Does (Dis)Similarity Information about a New Acquaintance Lead to Liking or Repulsion? An Experimental Test of a Classic Social Psychology Issue

Abstract

Decades of social psychology research has established the importance of similarity in leading to attraction. However, in response to early social psychology experiments demonstrating the similarity effect, Rosenbaum proposed the repulsion hypothesis, arguing that similarity does not lead to liking, but rather, dissimilarity leads to repulsion. Research to address whether dissimilarity carries more weight than similarity has generally involved participants’ reactions to sterile information about a bogus other whom they never meet. In contrast, in this study (N = 150), individuals first greeted another participant over Skype before they received manipulated (bogus) information on similarity or dissimilarity. In support of the similarity-attraction hypothesis, the two-step experimental design indicated that the participants in the similarity condition experienced an increase in liking and other positive reactions from before to after the receipt of the bogus similarity information. Participants in the dissimilarity condition, however, experienced no change (i.e., no repulsion effect).

Keywords

homophily interpersonal attraction repulsion effect similarity effect

Decades of social psychology literature in both sociology (Lazarsfeld and Merton 1954; McPherson, Smith-Lovin, and Cook 2001) and psychology (Byrne 1971; Montoya, Horton, and Kirchner 2008; Newcomb 1961) has established the important role of similarity in leading to attraction and affiliation. The similarity effect has come to be defined as one of the most robust findings in social psychology (Berscheid and Reis 1998; Finkel and Baumeister 2010; Layton and Insko 1974; McPherson et al. 2001) and is discussed in most social psychology textbooks. A related concept, homophily, refers to the principle that contact among similar others occurs at a greater rate than contact among dissimilar others (Lawrence and Shah 2017; McPherson et al. 2001). This greater contact among similar others (than dissimilar others) can occur in part because of the consequence of structural factors that bring similar people together in proximity. Beyond the baseline degree of homophily that occurs due to structural factors, people also show a preference for similar others. Combined, the forces of structural and choice homophily create homogeneous networks.

People can be similar to another on a number of dimensions, including attitudes and beliefs, personality, appearance, hobbies, and background/demographic characteristics (Baxter and West 2003; McCroskey, McCroskey, and Richmond 2006; Sprecher 1998). The sociological literature (e.g., Lazarsfeld and Merton 1954) distinguishes between status homophily (sociodemographic variables) and value homophily (based on values, attitudes, beliefs, internal states). Various theoretical explanations have been offered to explain the preference for similarity and why similarity leads to attraction, including that similarity (especially in attitudes and beliefs) is consensually validating and reinforcing and therefore feels good (Byrne 1971; Clore and Byrne 1974), similarity leads to uncertainty reduction and predictability (Berger and Calabrese 1974), and similarity can lead to the expectation of and the actual occurrence of enjoyable and rewarding interactions (Burleson and Denton 1992; Burleson, Kunkel, and Birch 1994; Sprecher et al. 2013; Werner and Parmelee 1979).

Although traditionally people did not learn about how they were (or were not) similar to someone until they met and self-disclosed, today people often judge the degree of similarity of a potential friend or romantic partner prior to meeting the person or sometimes after a very brief introduction, based on information presented in social media or internet dating profiles. They then make estimates of how rewarding the relationship will likely be and how much they will like the other person based on this information. According to predicted outcome theory (Sunnafrank 1986), people make initial judgments based on limited information (including degree of similarity) and forecast future outcomes for the relationship. These forecasts influence whether people pursue relationships with new acquaintances.

While early research in social psychology found strong and consistent evidence for the similarity-attraction effect and explained it primarily based on a reinforcement model in which similar others contribute to good feelings and dissimilar others lead to negative feelings (as summarized in Montoya and Horton 2014), an alternative perspective was presented, referred to as the repulsion hypothesis (Rosenbaum 1986). According to this perspective, dissimilarity leads to repulsion, whereas similarity is inconsequential. Early social psychological research was conducted to compare the validity of the two perspectives, with mixed results. The purpose of this experiment was to provide a unique, new test of the validity of the two predictions using a modified version of the bogus stranger paradigm (Byrne 1971), in which the stranger was real (not bogus) and there was a two-step process in which the two strangers greeted each other briefly (over Skype) and then received bogus information about each other. They completed reactions to each other after each of the steps, and thus change in liking could be examined.

The Bogus Stranger Paradigm to Assess the Similarity Effect

The main paradigm in social psychology that established the important role of similarity in leading to attraction was Byrne’s (1961) bogus stranger method, which has also been referred to as the “phantom other” approach. The basic procedure was as follows: (1) students completed an attitude survey early in the semester; (2) later they participated in another study and were told, in a cover story, that they would judge an individual from another class who had completed the same survey they had in the prior study; (3) in actuality, the participants were given a bogus survey that was made to be either similar or dissimilar to them or some variation in between; and (4) the participants then provided their reactions to the bogus other, which included a measure of liking. The participants never met the other because in the traditional stranger paradigm, the other was bogus and did not exist. Research using the bogus stranger paradigm found strong support for the similarity-attraction effect. The more similar the bogus other was to the participant, the more that the participant reported liking the other (Byrne 1971, 1997).

