predictors of friend approval for romantic -

Personal Relationships, (2012). Printed in the United States of America.Copyright © 2012 IARR; DOI: 10.1111/j.1475-6811.2012.01397.x

Predictors of friend approval for romanticrelationships

PAUL E. ETCHEVERRY,a BENJAMIN LE,b AND NICHOLAS G. HOFFMANa

aSouthern Illinois University Carbondale and bHaverford College

AbstractThree studies (1 correlational and 2 experiments) applied the investment model to explain social network approvalfor others’ romantic relationships. Study 1 supported the prediction that friends’ perceived satisfaction and perceivedalternatives were predictive of their approval for target relationships, while perceived investments was not asignificant predictor. Studies 2 and 3 employed experimental manipulations of perceived satisfaction to test its causalimpact on relationship approval, and in these studies, perceived satisfaction was a significant predictor of relationshipapproval. Taken as a whole, these findings supported the prediction that perceived relationship satisfaction is a causalfactor influencing friends’ approval of target romantic relationships.

The daily activities of romantic couples takeplace within larger social networks of friends,family, and acquaintances, and these socialnetworks influence those embedded roman-tic relationships (Etcheverry & Agnew, 2004;Leslie, Huston, & Johnson, 1986). Muchof the extant research on romantic relation-ships has taken an individualistic or dyadicapproach to understand romantic relationshipfunctioning. However, scholars have begunto recognize that to fully understand roman-tic relationship functioning, it is necessary toconsider social network influence (Berscheid,1999; Sprecher, Felmlee, Orbuch, & Wil-letts, 2001).

Although many aspects of social net-works have been found to be associatedwith relationship processes, including con-tact with kin (Burger & Milardo, 1995) and

Paul E. Etcheverry, Department of Psychology, SouthernIllinois University Carbondale; Benjamin Le, Departmentof Psychology, Haverford College; Nicholas G. Hoffman,Department of Psychology, Southern Illinois UniversityCarbondale.

Correspondence should be addressed to Paul E.Etcheverry, Department of Psychology, Southern Illi-nois University Carbondale, Life Sciences II Room281—Mail Code 6502, Carbondale, IL 62901, e-mail:[email protected].

network density (Kim & Stiff, 1991), by farthe most commonly studied variable is thedegree to which social network membersapprove or disapprove of the romantic rela-tionship (Sprecher et al., 2001). Social net-work approval is predictive of relationshipsatisfaction (Sprecher & Felmlee, 1992), com-mitment (Etcheverry, Le, & Charania, 2008),and the quality and stability of romantic rela-tionships (Le, Dove, Agnew, Korn, & Mutso,2010; Sprecher et al., 2001).

Despite the importance of social networkopinions, little research has examined the fac-tors that lead social network members toapprove or disapprove of a romantic rela-tionship. One study found that a change inrelationship status (i.e., becoming engaged)may prompt increases in support (Sprecher &Felmlee, 2000). Furthermore, Felmlee (2001)collected open-ended responses for why indi-viduals approve or disapprove of their bestfriends’ relationships and found that approvaltypically stemmed from the partners’ posi-tive personal characteristics and an equitableexchange of rewards. Disapproval was a func-tion of concerns about the friends’ well-beingin addition to the effect of the relationshipon the friendship. However, beyond these two

1

2 P. E. Etcheverry, B. Le, and N. G. Hoffman

studies, there is no direct research on the gen-esis of network approval or disapproval.

The lack of research examining the causesof social network approval or disapproval maybe due, in part, to the lack of a clear theo-retical basis for understanding the process ofnetwork approval. To this point, no theory hasbeen developed to explain how social networkmembers develop opinions regarding particu-lar romantic relationships. However, perspec-tives such as interdependence theory (Kelley& Thibaut, 1978) explain how people makedecisions about maintaining or ending theirown relationship. We suggest that interde-pendence theory can be extended to provideinsight into the process of network approval.

Interdependence theory

Although many perspectives have beenoffered for understanding interpersonal rela-tionships (Clark & Lemay, 2010), interde-pendence theory (Kelley & Thibaut, 1978)has been particularly effective in explain-ing how people evaluate the quality of theirinterpersonal relationships. Interdependencetheory assumes that people are motivated tomaximize rewards and minimize costs withintheir interpersonal relationships, includingromantic relationships. The theory posits thatpeople consider the level of outcomes in theirrelationships, defined as rewards minus costs,when evaluating those relationships. Theseoutcomes are compared with the expectedlevel of outcomes for the relationship, knownas the comparison level for outcomes, to forma judgment of satisfaction with the relation-ship (Rusbult, Arriaga, & Agnew, 2001). Ifrelationship outcomes exceed the expectedlevel of outcomes for a relationship, then theperson is satisfied with the relationship; ifoutcomes fall short of the comparison level,dissatisfaction will occur.

Members of relationships also considerthe level of rewards and costs available intheir best possible alternative relationships.The comparison level for alternatives repre-sents the expected level of rewards and coststhat could be received in an alternative rela-tionship, which is important in understandingthe extent to which one relies on a chosen

relationship over an alternative for the fulfill-ment of relationship needs.

Rusbult’s investment model

Designed as an addition to interdependencetheory, the investment model adds two con-structs: relationship investments and relation-ship commitment to understand relationships.Investments are those resources put into therelationship that would be lost if the relation-ship ends, including the loss of rewards thatcome from those investments and the costsassociated with replacing those investmentswith new relationships (Rusbult, 1980). In theinvestment model, commitment to a romanticrelationship refers to a psychological attach-ment to the relationship and an intention tomaintain the relationship (Arriaga & Agnew,2001). Commitment is argued to be a functionof higher levels of satisfaction, lower qualityof alternatives, and higher investments (Le &Agnew, 2003), and has been found to predictcognition and behavior within relationships(Rusbult & Agnew, 2010).

Interdependence theory, the investmentmodel, and social network approval

The key interdependence theory variablesdescribed above, satisfaction and alternatives,as well as the investment model variables,investment and commitment, point to the pos-sible rewards and costs of remaining in orleaving a relationship. It is possible that socialnetwork members are concerned with therewards and costs of a target person remainingin or leaving a relationship. Therefore, net-work members use their perceptions of thekey interdependence theory and investmentmodel variables to inform their approval ordisapproval for the target person’s relation-ship. To this point, no research has explic-itly tested whether a social network member’sperception of a person’s satisfaction, alterna-tives, and investments predicts approval ordisapproval of the relationship.

