Correlates of Couples ’ Perceived Similarity at the Initiation Stage and Currently

Although prior research (Montoya, Horton, & Kirchner, 2008) has indicated that perceived similarity is more important than actual similarity as a predictor of satisfaction and attraction, there is a lack of research on factors associated with couples’ perceptions of similarity in their relationship. In the present study, a sample of couples (both partners) provided ratings of the degree to which they perceived similarity in six areas (background characteristics, attitudes and beliefs, leisure pursuits and interests, communication style, personality, and physical attributes) for two stages in their relationship: currently and at the initiation stage (viewed retrospectively). The couples perceived greater similarity for the current stage of their relationship than for the beginning stage of their relationship. Factors found to be associated with perceived similarity included positive social network reactions, overlap in social networks (predictor of current perceived similarity only), perceived compatibility, and satisfaction and commitment (examined only for current perceived similarity). Of the various types of perceived similarity considered in this study, similarity in attitudes and beliefs was most consistently associated with the various predictor variables. Implications of these results for enhancing couple relationships are discussed.

Interpersona | An International Journal on Personal Relationships interpersona.psychopen.eu | 1981-6472 A distinction has been made between actual similarity (the degree to which two people are actually similar) and perceived similarity (the degree to which similarity is perceived with the other).Some researchers have argued that perceived similarity is much more important than actual similarity in generating attraction and relationship satisfaction (e.g., Condon & Crano, 1988;Duck & Barnes, 1992;Hoyle, 1993;Klohnen & Luo, 2003).Although there have been decades of research on similarity, and recent meta-analysis evidence of the greater importance of perceived than actual similarity in ongoing relationships (Montoya, Horton, & Kirchner, 2008, 2008), we know little about couples' perceptions of various types of similarity in their relationship, and the factors associated with the degree of perceived similarity.The present study, with a sample of romantic couples, focused on partners' perceptions of similarity with each other in several areas in regard to two times in their relationship: at the very beginning stage and currently.

Evidence that Perceived Similarity is More Important Than Actual Similarity
Research beginning years ago (Newcomb & Svehla, 1937) and continuing more recently (e.g., Luo & Klohnen, 2005;Watson et al., 2004) has indicated that existing spouse and friendship pairs are similar, and more similar than random pairs of individuals.Similarity in existing pairs, also referred to as homogamy, can occur for several reasons including that: (a) People are more likely to meet others who are similar based on where they live and work and the activities they engage in; (b) People may prefer similar others and therefore seek them over other possible partners within their environment; and (c) Partners may become more similar over time, as they spend time together (referred to as convergence).Researchers often assess actual similarity by comparing partners' scores on particular dimensions (e.g., attitudes, personality); smaller differences are an indication of greater similarity (Montoya et al., 2008).In addition, a profile approach has been used in some recent research (Luo & Klohnen, 2005), in which differences between partners can be examined for multiple dimensions at once.Perceived similarity has been assessed in multiple ways, too, including asking people how similar they are to their partner overall or in specific areas, or asking participants to complete a measure (e.g., personality) for self and also for the partner, with a smaller difference representing greater perceived similarity (Tidwell, Eastwick, & Finkel, 2013).
Partners are more similar than would occur if people paired up randomly, but partners tend to perceive that they are more similar than they actually are (Buunk & Bosman, 1986;Byrne & Blaylock, 1963).
Although there is strong evidence of the existence of similarity in actual couples, more equivocal is the effect of the degree of actual similarity (within the typical range found in romantic pairs) on relationship satisfaction and other indicators of relationship success.As noted above, the similarity-attraction hypothesis suggests that greater attraction and satisfaction will result from greater actual similarity.The evidence for this similarity-satisfaction hypothesis is mixed, however.Recently, Finkel, Eastwick, Karney, Reis, and Sprecher (2012), in a critical analysis of the emphasis on similarity matching at Internet matching services (e.g., eHarmony), concluded the following about the literature on similarity: "On one hand, most people do end up paired with partners who strongly resemble themselves, at least in terms of demography and attitudes.On the other hand, once paired, it remains unclear whether the degree of similarity within the couple is associated with more versus less successful relationships over time" (p.46).Finkel et al. (2012) cited that led to this conclusion was a meta-analysis of 313 laboratory and field studies on the effect of similarity (of attitudes and personality) on attraction and satisfaction (Montoya et al., 2008).In their meta-analysis, Montoya et al. considered three types of studies: no interaction studies (bogus stranger), brief-interaction studies (in which previously unacquainted individuals interacted), and studies of existing relationships.Montoya et al. (2008) found that actual similarity had a strong effect in laboratory studies that involved no interaction (bogus stranger paradigm), a moderate effect when there was a brief interaction in previously unacquainted pairs, but no effect in existing relationships.The effects of perceived similarity on attraction were strong in all three types of studies.Therefore, in actual relationships, the effect of perceived similarity on satisfaction is stronger than the effect of actual similarity.Example studies that have shown that perceived similarity is more strongly associated than actual similarity with relationship quality in existing couples include Acitelli, Douvan, and Veroff (1993), Hendrick (1981), Jones and Stanton (1988), and Levinger and Breedlove (1966).