Following this initial research, Byrne and his colleagues conducted many additional experiments with variations of the bogus stranger paradigm and found strong and consistent evidence for the similarity-attraction effect (Byrne 1971, 1997). The effect came to be defined as one of the strongest documented generalizations in social psychology (e.g., Berscheid and Reis 1998; Finkel and Baumeister 2010; Layton and Insko 1974; McPherson et al. 2001). Considerable research beyond the bogus stranger paradigm, including studies involving brief interactions of pairs of previously unacquainted individuals, also has shown support for the similarity-liking effect whether considering actual or perceived similarity (Montoya et al. 2008).

The Repulsion Hypothesis

In response to the early findings on the similarity effect conducted with the bogus stranger paradigm, Rosenbaum (1986) proposed the repulsion hypothesis as an alternative explanation. He argued that similarity has no consequences for liking, but rather, dissimilarity leads to repulsion. The crux of his argument is that the research based on the Byrne bogus stranger paradigm approach was missing an adequate control, a condition in which participants received no attitude information about the bogus other. Rosenbaum designed such an experiment by having a no-attitude condition in addition to the similarity and dissimilarity conditions. More specifically, participants received photograph information about a target so that all participants (even those in the no-attitude condition) had information on which to base their reactions. Rosenbaum found no differences in attraction to the target between those in the similarity condition and those in the no-attitude condition, but both groups reported more attraction than those in the dissimilarity condition.

In a second experiment (Rosenbaum 1986), participants reacted to the profile of a stranger who was described by personality traits and included a political affiliation either similar or dissimilar to the participant, or no party affiliation was mentioned (the participants were obtained from presidential caucuses in Iowa). Similar to the findings in the first experiment, Rosenbaum (1986) found no differences in liking for the target between the participants in the no-information condition (i.e., no information on political party affiliation) and those in the similarity condition (same affiliation) but found that the participants in both conditions liked the target more than those who believed that the target had a different political party affiliation.

In explaining why dissimilarity leads to repulsion, Rosenbaum (1986) argued that learning that someone is dissimilar leads to an aversive state (psychological discomfort) due to the inconsistency or imbalance (between the other’s traits, attitudes, or beliefs and one’s own). In the acquaintance process, once dissimilar others are eliminated due to repulsion, one is left attracted to or in proximity with those who are similar, according to Rosenbaum.

Follow-up Research

Smeaton, Byrne, and Murnen (1989) replied to Rosenbaum’s position by arguing that Rosenbaum’s premise that the no-attitude condition presents no information about similarity is unrealistic because similarity can still be assessed (or assumed) based on the limited information presented (e.g., the photograph) or even based on no information. They conducted additional experiments to emphasize that similarity leads to attraction rather than that dissimilarity leads to repulsion. For example, in one study, they had participants respond to one of three versions of a bogus attitude survey ostensibly completed by another student. Each version included eight items with dissimilar responses, but the number of similar responses (10, 28, 62) was varied. Their results indicated that with dissimilarity held constant, similarity independently affected attraction, thus disconfirming the repulsion hypothesis.

Research on this debate continued, including further examination of the outcomes (e.g., liking) of a no-information control group compared to similarity and dissimilarity information within the bogus stranger paradigm approach (Hoyle 1993; Pilkington and Lydon 1997; Singh and Ho 2000; Singh and Tan 1992; Tan and Singh 1995). Generally, this follow-up research demonstrated that people assume similarity in a no-attitude condition and therefore dissimilarity information may carry more weight than similarity information in attraction. These studies, however, involved participants reacting to bogus information about someone whom they did not meet because the other person did not exist (i.e., was a bogus stranger).

Purposes of this Study

In the current study, the debate of whether similarity leads to attraction or dissimilarity leads to repulsion is revisited, although with key differences from the prior bogus stranger paradigm. Therefore, this research adds to the classic attraction-repulsion debate in social psychology.

In most past bogus-stranger experiments, participants responded to measures of liking for the bogus other only once, which was after they received the bogus similarity information. As a consequence, it could not be examined whether similarity leads to an increase in attraction or repulsion leads to a decrease in attraction, a comparison that could be argued to be the more definitive test of the competing hypotheses. In one exception, Singh and Tan (1992, Study 1) had their participants initially think about a same-sex stranger (another student), who had ostensibly been randomly selected by a computer, and provide reactions to this stranger before receiving information about the stranger. Then, the participants were given the bogus similarity information about the stranger and asked to provide their reactions again. The researchers found that participants who received similarity information increased in their liking of the stranger, whereas participants who received dissimilarity information decreased in their liking. Their study, however, was low in ecological validity in that it would likely have been difficult for participants to provide reactions based on only thinking about a hypothetical stranger without any information about him or her. Therefore, the first unique aspect of the present study is that there was a two-step design. The first step of the procedure was a greeting between two real participants (thus, the other person and the pending interaction were made salient), after which they then completed measures of expectations of liking the other and enjoying the interaction. Then, at a second step, the participants received the bogus similarity or dissimilarity information about each other and completed reaction measures again. Because the participants completed measures that assessed their reactions to their assigned partner both immediately after they greeted each other over Skype and again after they received the similarity or the dissimilarity information, changes in reactions from before to after learning the other is dissimilar or similar could be examined.