Determinants of friend relationship approval

A basic assumption of social exchange the-ories in relationships is that people are

Predictors of friend approval 3

motivated to maximize the rewards and mini-mize costs received from relationships. How-ever, it is not explicitly clear that socialnetwork members should be concerned withthe rewards and costs of a target person’sromantic relationship as the social networkmember does not directly gain these rewardsand costs for themselves. The characteristicsof social networks can provide an insight hereas most social network members are friendsor family members of at least one memberof a romantic couple (Milardo, 1982, 1988).These social network members presumablycare for the target person and desire the high-est rewards and lowest costs for their friendor family member. Friendships that are com-monly based on mutual concern and care(Van Lear, Koerner, & Allen, 2006) shouldbe very likely to lead to friends desiringhigh rewards and low costs for a target per-son’s relationship.

The little past research on friend approvalfor romantic relationships (Felmlee, 2001)indicates that relationship approval is basedon partner characteristics and fair rewardswithin the relationship. Friend disapprovalfor a romantic relationship was partly basedon the friends’ well-being and the negativeaspects of the relationship. This prior studysupports the assumption that friends will beconcerned with the rewards and costs of a tar-get romantic relationship. Therefore, the cur-rent research focuses on friends as a sourceof relationship approval for a target romanticrelationship. If, as expected, friends use rela-tionship outcomes (rewards and costs) whenevaluating a target romantic relationship, thissupports the relevance of interdependence the-ory and investment model variables.

For a young adult sample involved pri-marily in dating relationships, friend approvalis likely to be of particular importance.Given the amount of time friends spendtogether at this age, they are potentially animportant source of influence on relation-ships. Prior research on a college-aged sam-ple supports this as friend approval was asignificant predictor of a person’s own rela-tionship commitment, even when controllingfor parent and sibling approval (Etcheverry &Agnew, 2004).

Satisfaction is partly based on the relation-ship outcomes (rewards minus costs) that arereceived from a relationship. For social net-work members, perceiving higher outcomesin a target relationship should lead to higherperceived satisfaction for that relationship andgreater approval for that relationship. Per-ceived alternatives are relevant to approvalor disapproval as they indicate the extent ofrelationship outcomes available in an alterna-tive relationship. If few alternatives are per-ceived, network members may approve morebecause there are no other viable relation-ships available. If many alternatives are per-ceived, approval for a current relationshipmay be lower because alternatives are seen asappealing. From the investment model, per-ceived investments in a relationship are rele-vant to relationship approval because the lossof investments is associated with lost rewardsand increased costs. Thus, network membersperceiving high investments may support thatrelationship to avoid the loss of those invest-ments.

It is less clear how the investment modelconstruct of commitment is related to friendrelationship approval. In the investment model,commitment is a product of satisfaction, alter-natives, and investments. However, unlikethese variables, commitment does not directlyprovide information about the level of rewardsand costs in a target relationship. If friendsare primarily concerned with relationship out-comes in a target relationship, then perceivedcommitment may not provide useful informa-tion over and above the other three variables.This suggests that perceived commitment maynot add to the prediction of relationshipapproval.

An alternative possibility is that friendswill consider the degree of commitment theyperceive a target person feels for his or herrelationship as a measure of the target’s desireto maintain the relationship. Little researchhas examined this issue, however, Sprecherand Felmlee (2000) report that when couplesbecame engaged, they perceived more supportfrom friends. This result may indicate thatthe increase in commitment associated withbecoming engaged is used by friends to adjustrelationship approval. Friends may choose to


take into account the target’s preferences, inwhich case if the friend perceives high targetcommitment, romantic relationship approvalshould increase, while low target commitmentshould lead to low relationship approval.

Developing network perceptions

Applying interdependence theory and theinvestment model to social network approvalassumes that network members can effec-tively develop perceptions of the charac-teristics of the target romantic relationship.Previous research including both experimentaland survey research has supported the abilityof social network members, including friends,to develop perceptions of a target roman-tic relationship. In an experimental study,Rusbult (1980) created hypothetical scenariosdepicting a romantic relationship character-ized by either high or low costs, either highor low quality of relationship alternatives, anddiffering levels of investments in the currentrelationship. Participants rated hypotheticalrelationships with lower costs as higher in sat-isfaction, while lower levels of alternativesand higher investments in the scenarios ledto perceptions of greater commitment. Theseresults suggest that social network membersare able to observe and develop evaluationsof romantic relationships.

In several cross-sectional studies, socialnetwork members have been found to developperceptions of the characteristics of targetrelationships (Agnew, Loving, & Drigotas,2001; Loving, 2006; MacDonald & Ross,1999). In all three of these studies, friendsof people involved in romantic relation-ships were able to develop perceptions ofcommitment, satisfaction, and investments inthe target relationship. The perceptions werefound to correlate with the measures ofthese variables collected from the romanticrelationship participants. In addition, theseperceptions were then found to be predictiveof relationship persistence (Loving, 2006).This information may come directly frommembers of the relationship, observing thecouple interact with each other, or discussingthe relationship with other social networkmembers (Agnew et al., 2001). Although

social network members may not be entirelyaccurate, they can develop perceptions ofcommitment to, satisfaction with, alternativesto, and investments in the romantic rela-tionship. Unlike the current research, noneof these previous studies tested how theseperceptions are related to friend’s approvalor disapproval for continuing a romanticrelationship.

Applying an interdependence theory ap-proach leads to the first two hypotheses:

H1: Perceiving a friend is satisfied withhis or her relationship will be pos-itively associated with a social net-work member’s approval of thatromantic relationship.

H2: Perceiving a friend has high qual-ity of alternatives to a romantic rela-tionship will be negatively associ-ated with a social network member’sapproval of that romantic relation-ship.

The next hypothesis is based on the invest-ment model’s inclusion of investments as apredictor of commitment.

H3: Perceiving a friend has invested ina romantic relationship will be pos-itively associated with a social net-work member’s approval of thatromantic relationship.

