivity.The present study sheds light on one aspect of support that may be important for itsbuffering properties. Perceived closeness to the friend was associated with decreased SBPreactivity. These results tie in nicely with the buffering effects of support that were foundspecifically for SBP. Perhaps one needs to feel fairly close to someone for social support tohave a protective role on cardiovascular reactivity. Future studies must be conducted todetermine if these results can be generalized to other behavioral stressors.In conclusion, the present study replicated two earlier studies in finding that thepresence of a non-evaluative friend tended to buffer SBP reactivity to a stressful laboratorytask. The current study was unique in using a different type of math task (written response toequations instead of oral serial subtraction) and in minimizing evaluation potential in a mannernot previously done. In addition, the visual analog scales provided a useful manipulationcheck and were able to detect differences among the groups on several state measures, thusconfirming the validity of the manipulation. The monitoring of the friends’ cardiovascularresponse was also a worthwhile addition to the study, in that it provides a further validation38check and describes the “cost” of providing support. In this case, being a support person wasnot associated with physiological arousal.Furthermore, our findings for the evaluative condition suggest that uncontrolledevaluation apprehension may eliminate the buffering effects of support. That is, a supportivesituation that has elements of evaluation may not adequately reduce cardiovascular reactivityto a stressful task. Thus, evaluation potential should be minimized in any support condition.Finally, our finding that non-evaluative subjects showed reduced SBP reactivity only atthe end of the task is an important one. Studies that use only means of the task period may bewashing out interesting time trends in the support process. Our data seem to indicate that theimpact of social support on reactivity to a stressful math task may not be an immediate one.Rather, it may take time for support to decrease cardiovascular response, possibly requiringtasks longer than five minutes. Whether this finding is specific to our task or true of otherbehavioral stressors is something that awaits future studies.6SDIGNMJV40Appendix 1DATA SHEET FOR SUBJECTDate of visit__________TimeSubject # InitialsGender F( ) M( )Condition: Non-evaluative____Evaluative_____ADAPTATION SBP DBP HRMinute 0_______ ____________Minute 4_Minute 5 - experimenter enters expt. room with additional instructions to subject (re: math task)TURN ON TV AND VCR (BUT DON’T PUT IN TAPE) - SHOULD SEE BLUE SCREENIn the friend conditions, the friend enters the room and is hooked to a bp cuffMinute 9Minute 13.5_ _____ ___Minute 14.5 - (over intercom) -30 second warning about math taskTASK SBP DBP HRMinute 15 - (over intercom) - tell subject to put down the questionnaires or comic book and that the taskwill now beginPUT ON MATH TASKMinute 16Minute 17.5_ ___Minute 19.5Minute 20 - (over intercom) - experimenter gives instructions about returning to questionnairesRECOVERY SBP DBP HRMinute 20Minute 2441DATA SHEET FOR FRIENDDate of visit___________Subject #_ _Gender F( ) M( )Condition: Non-evaluative_____Time________Initials__ __ __Evaluative__ADAPTATION SBP DBP HRMinute 5: FRIEND ENTERS ROOM AND IS HOOKED liP TO BP CUFFMinute 9Minute 13.5TASK SBP DBP HRMinute 16Minute 17.5Minute 19.5RECOVERY SBP DUP HRMinute 20_______Minute 2542Appendix 2: Friendship Questionnaire and Visual Analog Scales (for subject andfriend)Friendship QuestionnairePlease answer this questionnaire with regard to the friend that you brought with you to the labtoday.1. How long have you known this person?2. How long have you considered this person a close friend?For questions 3 to 5, place an X anywhere along the line to indicate how you feel or think.3. How often do you have contact (get together with or talk with on the telephone) with thisperson?Not very veryoften often4. How close do you feel towards your friend?Not close veryat all close5. How comfortable do you feel around your friend?Not comfortable veryat all comfortable43Visual Analog Scale for Subjects Perfonning the Math Task AloneFor each question below, place an X anywhere along the line to indicate how you felt duringthe math task?1. How stressed did you feel during the task?Not at all verystressed stressed2. How would you rate your performance during the task?poor excellent3. How confident did you feel in your ability to do the task?not at all veryconfident confident44Visual Analog Scale for Subjects performing the Math Task in the Presence of theirFriendFor each question below, place an X anywhere along the line to indicate how you felt duringthe math task?1. How stressed did you feel during the task?Not at all verystressed stressed2. How supported did you feel during the task?Not at all verysupported supported3. How comfortable did you feel with your friend present during the task?Not at all verycomfortable comfortable4. How evaluated or judged did you feel by your friend during the task?Not at all veryevaluated evaluated5. How competitive did you feel with your friend during the task?Not at all verycompetitive competitive45Visual Analog Scale for Subjects Performing the Math Task in the Presence of theirFriend (cont’d)6. How would you rate your performance during the task?poor excellent7. How confident did you feel in your ability to do the task?not at all veryconfident