This two-step design, involving actual participants who first greeted each other, mirrors how first impressions are now often formed in real-world contexts. For example, two people may meet briefly in a social setting and then learn information via Facebook or other online venues about each other before interacting again (Antheunis, Valkenburg, and Peter 2010; Yang, Brown, and Braun 2014). To further enhance ecological validity, similarity was operationalized through items referring to hobbies, interests, and self-descriptions, which represent the type of information that could potentially appear on social media or in a profile at an internet dating site.

The first purpose of the study was to examine whether the manipulated (bogus) similarity versus dissimilarity information, which was presented after the participants greeted each other, would nonetheless still affect initial liking. Because prior bogus experimental studies did not include an opportunity for two real people to greet each other before receiving the manipulated similarity information, it is an empirical question as to whether the bogus similarity or dissimilarity information will have an influence in the situation of participants responding to an actual person. An early study by Layton and Insko (1974) found that anticipation of future interaction with the other actually reduced the effect of similarity. Nonetheless, because there has been strong empirical evidence for the similarity effect in the absence of a lot of other information, the similarity-attraction effect is likely to be evidenced. Therefore, the first hypothesis is:

Hypothesis 1: Participants who receive bogus similarity information about another participant (after having a brief greeting) will report greater expectation of liking the other, greater expectation for an enjoyable interaction, and greater desire to interact than participants who receive bogus dissimilarity information.

The second and most unique purpose of this study was to distinguish between the effects of similarity versus dissimilarity on changes in expectations for liking and in other reactions between the assessments obtained pre-manipulation (of similarity) and assessments obtained post-manipulation. Because liking and the other initial reactions were obtained two times, first after the brief greeting and then again after the presentation of the bogus similarity or dissimilarity information, changes in the ratings can be compared between the two experimental groups. If the repulsion hypothesis receives more support, then those who receive dissimilarity information will have a decrease in their liking, and there will be no change in liking among those who receive similarity information. Conversely, if the original similarity-attraction hypothesis receives more support, those who receive similarity information will experience an increase in their liking, and there will be no change in liking for those who receive dissimilarity information. Of course, both predictions may be supported, if similarity information leads to an increase in liking and dissimilarity information leads to a decrease in liking. Because a comparison of the similarity effect versus the repulsion effect has not been previously conducted in a two-step design of this nature (i.e., two real participants who greet each other at a first step and then receive the similarity/dissimilarity information at a second step), a research question (rather than an hypothesis) is posed:

Research Question 1: Does the presentation of similarity information (after a greeting between two real people) lead to an increase in expectations of liking and other positive reactions, and/or does the presentation of dissimilarity information lead to a decrease in expectations of liking and other positive reactions?

Method

Participants

The participants were 150 undergraduate students (120 females and 27 males; no data on sex were available for 3 participants) who were obtained from the psychology participant pool and sociology classes at a U.S. Midwestern university.¹ The study was conducted over two semesters to run enough sessions; the final sample size was a function of how many sessions could be conducted across two semesters.

Participants who signed up for the study were scheduled in pairs for an experimental session based on their availability. Because a large majority of the participants were female, only female-female and female-male dyads were scheduled. Of the 75 dyads, 47 were female-female pairs, 25 were male-female pairs, and in 3 dyads, the sex composition could not be determined because no information was available on the sex of one of the members. The mean age of the participants was 20.04 years (SD = 2.99). A majority (76.7 percent) of the participants identified themselves as white, 12.0 percent were black, 7.3 percent were Hispanic/Latino, and 1.3 percent were Asian.

Procedure

The initial greeting

In each pair, the two participants were scheduled to report to the experiment at the same time, each to a different room in one area of a university building. Each participant was met by his or her own experimenter, who first directed the participant to read an informed consent form. Then the experimenters arranged for their participants to greet each other over Skype, with both audio and video turned on (this was very brief, almost like a hello wave). After the brief greeting, participants completed a rating form that made it possible to create a manipulation of similarity (to be discussed further below). Each participant subsequently completed an online survey that included many background questions and also measures of their initial reactions to the other based on the brief Skype introduction.

The similarity/dissimilarity manipulation

After the participants engaged in the video greeting with each other, they completed a self-descriptive form, which included several items that asked them to describe themselves with dichotomous responses. Some of the items began with “Which do you prefer?” (e.g., reality show vs. sitcom, shop alone vs. shop with others, big party vs. romantic dinner for two). Other items began, “Which best describes you?” (e.g., introvert vs. extrovert, sloppy vs. neat freak, dreamer vs. doer). The form had 17 items and has been used in prior research (e.g., Sprecher, 2014). The participants were told in advance that their form would be shared with the other participant.