One of the studies
The strong and consistent effects of perceived similarity on attraction (and satisfaction) may occur because both causal directions are operating.That is, not only is it likely that perceived similarity contributes to attraction (and satisfaction), but feelings of attraction and satisfaction may in turn lead to the perception of similarity.Two of the explanations referred to above for the effect of similarity on attraction (i.e., similarity is consensually validating, similarity leads to uncertainty reduction and predictability) require only that there is perceived similarity.Morry (2005) proposed the attraction-similarity model, which argues that in existing relationships, partners' satisfaction leads to perceived similarity.Morry's attraction-similarity model uses balance theory (Heider, 1958) to argue that people in relationships are motivated to perceive similarity in order to achieve balance.Her model is also based on the principles that: (a) People have "lay theories" that include beliefs that relationship partners are similar, and, more generally, that similarity and attraction go together; and (b) people project their own attitudes and traits onto others and especially those they like.
How Might Perceived Similarity Change Over Time?Morry's (2005) attraction-similarity hypothesis would suggest that perceived similarity increases over time in a relationship.As partners become closer, they are likely to increase in their motivation to perceive similarity and therefore engage in more projection.In addition, because perceived similarity is likely to be linked to actual similarity, if actual similarity changes over the course of a relationship, perceived similarity should change in a corresponding way.There are several reasons that actual similarity is likely to increase over time.Partners spend more time together as their relationship increases in duration, and therefore they are likely to develop more similar attitudes over time, which has been referred to as attitude alignment (Davis & Rusbult, 2001).It is more balanced to agree than to disagree with a close other (Heider, 1958;Newcomb, 1961).Furthermore, as pairs become interdependent, engage in the same activities, and live in the same environment, they are likely to develop similar interests (Gonzaga, 2009).Couples can even begin to look alike because of being influenced by each other's diets and exercise patterns (Christakis & Fowler, 2007).Overall, I expect to find that the participants will perceive more similarity with their partner currently than retrospectively when the relationship began.

Correlates of Perceived Similarity
Although I predict that relationship partners will perceive similarity in their relationship on several dimensions (e.g., attitudes and beliefs, communication styles), and will perceive more similarity currently than for the period when their relationship began, there is still likely to be variation among couples in perceived similarity, at both periods of the relationship.This variation in perceived similarity may be associated with several factors.Below, I discuss several such factors.