The investment model construct of com-mitment has a less clear relationship withfriend approval. It is possible that friendsmay use the commitment level of the targetperson when constructing their own approvalor disapproval of the relationship. Anotherpossibility is that friends are primarily con-cerned with the relationship outcomes that thetarget person receives from the relationship.The lack of prior research makes it difficultto develop a clear prediction regarding com-mitment; this leads to the following researchquestion regarding commitment:

RQ1: This research question will explorethe association between friend


perceived commitment and relation-ship approval.

Consistent with past findings on the struc-ture of the investment model (Le & Agnew,2003), perceived satisfaction in a romanticrelationship is expected to be the strongestpredictor of approval for that relationship,which fits with the theorized basis of satis-faction as the most direct evaluation of rela-tionship outcomes. Perceived alternatives andinvestments should also be predictive of socialnetwork approval of a relationship as thesevariables also involve a consideration of cur-rent and alternative relationship rewards andcosts.

Study 1

The first study employs data collected from amass-testing session to test the above hypo-theses. In this study, participants answeredquestions about their perceptions of a targetfriend’s romantic relationship. These ques-tions included perceptions of the friend’ssatisfaction, alternatives, investment, andcommitment.

Method

Participants and procedure

Data were collected from undergraduate stu-dents completing a questionnaire as partialfulfillment of research requirements in psy-chology classes. A total of 1,274 (mean age= 19 years, SD = 2.1; 54.9% female) peo-ple participated in the study. The majorityof participants reported being single (67.4%)and the remaining were in committed relation-ships (25.7%; 2.0% married, 2.0% living witha partner, 1.3% single with children, 0.1%divorced or separated, and 1.3% other), andmost participants self-reported their ethnicityas Caucasian (87.1%; 4.3% Asian Americanor Pacific Islander, 4.3% African Americanor Black, 1.8% Latino, 1.6% international stu-dent, 1.0% multiracial, 0.3% native American,and 0.4% other). Participants were asked toanswer questions about a friend’s romanticrelationship, thus being put in the role of thefriend providing their perceptions of a roman-tic relationship embedded in that network.

Participants were asked to choose a male orfemale friend who was involved in a roman-tic relationship based on the participant’s birthdate. If the participants’ birthday fell on anodd day (the first, third, fifth, etc.), theywere instructed to answer questions about amale friend; participants with birthdays on aneven day (the second, fourth, sixth, etc.) wereinstructed to answer about a female friend. Ofthe 1,274 participants, 28.3% were femalesreporting about female friends, 26.6% werefemales reporting about male friends, 20.9%were males reporting about female friends,and 24.2% were males reporting about malefriends. Thus, we were able to make compar-isons based on gender of the participants andfriends.

The target relationships participants answ-ered questions about were primarily exclusivedating relationships (77%; 7.5% nonexclusivedating, 7.5% cohabiting, 3.7% engaged butnot cohabiting, 2.2% engaged and cohabit-ing, and 2.3% married). Approximately 40%of the target relationships had existed forless than 1 year (31% were involved for1–2 years; 28% were involved for more than2 years), and the majority of target friendswere involved in heterosexual relationships(96.7%).

Measures

Perceived satisfaction. Participants first an-swered three questions regarding their percep-tions of the friend’s satisfaction with his or herromantic relationship. The items used werebased on the Investment Model Scale (IMS;Rusbult, Martz, & Agnew, 1998) and included“My friend is satisfied with his/her relation-ship,” “My friend’s relationship is much betterthan others’ relationships,” and “My friend’srelationship is close to ideal” (1–9 scale;α = .88).

Perceived alternatives. Participants answeredthree questions regarding their perceptions ofthe friend’s alternatives to his or her currentrelationship. The items used were based on theIMS (Rusbult et al., 1998) and included “Myfriend’s alternatives to his/her relationship areclose to ideal (dating another, spending time


with friends or on own, etc.),” “If they weren’tdating, my friend would do fine—he/shewould find another appealing person to date,”and “My friend’s alternatives are attractive(dating another, spending time with friends oron own, etc.)” (1–9 scale; α = .68).

Perceived investments. Participants respon-ded to three questions about their perceptionsof how much their friend had invested in therelationship. The items used were based onthe IMS (Rusbult et al., 1998) and included“My friend has put a great deal into his/herrelationship that he/she would lose if therelationship were to end,” “My friend is veryinvolved in the relationship—he/she has puta great deal into it,” and “Compared to otherpeople I know, my friend has invested a greatdeal in the relationship with his/her partner”(1–9 scale; α = .90).

Perceived commitment. Participants respon-ded to three questions, based on the IMS(Rusbult et al., 1998), about their percep-tions of the friend’s commitment. The itemsused included “My friend wants his/her rela-tionship to last a very long time,” “Myfriend is committed to maintaining his/herromantic relationship,” and “My friend wantshis/her relationship to last forever” (1–9scale; α = .68).

Approval or disapproval of friend’s roman-tic relationship. Participants answered fourquestions measuring their approval or dis-approval of their friend’s romantic relation-ship. These items included “I think my friend. . . Should Not/Should . . . continue in his orher current romantic relationship,” “I thinkmy friend . . . Does Not Have/Has . . . a cur-rent romantic relationship worth keeping,” “Ithink that this . . . Is Not/Is . . . a good cur-rent romantic relationship for my friend,” and“I am . . . Not Supportive/Supportive . . . ofmy friend’s current romantic relationship.”These items were reworded based on itemsoriginally designed to measure a memberof a romantic relationships perceptions ofsocial network member’s approval or disap-proval (Etcheverry & Agnew, 2004; Norma-tive Beliefs Scale; α = .94).

Control variables

Along with the above predictors, control vari-ables of participant gender (0 = female, 1= male) and length of target relationshipwere included in the analyses. Length of rela-tionship was coded such that 1 ≤ 1 month,2 = 1–2 months, 3 = 2–6 months, 4 =6–12 months, 5 = 1–2 years, 6 = 2–4 years,and 6 = 4+ years.