After the participant completed the self-descriptive rating form, the experimenter picked it up and stated that he or she would soon return with the other participant’s completed form. During this time, the participants completed an online survey that included questions about their background but also assessments of their initial reactions to the other based on the Skype greeting. While their participants were completing this online survey, the two experimenters met in the hallway between the experimental rooms and fabricated a new form for each of their participants. Depending on the condition to which their participant was randomly assigned,² the experimenters prepared each bogus version so that the pair had 3 of the 17 items either similar or dissimilar.³ Items selected for the similar responses (in the dissimilarity condition) or for the dissimilar responses (in the similarity condition) were randomly determined. After the fabricated information was exchanged between the dyad members and they reacted to the information, the dyad then proceeded to an actual interaction (data not considered in this study).

Measures

Measures after the brief greeting

As noted previously, participants were asked a few questions in the initial online survey, administered after the brief Skype greeting, about their first impressions of their interaction partner. More specifically, they were asked how much they thought they would like the other, how much they thought they would enjoy the interaction, and how eager they were to interact with the other. Each question was followed by a 1 (not at all) to 7 (a great deal) response scale.

Measures after the exchange of bogus similarity information

After receiving the bogus information, participants completed a brief reaction to the other again. This brief online survey included the same measures as previously described—assessing the expectation of liking the other, expectation of enjoying an interaction with the other, and desire to interact with the other.⁴ Participants were also asked two questions about their perceptions of similarity with the other person: “How much do you think you have in common with the other person?” (1 = nothing or almost nothing, to 7 = a great deal) and “How similar do you think you and the Other are likely to be?” (1 = not at all, to 7 = a great deal). A composite was created of the two similarity items, which had a Cronbach’s alpha of .91.

Results

Preliminary Information: Ratings after the Brief Introduction

After the brief introduction with their interaction partner, participants reported, on average, a moderate degree of expecting to like the other (M = 4.61, SD = .91), expecting to enjoy the interaction (M = 4.36, SD = 1.02), and eagerness to interact with the other (M = 4.11, SD = 1.17). These ratings were obtained before the receipt of the bogus similarity or dissimilarity information, and therefore not surprisingly, these ratings were not significantly different between the conditions (ts ranged from .51 to 1.53, all p > .13).⁵ All analyses below are based on the total sample of dyads with data from both members.

Ratings after the Bogus Similarity or Dissimilarity Information: Test of Hypothesis 1

To test Hypothesis 1, the group assigned to the similarity information was compared with the group assigned to the dissimilarity information in their reactions after receiving the bogus information, via an independent samples t-test. (As indicated in note 2, the intraclass correlations for the dyad members were low, suggesting independence between dyadic members, and thus individual analyses were conducted.). In support of the hypothesis, those in the similarity condition, compared to those in the dissimilarity condition, reported greater expectations of liking the other, greater expectations of enjoying the interaction, and greater eagerness to interact (see the means in Table 1). In addition, the composite score of perceived similarity was significantly higher for those in the similarity condition than for those in the dissimilarity condition, which provides support for the success of the manipulation.

Table 1.

Reactions after Receiving the Bogus Similarity Versus Dissimilarity Information

	Similarity Condition(n = 77)	Dissimilarity Condition(n = 73)
	M (SD)	M (SD)	t	d	95% CI
Expectation of liking the other	5.27 (1.03)	4.42 (.80)	5.60***	.92	[.549, 1.147]
Expectation of enjoying the interaction	5.03 (1.06)	4.33 (.96)	4.21***	.59	[.370, 1.024]
Eagerness to interact	4.71 (1.17)	4.22 (1.06)	2.72**	.44	[.135, .855]
Similarity composite (manipulation check)	5.66 (1.04)	2.69 (1.00)	17.83***	2.91	[2.641, 3.300]

Note: d = Cohen’s d (effect size); CI = confidence interval.

p < .01, ***p < .001.

Change in Reactions as a Function of the Similarity Manipulation: Research Question 1

The differences found in expectations for liking and other reactions after the receipt of bogus similarity information (as reported previously) could occur for multiple reasons. The differences could occur because similarity information led to an enhancement of positive evaluations or the differences could occur because dissimilarity information led to a decrease in positive evaluations. Another possibility is that both shifts were occurring. To address Research Question 1, a mixed-model analysis of variance was conducted in which the similarity manipulation was the between-subjects variable, and timing of assessment (before or after the receipt of the bogus information) was the within-subjects variable.

This analysis yielded significant effects for assessment (pre– vs. post–bogus information) for all three ratings (all p < .001), with the ratings higher at the second assessment than at the first assessment: expectation of liking the other (M = 4.61, SD = .91 vs. M = 4.87, SD = 1.02), expectation of enjoying the interaction (M = 4.36, SD = 1.02 vs. M = 4.69, SD = 1.07), and eagerness to interact (M = 4.11, SD = 1.17 vs. M = 4.47, SD = 1.15). In addition, and in support of Hypothesis 1, the similarity manipulation was significant for all three reactions (expected liking, expected enjoyment, eagerness to interact); see means in Table 1. These main effects, however, were qualified by a significant assessment × similarity manipulation interaction for each of the three ratings (see F values in Table 2). Therefore, follow-up simple effects analyses within each condition were conducted, which address Research Question 1, the question of whether change in participants’ reactions from before to after the receipt of the bogus information differs as a function of whether they were in the similarity condition or the dissimilarity condition.