confident46Visual Analog Scale for FriendsFor each question below, place an X anywhere along the line to indicate how you felt duringthe math task.1. How stressed did you feel while your friend was doing the math task?Not at all verystressed stressed2. How supportive of your friend did you feel during the task?Not at all verysupportive supportive3. To what extent did you feel that you were evaluating your friend while she was doing thetask?Not at all veryevaluating evaluating4. How competitive did you feel with your friend during the task?Not at all verycompetitive competitive47Appendix 3Instructions to FriendEvaluative friend:You will be sitting in the room with your friend, while your ffiend does a stressful 5minute math task. Like your friend, you will be hooked up to a recording device that monitorsyour bp and heart rate. First you will relax and fill out questionnaires and then you will bethere while your friend does the 5 minute math task... Sign consent formBefore the task, both you and your friend will be filling out questionnaires. Here is thestack of questionnaires that we’d like for you to fill out. For now, we’d like you to start withthese questionnaires and stop when you get to the blue page which tells you to stop. Thereare instructions on top of each questionnaire explaining how to fill them out. Some are self-explanatory, but please read them carefully. Don’t take too much time to think about eachquestion, but say whatever immediately comes to mind, as you read them.... If, for somereason, you finish the questionnaires, please feel free to read the Herman Comic book.During the task, you will be able to see the questions on t.v., and the answers thatyour friend writes will be on a white board, within your view. Your job is to give your friendnon-verbal support during the math task. In other words, we would like you to refrain fromtalking with or touching your friend in a supportive way, while in the room. But you cansupport your friend in other ways that you feel comfortable with. For example, you can giveyour friend a warm smile from time to time, you can sit with a more open position, withoutyour arms crossed at your chest (DEMONSTRATE), or just being there for them and doingwhatever you think constitutes support.Does this make sense?We would like you, however, to not stare too often at your friend. Certainly feel freeto glance at her but don’t stare at her. Also don’t nod when she writes down her answer. Inother words, don’t do anything which might let her know if her answer is correct or not. Also48please try to refrain from fidgeting in your chair, yawning in an obvious manner or doinganything that might distract your friend.We would like you to attend to the questions and answers that your friend provides, aswell as attend to your friend. As soon as the math task is finished, I will come into the roomand unhook you and your part in the study will be finished.Oh also - please don’t talk or interact with your friend in any way during the periodprior to the task or during the task itself. While you’re waiting for the task to begin, just focuson the questionnaires and if you finish early, please read the comic book.Non-evaluative friend:You will be sitting in the room with your ffiend, while your friend does a stressful 5minute math task. Like your friend, you will be hooked up to a recording device that monitorsyour bp and heart rate. First you will relax and fill out questionnaires and then you will bethere while your friend does the 5 minute math task.... sign consent form.Before the task, both you and your friend will be filling out questionnaires. Here is thestack of questionnaires that we’d like for you to fill out. For now, we’d like you to start withthese questionnaires and stop when you get to the blue page which tells you to stop. There areinstructions on top of each questionnaire explaining how to fill them out. Some are self-explanatory, but please read them carefully. Don’t take too much time to think about eachquestion, but say whatever immediately comes to mind, as you read them.... If, for somereason, you finish the questionnaires, please feel free to read the Comic book.During the task, you will not be able to see the questions that your friend is given oryour friend’s answers (and please do not make any attempt to do so). Your job, however, isto give your friend non-verbal support during the math task. In other words, we would likeyou to refrain from talking with or touching your friend in a supportive way, while in theroom. But you can support your friend in other ways that you feel comfortable with. Forexample, you can give your friend a warm smile from time to time, you can sit with a more49open position, without your arms crossed at your chest (DEMONSTRATE), or just beingthere for them and doing whatever you think constitutes support.Does this make sense?We would like you, however, to not stare too often at your friend. Certainly feel freeto glance at her but don’t stare at her. Also please try to refrain from fidgeting in your chair,yawning in an obvious manner or doing anything that might distract your friend.Also, we will have some picture books for you to look through when your friend isdoing the math task. We would like for you to look through them, because it will help yourfriend feel more comfortable if she knows you have something to do during the task also.As soon as the math task is finished, I will come into the room and unhook you andyour part in the study will 1 be finished.Oh also - please don’t talk or interact with your friend