Results

An initial analysis of variance (ANOVA)examined the interaction and main effectsof participant and target gender on rela-tionship approval, perceived satisfaction, per-ceived alternatives, and perceived invest-ments. Only the effect of participant gen-der on perceived investments was signif-icant, F (1, 1273) = 6.86, p < .01, withfemale participants perceiving higher invest-ments (M = 6.70, SD = 0.07) than male par-ticipants (M = 6.42, SD = 0.08).

To test Hypotheses 1, 2, and 3 andResearch Question 1, a multiple regressionwas run with perceived satisfaction, per-ceived alternatives, perceived investments,and perceived commitment predicting partic-ipant approval of the target romantic rela-tionship (Table 1). Perceived satisfaction andalternatives were significantly associated withapproval; however, perceived investmentswere not. Perceived commitment did not addto the prediction of relationship approvalwhen assessed with the other hypothesizedpredictors.

Although both perceived satisfaction andalternatives were significant, perceived sat-isfaction was by far the stronger predictor,accounting for 52% of the variance in rela-tionship approval, while perceived alterna-tives explained less than 1% of the variance.These results highlight perceived satisfactionas the most prominent predictor of relation-ship approval.

Although no hypotheses were proposedregarding gender, gender differences weredeemed to be of general interest. There-fore, analyses were conducted examining boththe gender of the participant and the gen-der of the target friend interacting with


Table 1. Study 1 perceived investment model variables predicting romantic relationshipapproval

Model β T p < R2 F df p <

529 375.76 61,264 .001Participant gender −.018 0.25 .81Length of relationship .034 1.27 .21Perceived satisfaction .725 28.14 .00Perceived alternatives −.039 1.96 .05Perceived investments .036 1.25 .21Perceived commitment −.045 −0.82 .41

the investment model variables predictingrelationship approval. Perceived commitmentwas not directly hypothesized nor was it a sig-nificant predictor in the previous model andtherefore was not included in this analysis.

In these analyses, perceived satisfaction,perceived alternatives, perceived investments,participants’ gender, and gender of the friendwere entered into a multiple regressionanalysis predicting relationship approval. Inaddition, each of the perceived investmentmodel variables was included in two-wayinteractions with participant’s gender andfriend’s gender, respectively. Finally, three-way interaction terms were created for eachof the investment model variables interact-ing with both participant’s gender and friend’sgender.

This multiple regression indicated a sig-nificant interaction between perceived satis-faction and participant gender. Examinationsof the simple slopes for perceived satisfac-tion indicated that perceived satisfaction wasa significant predictor of relationship approvalfor female participants (β = .762, t = 19.41,p < .01) and male participants (β = .605, t =10.40, p < .01) but satisfaction was a sig-nificantly stronger predictor for female ascompared to male participants (t = −2.24,p < .05). The interaction of perceived invest-ments and participant gender was also signif-icant (t = −2.59, p < .01), with investmentsbeing a significant predictor for female partic-ipants (β = .127, t = 2.29, p < .05) but notfor male participants (β = −.098, t = −1.27,p = .21). None of the other two-way or three-way interactions was significant.

Discussion

The results of Study 1 indicate that interde-pendence theory variables contribute to theprediction of friend relationship approval inthat both perceived satisfaction and alter-natives were significantly associated withapproval. As anticipated, perceived satis-faction was a strong positive predictor ofapproval (vs. disapproval) of the friend’srelationship. Similarly, perceived alternativeswere a negative predictor of approval of thefriend’s romantic relationship. Although bothperceived satisfaction and alternatives weresignificant predictors of relationship approval,perceived satisfaction explained a substan-tially larger proportion of the variance.

In contrast, the investment model addi-tions of investment and commitment wereboth nonsignificant predictors of friend rela-tionship approval for the entire sample. How-ever, investments were a small but significantpredictor of relationship approval for femalebut not male participants. For female partic-ipants, investments were a significant predic-tor regardless of whether the target person’sgender was male or female. This result sug-gests that women may be more concernedwith what a friend has put into a relation-ship that is lost if the relationship ends.Male friends appear not as concerned withinvestments in others’ relationships. Even forfemale participants, investments explained asubstantially smaller amount of variance thanperceived satisfaction. Given that participantswere college-aged and so were most of thetarget relationship members, it is possiblethat for this age group, investments are not


valued in terms of determining relationshipapproval. For older participants with differ-ent priorities, investments such as sharedproperty and children may have a strongerimpact on approval. Another possibility isthat different types of investments are differ-entially influential on relationship approval.Prior research has argued that there are manydifferent types of investments (Goodfriend &Agnew, 2008), including tangible and intangi-ble investments. It is possible that more tangi-ble investments that are more easily observedby the friend will be more predictive thanintangible investments; however, the currentmeasure of investments failed to make thisdistinction.

An important result of Study 1 is thatperceived commitment was not significantlyassociated with perceived approval. The cor-relations between perceived commitment andrelationship approval (Table 2) indicate thatcommitment and relationship approval arepositively associated. However, after control-ling for perceived satisfaction and alterna-tives, perceived commitment is no longer asignificant predictor of relationship approval.This finding supports the possibility thatalthough network members can develop per-ceptions of a target person’s commitment, itis the relationship outcomes of the currentrelationship and potential alternatives that aremost relevant to the participant’s relationshipapproval.

The results of Study 1 indicate that per-ceived satisfaction is the strongest predictor

of social network members’ approval ofromantic relationship. The contribution ofperceived alternatives, while significant, wasvery small suggesting that this variable has amuch smaller impact on relationship approval.Although investments were a significant pre-dictor of approval for female participants, italso explained much less variance than satis-faction. Therefore, the following two studiesfocus on developing a better understanding ofthe association of perceived satisfaction withrelationship approval.

Despite the strong association between per-ceived satisfaction and approval in Study1, the correlational design precludes makingcausal inferences. It is possible that when asocial network member approves of a friend’srelationship, they then assume that the friendis satisfied in the relationship. In addition,other variables could be causing both per-ceived satisfaction and approval. For thisreason, Studies 2 and 3 experimentally manip-ulate perceived relationship satisfaction toconfirm its causal impact on approval for theromantic relationship.

H4: Participants in the high perceived sat-isfaction experimental condition willreport more approval for a targetromantic relationship than partici-pants in the low perceived satisfactionexperimental condition.