Table 2.

Change in Reactions to Partner from before to after the Presentation of the Bogus Similarity or Dissimilarity Information

	Similarity condition(n = 75)^a		Dissimilarity condition(n = 73)		F Pre vs. post	F Similarity manipulation	F interaction
	Before bogus info	After bogus info	Before bogus info	After bogus info	F Pre vs. post	F Similarity manipulation	F interaction
Expectation of liking the other	4.72 (1.01)	5.29 (1.04)	4.49 (.78)	4.42 (.80)	13.63***	16.69***	22.03***
					η² = .085	η² = .103	η² = .131
Expectation of enjoying the interaction	4.40 (1.09)	5.04 (1.07)	4.32 (.94)	4.33 (.96)	25.57***	6.65*	23.47***
					η² = .149	η² = .044	η² = .138
Eagerness to interact	4.04 (1.36)	4.72 (1.18)	4.18 (.93)	4.22 (1.06)	17.00***	1.18	13.35***
					η² = .104	η² = .008	η² = .084

Two participants were missing from this analysis in the similarity condition because there were technological difficulties in their completion of the online survey containing the prebogus measures, although they were included in the postmanipulation analyses (Table 1).

p < .05, **p < .01, p < .001.

As the means for this analysis in Table 2 show, participants in the similarity condition experienced an increase in their expectation of liking the other, expectation of enjoying the interaction, and eagerness to interact from before to after the receipt of the (bogus) similarity information about the other. All of the differences were significant, as indicated by the simple effects comparisons (ts all >5.00, p < .001). Thus, support was found for the similarity-liking effect. However, for participants in the dissimilarity condition, the means before and after the presentation of the bogus (dissimilarity) information were similar and not significantly different (ts < 1.00, all p ns). Therefore, no change (i.e., repulsion) was experienced for those in the dissimilarity condition. See Figure 1 for changes more specifically in liking between the two conditions.

Discussion

Today, people often meet potential friends and romantic partners briefly in certain settings (a party, Starbucks, the supermarket, a Tinder-generated brief meeting) and then later learn more information about the other through social media. People’s first impression of another formed after a brief greeting may include a general forecast of how much they think they would enjoy further interaction and like the person, but this initial impression is likely to change with the acquisition of additional information. For example, after an initial meeting, information can be obtained through social media about a new acquaintance’s interests, hobbies, and attitudes, and this additional information can lead to modified judgments of similarity or dissimilarity to the other. Following acquisition of (new) similarity information about another, do people enhance their liking (Byrne 1971)? Furthermore, does the acquisition of (new) dissimilarity information lead people to experience a decrease in their liking in comparison to their initial reaction? These were the questions addressed in this research.

This study was modeled after the old phantom-other similarity studies, although with a few important differences to address the issues raised previously. First, the “other” was not bogus but was a real participant, and participants knew they would be later interacting with the other. Second, the similarity manipulation was preceded by a greeting between the two real participants, which added ecological validity to the study. Third, measures of reactions (e.g., expectation of liking) of the dyad members to each other were administered after the initial greeting and then again after the presentation of the similarity manipulation, which allowed for a full pretest, posttest, control/experimental design (Carlsmith, Ellsworth, and Aronson 1976). This two-step design made it possible to compare changes in the participants’ reactions from before to after the receipt of the similarity or dissimilarity information—a stringent test of the relative validity of the similarity-liking hypothesis versus the repulsion hypothesis.

After the brief Skype greeting, the participants reported a relatively high degree of expectation of liking the other and enjoying the interaction. These reactions are likely influenced by the “thin slices” of nonverbal behavior (e.g., a wave, smile) that would have occurred in the brief Skype greeting (Ambady and Rosenthal 1993). In addition, just the mere exposure to another can lead to liking (Zajonc 1968). Then, after the presentation of the bogus (similarity or dissimilarity) information, and in support of the similarity effect found in many prior experimental studies (e.g., Byrne 1971), participants in the similarity condition reported more positive reactions than participants in the dissimilarity condition. However, this study used a more ecologically valid design (i.e., real participants) compared to the prior experiments. The most unique aspect of this study, however, was the ability to examine change in reactions to the other from before to after the receipt of the similarity/dissimilarity information. In the stringent two-step design of this study, evidence was found for the similarity effect, but no evidence was found for the repulsion hypothesis. For Rosenbaum’s (1986) hypothesis to be supported, the participants’ evaluation of the other would need to be more negative after the introduction of the dissimilarity information than before. Instead, dissimilarity information had no consequences—positive or negative. That is, the participants in the dissimilarity condition experienced no change in expectations of liking and enjoyment of interaction or desire to interact from before to after the receipt of the dissimilarity information. However, the similarity effect was supported by a comparison of the reactions from before to after the receipt of the bogus (similarity) information. Participants in the similarity condition experienced an increase in their positive evaluations of the other.