in any way during the periodprior to the task or during the task itself. While you’re waiting for the task to begin, just focuson the questionnaires and if you finish early, please read the comic book.50ReferencesMien, K., Blascovich, J., Tomaka, J., & Kelsey, R. (1991). Presence of human friends and petdogs as moderators of autonomic responses to stress in women. Journal ofPersonality and Social Psychology, 582-589.Berkman, L., Leo-Summers, L., & Horwitz, R. (1992). Emotional support and survival aftermyocardial infarction. Annals of Jnternal Medicine, 117, 1003-1009.Berkman, L. & Syme, S.L. (1979). Social networks, host resistance and mortality: a nine yearfollow-up study of Alameda County residents. American Journal Epidemiology, i,186-204.Blazer, D. (1982). Social support and mortality in an elderly community population.American Journal of Epidemiology, 11, 684-694.Case, R., Moss, A., Case, N., McDermott, M., & Eberly, S. (1992). Living alone aftermyocardial infarction. JAMA, 7, 515-520.Cassel, J. (1976). The contribution of the social environment to host resistance. AmericanJournal Epidemiology, 104, 107-123.Cohen, S. (1988). Psychosocial models of the role of social support in the etiology ofphysical disease. Health Psychology, 7, 269-297.Cohen, S. & McKay, G. (1984). Social support, stress, and the buffering hypothesis: Atheoretical analysis. In A. Baum, S.E. Taylor, & J.E. Singer (Eds.), Handbook ofPsychology and Health (pp. 253-267). Hillsdale, N.J.: Lawrence Eribaum Associates,Inc.Cohen, S., Mermelstein, R., Kamarck, T., & Hoberman, H.N. (1985). Measuring thefunctional components of social support. In I. Sarason & B. Sarason (Eds.), SocialSupport: Theory. Research. and Applications (pp. 73 -94). Dordrecht, NEtherlands:Martinus Nijhoff.Cohen, S. & Wills, T.A. (1985). Stress, social support, and the buffering hypothesis.Psychological Bulletin, , 310-357.Drescher, V.M., Gant, W.H., & Whitehead,W.E. (1980). Heart rate response to touch.Psychosomatic Medicine, 42, 559-565.Edens, J., Larkin, K., & Abel, J. (1992). The effect of social support and physical touch oncardiovascular reactions to mental stress. Journal of Psychosomatic Research, ,371-382.51Haynes, S. & Feinleib, M. (1980). Women, work, and coronary heart disease: Pprospectivefindings from the Framingham Heart Study. American Journal of Public Health, 7Q,133- 14 1.House, J. (1981). Work. Stress, and Social Support. Reading, MA: Addison-Wesley.House, J., Robbins, G., & Metzner, H. (1982). The association os social relationships andactivities with mortality: Prospective evidence from the Tecumseh Community HealthStudy. American Journal Epidemiology, il, 123-140.Joseph, J. (1980). Social affihiiation. risk factor status, and coronary heart disease: A cross-sectional study of Japanese-American men. Unpublished Ph.D. Thesis, University ofCalifornia, Berkeley.Kamarck, T., Manuck, S., & Jennings, R. (1980). Social support reduces cardiovascularreactivity to psychological challenges. Psychosomatic Medicine, , 42-58.Kaplan, B.H., Cassel, J.C., & Gore, 5. (1977). Social support and health. Medical Care, j.,47-58,Kaplan, G., Salonen, J., Cohen, R., Brand, R., Syme, L., & Puska, P. (1988). Socialconnections and mortality from all causes and from cardiovascular disease:Prospective evidence from Eastern Finland. American Journal of Epidemiology. i2,370-380.Lazarus, R.S., & Folkman, S. (1984). Stress. appraisal. and coping. New York:Springer.Leary, M.R. (1983). A brief version of the Fear of Negative Evaluation Scale. Personality andSocial Psychology Bulletin. , 37 1-375.Linden, W. (1991). What do arithmetic stress tests measure? Protocol variations andcardiovascular responses. Psychophysiology, 2, 9 1-102.Linden, W., Chambers., L., Maurice, J., & Lenz, J.(1993). Sex differences in social support,self-deception, hostility, and ambulatory cardiovascular activity. Health Psychology,j., 376-3 80.Manuck, S., & Krantz, D. (1986). Psychophysiological reactivity in coronary heart diseaseand essential hypertension. In K.A. Mathews, S.M. Weiss, T.Detre, T.M. Dembroski,B. Falkner, S. Manuck, & RB. Williams. (Eds.) Handbook of Stress. Reactivity andCardiovascular Disease (pp. 11-34). New York: Wiley.Medalie, J. & Goldbourt, V. (1976). Angina pectoris among 10,000 men: Psychosocial andother risk factors as evidenced by a multivariate analysis of a five year incidence study.American Journal of Medicine, Q, 910-921.52Orth-Gomer, K. & Johnson, J. (1987). Social network interaction and mortality: a six yearfollow-up study of a random sample of the Swedish population. J. Chron. Dis, 4,949-957.Reed, D., McGee, D., Yano, K., & Feinleib, M. (1983). Social networks and coronary heartdisease among Japanese men in Hawaii. American Journal ofEpidemiology, 111, 384-396.Rosenberg, M. (1965). Society and the Adolescent Self-Image. Princeton, NJ: princetonUniversity Press.Rotter, J.B. (1966). Generalized expectancies for internal versus external control ofreinforcement. Psychological Monographs, 80, 1-28.Schoenback, V., Kaplan, B., Fredman, L., & Kleinbaum, D. (1986). Social ties and mortalityin Evans County Georgia. American Journal of Epidemiology, m, 577-591.Seeman, T. & Syme, L. (1987). Social networks and coronary artery disease: A comparisonof the structure and ftinction of social relations as predictors of disease.Psychosomatic Medicine, 4, 341-354.Sheffield, D. & Carroll, D. (1993). Social support and cardiovascular reactions to activelaboratory stressors. Psychology and Health, In press.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items