Studies 2 and 3 each employs distinctmanipulations of perceived satisfaction to

Table 2. Correlations of multiregression model variables for Study 1

1 2 3 4 5 6

1. Perceived satisfaction —2. Perceived alternatives 0.19∗∗ —3. Perceived investments 0.55∗∗ 0.16∗∗ —4. Perceived commitment 0.63∗∗ 0.06∗ 0.70∗∗ —5. Relationship approval 0.73∗∗ 0.10∗∗ 0.42∗∗ 0.46∗∗ —6. Participant gender −0.02 −0.04 −0.08 0.12∗∗ −0.02 —7. Length of relationship 0.20∗∗ 0.01 0.37∗∗ 0.32∗∗ 0.17 0.01M 5.93 6.11 6.57 7.08 0.45 4.64SD 2.01 1.64 1.89 1.99 0.50 1.42

∗p < .05. ∗∗p < .01.


create independent tests of the ability of per-ceived satisfaction to influence relationshipapproval. Using two separate manipulationsdecreases the likelihood of any one confoundproviding a reasonable alternative explanationof both results. Therefore, Studies 2 and 3use unique manipulations but the same mea-sures of relationship approval and perceivedsatisfaction.

Study 2

The second study manipulates social networkmembers’ perceived satisfaction by havingparticipants assimilate or contrast their per-ceptions of the friends’ relationships with anideal romantic relationship based on princi-ples of the feature-matching model (Tyver-sky, 1977). A similar technique was used byBroemer and Diehl (2003) to manipulate par-ticipants’ satisfaction with their own relation-ships, and in the current study, we used thisapproach to manipulate friends’ perception ofrelationship satisfaction.

In the Broemer and Diehl (2003) study,participants were asked to compare their ownromantic relationship with the characteristicsof the ideal romantic relationship. The feature-matching model indicates that when beingasked how Object A is similar to Object B,Object A is the target of comparison andObject B acts as the referent for compari-son (Holyoak & Gordon, 1983). Prior researchon the feature-matching model has found thatwhen people are more familiar with an object,they are better able to identify unique char-acteristics of that object. The more uniquean object is, the more dissimilar it will beconsidered from another object. However, thefeature-matching model argues that the effectsof the unique characteristics of an object onperceptions of dissimilarity are nonsymmetri-cal such that uniqueness will have a strongerimpact on dissimilarity when that object isthe target of comparison than when it is thereferent for comparison (Holyoak & Gordon,1983). Broemer and Diehl used this non-symmetrical characteristic of comparisons byasking participants to compare their relation-ship with the ideal relationship, where forsome participants, the romantic relationship

was the target of comparison, and for someparticipants, the romantic relationship wasthe referent. Specifically, Broemer and Diehlinstructed half of the participants to considerhow their own romantic relationship was sim-ilar to the ideal romantic relationship, whichleads to perceive less similarity between thetwo objects or a contrasting effect. The otherhalf of participants were instructed to comparehow the ideal romantic relationship was sim-ilar to their own romantic relationship, whichleads to perceive more similarity betweenthe two objects or an assimilation effect. Asexpected, participants in the contrastive con-dition reported lower levels of relationshipsatisfaction as compared to participants in theassimilative condition.

Study 2 applies a similar approach butinstead changes whether the target romanticrelationship is the target of comparison orthe referent. For Study 2, participants wereasked to indicate how similar the friend’sromantic relationship was to the ideal roman-tic relationship (contrast condition) or howan ideal romantic relationship was similar totheir friend’s romantic relationship (assimi-lation condition). These manipulations wereintended to create high perceived satisfaction(assimilation condition) and low perceivedsatisfaction (contrastive condition) groups.

Method

Participants

Undergraduate students participated in thisresearch in partial fulfillment of the researchrequirement in a psychology course. Partic-ipants took the perspective of social net-work members and were asked to think of afriend who is involved in a romantic relation-ship. Participants were randomly assigned toanswer about either a male or female friend.A total of 48 participants (25 females; 85%Caucasian, 8% African American, 7% other)were randomly assigned to either the assimi-lation or contrast condition.

Procedure and measures

Participants provided open-response answersto the following questions designed to promote


thinking about the ideal relationship: “What isideal romantic love?” “What are ideal roman-tic behaviors?” “What attributes does an idealpartner possess?” “What are everyday activi-ties for individuals in ideal romantic relation-ships?” and “What are traditional feelings andemotions of those in ideal romantic relation-ships?” All participants completed this task.

After answering these questions aboutideal relationships, participants were asked tothink about a male or female friend who isinvolved in a romantic relationship. Partici-pants were instructed to either compare thefriend’s relationship with the ideal roman-tic relationship (assimilation condition) or tocompare the ideal romantic relationship withthe friend’s romantic relationship (contrastcondition). Participants were given 5 min towrite their thoughts regarding these compar-isons and afterward were asked to indicatetheir approval of the friend’s romantic rela-tionship using the same four items from Study1 (α = .97). Finally, participants completed amodified version of the entire IMS (Rusbultet al., 1998) reframed to be answered aboutthe friend’s romantic relationship, includingmeasures of perceived relationship satisfac-tion (α = .94), perceived alternatives (α =.88), perceived investments (α = .83), andperceived commitment (α = .94). Some of theitems in the IMS were measured on differentmetrics so responses were transformed intoz-scores before being averaged across items.

Results

A manipulation check investigated if theexperimental conditions (two levels: assim-ilation to ideal and contrast from ideal)differed in perceived relationship satisfac-tion. The manipulation check indicated thatthe conditions differed significantly, F (1,44) = 4.46, p < .05, with the assimila-tion condition having significantly higherlevels of perceived satisfaction (M = 0.30,SD = 0.70) than the contrast condition (M =−0.19, SD = 0.81). Importantly, the manip-ulation did not cause a significant changein perceived alternatives, F (1, 44) = 1.87,p = .18; perceived investments, F (1, 44) =0.00, p = .96; or perceived commitment, F (1,

44) = 0.04, p = .84. Examination of theresponses provided indicated that all partic-ipants completed the comparison with theideal relationship and that the length of theresponses was similar across conditions.