Why is similarity an important positive influence on initial reactions to a stranger? The literature abounds with explanations—some of which are only speculations and some supported by empirical evidence—of why similarity leads to attraction. These explanations include that similarity leads to liking because it increases positive affect and reinforcement (Byrne and Clore 1970), uncertainty reduction and predictability (Berger and Calabrese 1974), and expectations of enjoying interactions (Sprecher et al. 2013). Indeed, the present study had evidence for the latter explanation (i.e., participants in the similarity condition had a higher expectation of enjoying the interaction than those in the dissimilarity condition).

Although the results suggest that dissimilarity has no consequences in this study as evidenced by the comparison of changes in reactions from before to after the similarity manipulation, there can be another interpretation. Generally, more information leads to greater familiarity and liking in the stage of awareness (the first stage in Levinger and Snoek’s [1972] model of relatedness), although perhaps not in all situations (Finkel et al. 2015; Norton, Frost, and Ariely 2007, 2013; Reis et al. 2011). The context of dissimilarity may be one such situation in which greater information does not lead to liking. Recent research using a trait information paradigm has found that when participants are given a list of traits about a bogus other, those who receive a smaller number of randomly selected traits express more liking for the other than those who receive a greater number of traits (Norton et al. 2007, 2013). As noted by Finkel et al. (2015), however, another paradigm in social psychology—the mere exposure paradigm—has found that the more frequently people are exposed to a phenomenon (e.g., photos of unknown others, students sitting in a class; see Moreland and Beach 1992; Zajonc 1968), the more attraction is experienced. The present study’s design included aspects of both the mere exposure paradigm and the trait information paradigm. After the mere exposure provided through the Skype greeting, the participants reported an overall moderate degree of expectation of liking the other and enjoyment of the interaction. Then, in both conditions, the participants received trait information about the target. In the similarity condition, the trait information obtained was in the context of learning that the other was similar. In the dissimilarity condition, the trait information obtained was in the context of learning that the other was dissimilar. In the dissimilarity condition, the increase in familiarity through the trait information may have prohibited any enhancement of liking that generally occurs with the addition of new information. Thus, dissimilarity may have had consequences but in the form of preventing the familiarity effect from occurring in the awareness stage of relationship development. That is, greater information (i.e., familiarity) may operate when the information obtained about another indicates similarity but does not operate when the information indicates dissimilarity.

There were many strengths of this study, namely the two-step experimental design with pre-manipulation and post-manipulation measures of liking and other reactions, and the ecological validity provided by having a real (not bogus) other. There were also limitations, however, which give future investigators the opportunity to continue this line of research. First, there may be additional variations of the experimental design that could be implemented to further understand the role of acquisition of similarity versus dissimilarity information in the early acquaintance process. This could include varying the length of the initial greeting or manipulating whether a greeting even occurs. In addition, future research could manipulate the type of similarity information presented (i.e., the more mundane and noncontroversial attributes considered in this study vs. more impactful and controversial attitudes), or even whether similarity/dissimilarity information is presented at all (i.e., having a control condition with no trait information presented). Although the repulsion effect did not occur in the context in which it was examined in this study, it is possible that in other contexts (e.g., no greeting between participants, similarity/dissimilarity manipulated with attitudes and attributes that are especially salient or meaningful to the participants), a repulsion effect would be found.

Second, additional types of samples could be used. Most of the dyads in this study were female-female; many fewer were male-female, and none were male-male. Although there are no reasons to expect that the general findings (strong effect for the similarity effect and no support for the repulsion hypothesis) would differ in male-male dyads or in a study focused on romantic attraction conducted with participants matched for sexual orientation, these are empirical questions for future investigation. Furthermore, as is true of most laboratory experiments conducted in social psychology, college students were the participants. Further research could examine the influence of similarity and dissimilarity information on changes in attraction in older samples as the role of homophily may change over the life course (Thomas 2019).

The debate about whether similarity leads to attraction or dissimilarity leads to repulsion is an old one in social psychology but has new implications for the diverse ways in which people now develop acquaintances and learn information about them. Structural factors have historically brought people together with a baseline degree of homophily, but people are now meeting new acquaintances outside of geographical proximity and with information acquired about the other from diverse sources—including through both face-to-face encounters and digital forms in social media. This study suggests that the acquisition of similarity information about another in regard to noncontroversial traits and interests leads to an increase in liking from an initial baseline, whereas the acquisition of dissimilarity information on the same topics has no consequences—or possibly reduces the likelihood that a familiarity effect coming from the acquisition of new information leads to enhanced liking.

Footnotes

Acknowledgements

The author would like to acknowledge the following students who were on the author’s research teams during the semesters that this study was conducted: Ashley Campbell (lab co-director), Andrew Clesen (lab co-director), Jillien Ewoldt, Abrecia Hailaselise-Mankelklote, Adam Hampton (lab co-director), Caitlin Hannigan (lab co-director), Ryan Jacobsen, Madeline Luniewicz, Lauren Maass, Michelle Malik, Jalisa Nailon, Kyla Peal, Jennifer Sandoval, Brittany Weber, and Chaunte Wilder. A prior version of this research was presented at the Biennial meeting of the International Association for Relationship Research (July 2018, Fort Collins, Colorado).