In the second set of analyses, the assimila-tion/contrast manipulation significantly affect-ed participants’ approval or disapproval offriends’ relationships in the predicted direc-tion, F (1, 40) = 4.67, p < .05, R2 = .10,with the assimilation condition having higherlevels of approval for the friend’s relation-ship than the contrast condition (5.88 vs. 4.55,SDs = 1.77 and 2.17, respectively). Thus,changes in perceived satisfaction caused a cor-responding change in relationship approval.

To further test that the change in relation-ship approval was due to change in relation-ship satisfaction, bootstrapping mediationalanalyses were conducted (Preacher & Hayes,2008). The results of these analyses indicatedthat there was a significant indirect effectthrough perceived satisfaction (indirect effect= 1.16, SE = 0.52, 95% CI [0.21, 2.15]).Once the indirect mediational effect of sat-isfaction was taken into account, the directeffect of assimilation/contrast manipulationon relationship approval was not significant(β = .17, t = 0.57, p = .57). The sample sizeof Study 2 was chosen to test the causaleffect of satisfaction based on the effect sizedetermined in Study 1. Tests for interactionswith participant gender were not conducted asthe sample size for the current study wouldnot provide enough power to test interactiveeffects.

Study 3

Study 3 used a manipulation based on theavailability heuristic that was used by Broe-mer (2001) in a prior study to manipu-late interpersonal closeness. The availabilityheuristic refers to a cognitive shortcut thatpeople use when attempting to judge theprevalence or uniqueness of an event or objectin the environment (Tyversky & Kahneman,1973). The availability heuristic refers to thecommon tendency that when asked to indicatehow prevalent some object or event is, peopleoften use the ease of recall or how easy it is


to think of instances of that object or eventas a guide for how prevalent the object orevent is. Researchers have found that by turn-ing this heuristic around and systematicallymanipulating the ease of recall or availabilityof an object or event, it is possible to exper-imentally change participant’s judgments ofhow common or uncommon the object orevent is (e.g., Schwarz et al., 1991). Apply-ing this to relationships, Broemer employedthe availability heuristic by asking partici-pants to list either 5 (easy recall condition) or10 (hard recall condition) end-states or eventsin their romantic relationship. When askedto list fewer end-states or events, participantsviewed these events to be more common thanthose asked to list more end-states or events.This research found that by manipulating howcommon or uncommon these relationship end-states were perceived to be, participants feltthat closeness to a romantic partner could bemanipulated.

In Study 3, we sought to replicate theresults of Study 2 using the availability heuris-tic to manipulate perceived satisfaction in thefriend’s romantic relationship. Consistent withpast work (Broemer, 2001), participants wererandomly assigned to list few (easy recall con-dition) or many (hard recall condition) reasonswhy their friend is satisfied with the friend’scurrent romantic relationship.

Method

Participants and procedure

As in the previous studies, participants wererandomly assigned to think of either a femaleor male friend who was currently involvedin a romantic relationship. In total, 44 par-ticipants (22 women; 84% Caucasian, 4%African American, 4% Asian American) wererandomly assigned to either the easy or hardrecall condition and asked to provide reasonswhy their friend is satisfied with their currentromantic relationship. Participants in the easyrecall condition were asked to list three rea-sons their friend is satisfied, while participantsin the hard recall condition were asked tolist eight reasons. All participants in the easyrecall/list three conditions listed three reasons.

Seventy percent of participants in the hardrecall/list eight listed eight reasons. Partici-pants who listed fewer than eight reasons inthe hard recall condition were retained in anal-yses as not listing all eight asked for reasonswas viewed as evidence of the difficulty of thetask in this condition, which is consistent withthe rationale underlying this manipulation.

Measures

Following the recall task, participants com-pleted two questions regarding how hard itwas to recall the reasons why their friendwas satisfied. These items were “I found itvery easy/hard to come up with three reasonswhy my friend is satisfied” and “I could eas-ily generate more than the three reasons whymy friend is satisfied” (reverse coded). Thesetwo items were measured on a 1–7 scale withhigher numbers indicating a greater subjectivedifficulty of recall (correlation between itemsr = .77) and were averaged to create a mea-sure of difficulty of recall.

Next, participants completed the samemeasures of participant approval (α = .98) ofthe friend’s romantic relationship as used inStudies 1 and 2. Finally, we administeredthe modified IMS (used in Study 2; Rusbultet al., 1998) to assess perceived relationshipsatisfaction (α = .92), perceived alternatives(α = .75), perceived investments (α = .84),and perceived commitment (α = .94). Similarto the previous study, IMS items were trans-formed to z scores.

Results

An initial analysis of the difficulty of recallmeasure indicates that participants asked tolist three reasons why their friend was sat-isfied reported the task was significantlyeasier than participants asked to list eightreasons (Ms = 3.40 vs. 5.24, SDs = 1.67and 1.43, respectively), F (1, 43) = 14.94,p < .001. A second manipulation check wasrun to confirm that participants in the “listthree” condition reported significantly higherlevels of perceived friend satisfaction thanin the “list eight” condition (Ms = 0.26 vs.−0.38, SDs = 0.91 and 1.02, respectively),


F (1, 43) = 5.17, p < .05. Additional analy-ses showed that the condition variable was nota significant predictor of perceived alterna-tives, F (1, 43) = 1.93, p = .18, or perceivedinvestments, F (1, 43) = 2.43, p = .13, butwas a significant predictor of perceived com-mitment, F (1, 43) = 4.90, p < .05.

An ANOVA was conducted to determine ifparticipants’ relationship approval varied bycondition. The analysis indicated that those inthe “list three” condition reported significantlyhigher levels of approval for their friends’relationship compared to participants in the“list eight” condition (Ms = 6.16 vs. 5.17,SDs = 1.27 and 1.95, respectively), F (1, 43)= 4.29, p < .05.