Notes

ORCID iD

Susan Sprecher

Bio

Susan Sprecher is a Distinguished Professor of Sociology at Illinois State University. Her research, which has spanned over 35 years, has focused on a number of issues about close relationships, including attraction, relationship initiation, love, and sexuality. She has been actively involved in various positions in the International Association for Relationship Research and was the recent President.

References

Ambady

Nalini

Rosenthal

Robert

. 1993. “Half a Minute: Predicting Teacher Evaluations from Thin Slices of Nonverbal Behavior and Physical Attractiveness.” Journal of Personality and Social Psychology 64:431–41.

Antheunis

Marjolijn L.

Valkenburg

Patti M.

Peter

Jochen

. 2010. “Getting Acquainted through Social Network Sites: Testing a Model of Online Uncertainty Reduction and Social Attraction.” Computers in Human Behavior 26:100–09.

Baxter

Leslie A.

West

Lee

. 2003. “Couple Perceptions of Their Similarities and Differences: A Dialectical Perspective.” Journal of Social and Personal Relationships 20:491–514.

Berger

Charles R.

Calabrese

Richard J

. 1974. “Some Explorations in Initial Interaction and beyond: Toward a Developmental Theory of Interpersonal Communication.” Human Communication Research 1:99–112.

Berscheid

Ellen

Reis

Harry T

. 1998. “Attraction and Close Relationships.” Pp. 193–281 in The Handbook of Social Psychology, edited by Gilbert

D. T.

Fiske

S. T.

Lindzey

. New York, NY: Oxford University Press.

Burleson

Brant R.

Denton

Wayne H

. 1992. “A New Look at Similarity and Attraction in Marriage: Similarities in Social-Cognitive and Communication Skills as Predictors of Attraction and Satisfaction.” Communications Monographs 59:268–87.

Burleson

Brant R.

Kunkel

Adrianne W.

Birch

Jennifer D

. 1994. “Thoughts about Talk in Romantic Relationships: Similarity Makes for Attraction (and Happiness, Too).” Communication Quarterly 42:259–73.

Byrne

Donn

. 1961. “Interpersonal Attraction and Attitude Similarity.” The Journal of Abnormal and Social Psychology 62:713–15.

Byrne

Donn

. 1971. The Attraction Paradigm. New York, NY: Academic Press.

10.

Byrne

Donn

. 1997. “An Overview (and Underview) of Research and Theory within the Attraction Paradigm.” Journal of Social and Personal Relationships 14:417–31.

11.

Byrne

Donn E.

Clore

Gerald L

. 1970. “A Reinforcement Model of Evaluative Responses.” Personality: An International Journal 1:103–28.

12.

Carlsmith

J. Merrill

Ellsworth

Phoebe C.

Aronson

Elliot

. 1976. Methods of Research in Social Psychology. Reading, MA: Addison-Wesley Publishing Company.

13.

Clore

Gerald L.

Byrne

Donn

. 1974. “A Reinforcement-Affect Model of Attraction.” Pp. 143–70 in Foundations of Interpersonal Attraction, edited by Huston

T. L

. New York: Academic Press.

14.

Finkel

Eli J.

Baumeister

Roy F

. 2010. “Attraction and Rejection.” Pp. 419–59 in Advanced Social Psychology: The State of the Science, edited by Baumeister

R. F.

Finkel

E. J

. New York, NY: Oxford University Press.

15.

Finkel

Eli J.

Norton

Michael I.

Reis

Harry T.

Ariely

Dan

Caprariello

Peter A.

, et al. 2015. “When Does Familiarity Promote Versus Undermine Interpersonal Attraction? A Proposed Integrative Model from Erstwhile Adversaries.” Perspectives on Psychological Science 10:3–19.

16.

Hoyle

Rick H.

1993. “Interpersonal Attraction in the Absence of Explicit Attitudinal Information.” Social Cognition 11:309–20.

17.

Kenny

David A.

Kashy

Deborah A.

Cook

William

. 2006. Dyadic Data Analysis. New York, NY: Guildford Press.

18.

Lawrence

Barbara S.

Shah

N. P

. 2017. “Homophily: Measures and Meaning.” Working paper, University of California, Los Angeles.

19.

Layton

Bruce D.

Insko

Chester A

. 1974. “Anticipated Interaction and the Similarity-Attraction Effect.” Sociometry 37:149–62.

20.

Lazarsfeld

Paul

Merton

Robert

. 1954. “Friendship as Social Process”. Pp. 18-66 in Freedom and Control in Modern Society, edited by Berger

Morroe

Abel

Theodore

Page

Charles

. New York: Octagon.

21.

Levinger

George Klaus

Snoek

J. Diedrick

. 1972. Attraction in relationship: A New Look at Interpersonal Attraction. New York: General Learning Press.

22.

McCroskey

Linda L.

McCroskey

James C.