Mediational analyses were conducted usingthe Preacher and Hayes (2008) bootstrappingmethod to examine if perceived satisfactionmediated the effect of the manipulation in pre-dicting relationship approval. Perceived com-mitment was significantly predicted by condi-tion; therefore, commitment was included asan additional mediator along with perceivedsatisfaction. The results of this bootstrappingmediation analysis indicated that there was asignificant indirect effect through perceivedsatisfaction (indirect effect = 0.20, SE =0.10, 95% CI [0.03, 0.43]). However, the indi-rect effect of condition through commitmentwas not significant (indirect effect = 0.02,SE = 0.04, 95% CI [−0.02, 0.15]). Once theindirect mediational effect of satisfaction andcommitment was taken into account, the directeffect of three/eight listing condition on rela-tionship approval was not significant (β = .02,t = 0.40, p = .69). As with Study 2, the sam-ple size for the current study lacked power totest for interactions with participant gender topredict relationship approval.

Studies 2 and 3 discussion

Studies 2 and 3 complement the results ofStudy 1 and provide further support for therole of perceived satisfaction in social net-work approval. Using different manipulations,these two studies provide strong evidence forthe causal association between perceived sat-isfaction and approval for a friend’s romanticrelationship.

General Discussion

The three studies provide strong evidence forthe role of perceived satisfaction in predictingsocial network approval for a romantic rela-tionship. The results of Study 1 indicate thatperceived satisfaction is the strongest con-tributor of the investment model variables tofriend approval for a romantic relationship.Studies 2 and 3 support the results of Study1 by demonstrating perceived satisfaction’scausal association with relationship approval.Perceived alternatives were a significant pre-dictor in Study 1, and investments were asignificant predictor for female participants,although the amount of variance in relation-ship approval explained by these variableswas small when compared to perceived sat-isfaction. Perceived commitment was a non-significant predictor of approval for a roman-tic relationship.

These results suggest that friends, suchas members of the romantic relationship, areconcerned with the level or rewards and costsor outcomes in a relationship. Therefore, per-ceived satisfaction, as the best indicator ofrelationship outcomes, is a predominate pre-dictor of relationship approval. This find-ing fits with interdependence theories’ focuson relationship outcomes as a primary toolfor evaluating the quality of relationships.Assuming that friends evaluate relationshipsin a similar manner as the people in thoserelationships, it makes sense for satisfactionto guide relationship approval.

It could be argued when examining themanipulations used in Studies 2 and 3 thatthese manipulations affected perceived satis-faction, not the underlying rewards and costsargued to be important in interdependence the-ory (Kelley & Thibaut, 1978). Although thecurrent research supports the importance ofperceived satisfaction, additional research isneeded to confirm that friends are concernedwith rewards and costs in a target relationship,not just the general happiness level of a targetperson.

It is worth noting that the measures usedin this study are only perceived relationshipsatisfaction, alternatives, and investments. Itis unclear how accurate friend members are


regarding these variables. However, it isassumed that whether they are inaccurate oraccurate, it is these perceptions of relationshipvariables that determine friend approval for aromantic relationship.

Alternative bases for relationship approval

The current research indicates the impor-tance of the interdependence theory variables,especially perceived satisfaction in predictingrelationship approval. The current results donot leave out the possibility that other fac-tors might predict relationship approval. Forexample, in Felmlee’s (2001) research, someparticipants indicated disapproval of a roman-tic relationship because the romance is seen asdisrupting the friendship with the social net-work member. Social comparison processes(Buunk & Oldersma, 2001) might also mat-ter, with friends using a target relationship aseither an inspirational target for upward socialcomparison or a source of positive feelingsas a target for downward social comparison.Jealousy may also be an important variable,especially if the friend wishes to date one ofthe target relationship members.

The current research was not designed totest these alternative predictors of relation-ship approval, leaving open the possibilitythat these other variables matter. However, itmight also be worth considering that socialcomparison processes, jealousy, and otherfactors might actually influence perceived sat-isfaction and alternatives, which then pre-dict relationship approval. These unansweredquestions suggest that predictors of social net-work approval for a relationship is an impor-tant topic for relationship research to study.

Future directions and conclusions

Very little research has examined the deter-minants of social network approval. There-fore, the current research is only an initialattempt to understand what is likely a complexaspect of social network influence on romanticrelationships. However, the current researchdoes suggest the importance of perceived sat-isfaction and underlying assumptions of socialexchange models (i.e., interdependence the-ory) in guiding future research in this area.

Several key areas needing future researchare identified. First, it will be important toreplicate these results with older and morediverse samples, other types of social net-work members such as parents and familymembers, and more diverse types of relation-ships (e.g., marriages). Assuming that par-ents, family members, and other social net-work members also desire the best outcomesfor the target romantic relationship member,we might expect an interdependence theoryapproach to predict that relationship approvalwill be effective. However, it is likely for par-ents and family members that other factorsare associated with relationship approval, suchas the long-term viability of a relationship.Future research should explore these possibil-ities. In addition, future work should explorepotential moderators of the association ofinterdependence theory and investment modelvariables with relationship approval. Severalpossible moderators include the nature of thetarget romantic relationship (dating, cohabitat-ing, and married) and the degree of closenessbetween the friend and the member of the tar-get relationship.

These studies can be used to test if per-ceived satisfaction remains such a strongpredictor, if perceived investments remain asignificant predictor for women only, and ifperceived alternatives ever explain more ofthe variance in relationship approval. In addi-tion, it will be important to examine situationswhere the assumption that social networkmembers approve more of romantic relation-ships with high rewards and low costs may beviolated.

Conclusion

The findings from the three studies, one cor-relational and two experimental, support therelevance of interdependence theory to under-stand the factors that contribute to friendapproval for a target romantic relationship.Friend-perceived satisfaction was a particu-larly strong predictor of relationship approval.This research represents an initial step towarddeveloping a broader understanding of thefactors that contribute to friend and social net-work approval of romantic relationships.


References

Agnew, C. R., Loving, T. J., & Drigotas, S. M. (2001).Substituting the forest for the trees: Social networksand the prediction of romantic relationship state andfate. Journal of Personality and Social Psychology,81, 1042–1057.

Arriaga, X. B., & Agnew, C. R. (2001). Being commit-ted: Affective, cognitive, and conative components ofrelationship commitment. Personality and Social Psy-chology Bulletin, 27, 1190–1203.

Berscheid, E. (1999). The greening of relationship sci-ence. American Psychologist, 54, 260–266.

Broemer, P. (2001). Ease of recall moderates the impactof relationship-related goals on judgments of inter-personal closeness. Journal of Experimental SocialPsychology, 37, 261–266.

Broemer, P., & Diehl, M. (2003). What you think is whatyou get: Comparative evaluations of close relation-ships. Personality and Social Psychology Bulletin, 29,1560–1569.

Burger, E., & Milardo, R. M. (1995). Marital interde-pendence and social networks. Journal of Social andPersonal Relationships, 12, 403–415.

Buunk, B. P., & Oldersma, F. L. (2001). Social com-parison and close relationships. In G. J. O. Fletcher& M. S. Clark (Eds.), Blackwell handbook of socialpsychology: Interpersonal processes (pp. 388–408).Malden, MA: Blackwell.

Clark, M. S., & Lemay, E. P., Jr. (2010). Close relation-ships. In S. T. Fiske, D. T. Gilbert, & G. Lindzey(Eds.), The handbook of social psychology(pp. 898–940). Hoboken, NJ: Wiley.

Etcheverry, P. E., & Agnew C. R. (2004). Subjectivenorms and the prediction of romantic relationshipstate and fate. Personal Relationships, 11, 409–428.

Etcheverry, P. E., Le, B., & Charania, M. R. (2008). Per-ceived versus reported social referent approval andromantic relationship commitment and persistence.Personal Relationships, 15, 281–295.

Felmlee, D. H. (2001). No couple is an island: Asocial network perspective on dyadic stability. SocialForces, 4, 1259–1287.

Goodfriend, W., & Agnew, C. R. (2008). Sunken costsand desired plans: Examining different types ofinvestments in close relationships. Personality andSocial Psychology Bulletin, 34, 1639–1652.

Holyoak, K. J., & Gordon, P. C. (1983). Social referencepoints. Journal of Personality and Social Psychology,44, 881–887.

Kelley, H. H., & Thibaut, J. W. (1978). Interpersonalrelations: A theory of interdependence. New York,NY: Wiley.

Kim, H. J., & Stiff, J. B. (1991). Social networks and thedevelopment of close relationships. Human Commu-nication Research, 18, 70–91.

Le, B., & Agnew, C. R. (2003). Commitment and its the-orized determinants: A meta-analysis of the invest-ment model. Personal Relationships, 10, 37–57.

Le, B., Dove, N. L., Agnew, C. R., Korn, M. S., &Mutso, A. A. (2010). Predicting non-marital romantic

relationship dissolution: A meta-analytic synthesis.Personal Relationships, 17, 377–390.

Leslie, L. A., Huston, T. L., & Johnson, M. P. (1986).Parental reactions to dating relationships: Do theymake a difference? Journal of Marriage and theFamily, 48, 57–66.

Loving, T. J. (2006). Predicting dating relationship fatewith insiders’ and outsiders’ perspectives: Who andwhat is asked matters. Personal Relationships, 13,349–362.

MacDonald, T. K., & Ross, M. (1999). Assessing theaccuracy of predictions about dating relationships:How and why do lovers’ predictions differ fromthose made by observers? Personality and SocialPsychology Bulletin, 25, 1417–1429.

Milardo, R. M. (1982). Friendship networks in develop-ing relationships: Converging and diverging socialenvironments. Social Psychology Quarterly, 45,162–172.

Milardo, R. M. (1988). Families and social networks: Anoverview of theory and methodology. In Milardo, R.M. (Ed.), Families and social networks (pp. 13–47).Thousand Oaks, CA: Sage.

Preacher, K. J., & Hayes, A. F. (2008). Asymptotic andresampling strategies for assessing and comparingindirect effects in multiple mediator models. BehaviorResearch Methods, 40, 879–891.

Rusbult, C. E. (1980). Commitment and satisfaction inromantic associations: A test of the investmentmodel. Journal of Experimental Social Psychology,16, 172–186.

Rusbult, C. E., & Agnew, C. R. (2010). Prosocial moti-vation and behavior in close relationships. In M.Mikulincer & P. R. Shaver (Eds.), Prosocial motives,emotions, and behavior: The better angels of ournature (pp. 327–345). Washington, DC: AmericanPsychological Association.

Rusbult, C. E., Arriaga, X. B., & Agnew, C. R. (2001).Interdependence in close relationships. In G. J. O.Fletcher & M. S. Clark (Eds.), Blackwell handbookof social psychology: Interpersonal processes (Vol. 2,pp. 359–387). Oxford, UK: Blackwell.

Rusbult, C. E., Martz, J. M., & Agnew, C. R. (1998).The Investment Model Scale: Measuring commit-ment level, satisfaction level, quality of alterna-tives, and investment size. Personal Relationships, 5,357–391.

Sprecher, S., & Felmlee, D. (1992). The influence ofparents and friends on the quality and stability ofromantic relationships: A three-wave longitudinalinvestigation. Journal of Marriage and the Family, 54,888–900.

Sprecher, S., & Felmlee, D. (2000). Romantic partners’perceptions of social network attributes with thepassage of time and relationship transitions. PersonalRelationships, 7, 325–340.

Sprecher, S., Felmlee, D., Orbuch, T. L., & Willetts,M. C. (2001). Social networks and change in per-sonal relationships. In A. L. Vangelisti, H. T. Reis,& M. A. Fitzpatrick (Eds.), Stability and changein relationships. Advances in personal relationships


(pp. 257–284). New York, NY: Cambridge Univer-sity Press.

Schwarz, N., Bless, H., Strack, F., Klumpp, G., Rittenauer-Schatka, H., & Simons, A. (1991). Ease of retrieval asinformation: Another look at the availability heuris-tic. Journal of Personality and Social Psychology, 61,195–202.

Tyversky, A. (1977). Features of similarity. Psychologi-cal Review, 84, 327–352.

Tyversky, A., & Kahneman, D. (1973). Availability: Aheuristic for judging frequency and probability. Cog-nitive Psychology, 4, 207–232.

Van Lear, A., Koerner, A. F., & Allen, D. (2006). Rela-tionship typologies. In A.Vangelisti & D. Perlmann(Eds.), The Cambridge handbook of personal rela-tionships (pp. 91–111). Cambridge, UK: CambridgeUniversity Press.

predictors of friend approval for romantic -

Documents