Richmond

Virginia P

. 2006. “Analysis and Improvement of the Measurement of Interpersonal Attraction and Homophily.” Communication Quarterly 54:1–31.

23.

McPherson

Miller

Smith-Lovin

Lynn

Cook

James M

. 2001. “Birds of a Feather: Homophily in Social Networks.” Annual Review of Sociology 27:415–44.

24.

Matthew

Montoya R.

Horton

Robert S

. 2014. “A Two-Dimensional Model for the Study of Interpersonal Attraction.” Personality and Social Psychology Review 18:59–86.

25.

Montoya

R. Matthew

Horton

Robert S.

Kirchner

Jeffrey

. 2008. “Is Actual Similarity Necessary for Attraction? A Meta-Analysis of Actual and Perceived Similarity.” Journal of Social and Personal Relationships 25:889–922.

26.

Moreland

Richard L.

Beach

Scott R

. 1992. “Exposure Effects in the Classroom: The Development of Affinity Among Students.” Journal of Experimental Social Psychology 28:255–76.

27.

Newcomb

Theodore M.

1961. The Acquaintance Process. New York: Holt, Rinehart, and Winston.

28.

Norton

Michael I.

Frost

Jeana H.

Ariely

Dan

. 2007. “Less Is More: The Lure of Ambiguity, or Why Familiarity Breeds Contempt.” Journal of Personality and Social Psychology 92:97–105.

29.

Norton

Michael I.

Frost

Jeana H.

Ariely

Dan

. 2013. “Less Is Often More, but Not Always: Additional Evidence that Familiarity Breeds Contempt and a Call for Future Research.” Journal of Personality and Social Psychology 105:921–23.

30.

Pilkington

Neil W.

Lydon

John E

. 1997. “The Relative Effect of Attitude Similarity and Attitude Dissimilarity on Interpersonal Attraction: Investigating the Moderating Roles of Prejudice and Group Membership.” Personality and Social Psychology Bulletin 23:107–22.

31.

Reis

Harry T.

Maniaci

Michael R.

Caprariello

Peter A.

Eastwick

Paul W.

Finkel

Eli J

. 2011. “Familiarity Does Indeed Promote Attraction in Live Interaction.” Journal of Personality and Social Psychology 101:557–70.

32.

Rosenbaum

Milton E.

1986. “The Repulsion Hypothesis: On the Nondevelopment of Relationships.” Journal of Personality and Social Psychology 51:1156–66.

33.

Singh

Ramadhar

Yan Ho

Soo

. 2000. “Attitudes and Attraction: A New Test of the Attraction, Repulsion and Similarity-Dissimilarity Asymmetry Hypotheses.” British Journal of Social Psychology 39:197–211.

34.

Singh

Ramadhar

Tan

Lynne SC

. 1992. “Attitudes and Attraction: A Test of the Similarity-Attraction and Dissimilarity-Repulsion Hypotheses.” British Journal of Social Psychology 31:227–38.

35.

Smeaton

George

Byrne

Donn

Murnen

Sarah K

. 1989. “The Repulsion Hypothesis Revisited: Similarity Irrelevance or Dissimilarity Bias?” Journal of Personality and Social Psychology 56:54–59.

36.

Sprecher

Susan

. 1998. “Insiders’ Perspectives on Reasons for Attraction to a Close Other.” Social Psychology Quarterly 61:287–300.

37.

Sprecher

Susan

. 2014. “Effects of Actual (Manipulated) and Perceived Similarity on Liking in Get-acquainted Interactions: The Role of Communication.” Communication Monographs 81: 4-27.

38.

Sprecher

Susan

Treger

Stanislav

Hilaire

Nicole

Fisher

Amanda

Hatfield

Elaine

. 2013. “You Validate Me, You Like Me, You’re Fun, You Expand Me: ‘I’m Yours!.’” Current Research in Social Psychology 22:22–34.

39.

Sunnafrank

Michael

. 1986. “Predicted Outcome Value during Initial Interactions: A Reformulation of Uncertainty Reduction Theory.” Human Communication Research 13:3–33.

40.

Tan

Diana Tze Yeong

Singh

Ramadhar

. 1995. “Attitudes and Attraction: A Developmental Study of the Similarity-Attraction and Dissimilarity-Repulsion Hypotheses.” Personality and Social Psychology Bulletin 21:975–86.

41.

Thomas

Reuben J.

2019. “Sources of Friendship and Structurally Induced Homophily across the Life Course.” Sociological Perspectives. doi/10.1177/0731121419828399

42.

Werner

Carol

Parmelee

Pat

. 1979. “Similarity of Activity Preferences among Friends: Those Who Play Together Stay Together.” Social Psychology Quarterly 42:62–66.

43.

Yang

Chia-chen

Brown

B. Bradford

Braun

Michael T

. 2014. “From Facebook to Cell Calls: Layers of Electronic Intimacy in College Students’ Interpersonal Relationships.” New Media & Society 16:5–23.

44.

Zajonc

Robert B.

1968. “Attitudinal Effects of Mere Exposure.” Journal of Personality and Social Psychology 9:1–27.