Relationship Satisfaction and Perceived Similarity
One factor that has been found to be associated with perceived similarity, as discussed above, is relationship satisfaction.Although the classic similarity-attraction hypothesis suggests that similarity leads to attraction and satisfaction, equal or stronger evidence suggests that attraction/satisfaction leads to perceived similarity (Morry, 2005(Morry, , 2007;;Morry, Kito, & Ortiz, 2011).Those who are more satisfied in their relationship perceive greater similarity, and conversely, dissatisfaction can lead to perceptions of dissimilarity.Therefore, I hypothesize that greater satisfaction (and commitment) are associated with greater perceived similarity.However, because not all types of perceived similarity are likely to be equally important in contributing to relationship satisfaction (Lutz-Zois, Bradley, Mihalik, & Moorman-Eavers, 2006), I will also explore whether the association between perceived similarity and relationship satisfaction depends on the type of similarity.

Compatibility and Perceived Similarity
A related factor considered in this study is perceived compatibility of the relationship.Relationship compatibility has become popular concept in the media due to the claim by many Internet dating sites that they can find "compatible matches" for single people (Sprecher, 2011).Although compatibility and compatible matches are not central concepts in relationship science, when "compatibility" and "compatible matches" are discussed in the literature, they are often associated with the similarity principle (see discussion by Sprecher, 2011).That is, it is often assumed that similar partners make compatible pairs.In addition, stage models of relationship development (Cate & Lloyd, 1992) suggest that similarity increases partner compatibility, which then contributes to attraction.Therefore, I hypothesize that perceptions of similarity are associated positively with perceptions of being a compatible match.

Relationship Onset (Being Friends First) and Perceived Similarity
Not all romantic relationships begin in the same way.Some relationships emerge out of a friendship, whereas others begin because the two individuals had instant attraction soon after meeting.How might degree of perceived similarity be associated with the type of onset of the relationship?On the one hand, partners who were friends first may have had similar backgrounds and opportunities to develop many other types of similarities (e.g., leisure pursuits) even prior to forming a romantic relationship.In fact, the romantic relationship may have developed out of friendship because of similarities.Romantic partners who were friends first may also have had more time to discover their existing similarities.On the other hand, the intense rush of "falling in love" can lead to positive illusions about one's partner, and the illusions can include perceptions of similarity with the partner.Furthermore, the belief that one has found one's soul mate (i.e., someone very similar) may lead to the onset of rapid attraction.
To my knowledge, only one study has considered the issue of how type of relationship onset is associated with similarity.Barelds and Barelds-Dijkstra (2007) predicted that those who fell in love rapidly (vs.those who were friends first) would be less likely to have a partner who was similar.They argued that this occurs because "friends first" romantic pairs have time to get to know each other's personality, which enables them to "not only select a mate who they perceive to be similar to themselves but also one who is similar to themselves, increasing the chances of partner personality homogamy" (p.483).With a sample of 137 married and cohabiting couples, Barelds and Barelds-Dijkstra (2007) compared three types of couples: (a) those who said they were friends before being lovers; (b) those who experienced love at first sight; and (c) those who were intermediate (had known each other for a brief period before falling in love).In support of their prediction, they found that partners who fell in love first had less similar personalities (relative to those who were friends first), particularly in regard to extroversion, emotional stability, and autonomy (among the Big Five Personality traits).However, in discussing why the "lovers first group" did not differ from the "friends first" group in relationship quality, Barelds and Barelds-Dijkstra stated that the lovers first group may perceive themselves to be similar and concluded, "relationship success may be more strongly related to spouses' perceptions of similarity than to their actual degree of similarity."(p.491).Barelds and Barelds-Dijkstra, however, did not measure perceived similarity in their study.Based on their research and Interpersona  2013, Vol. 7(2), 180-195  doi:10.5964/ijpr.v7i2.126 the above arguments, I hypothesize that romantic dyads who were friends first will perceive greater similarity retrospectively for when they entered their relationship than dyads who were not friends first.

Social Networks and Perceived Similarity
Romantic relationships do not exist in a vacuum; they are embedded in a social network of friends, family members, and acquaintances (Parks, 2006;Sinclair & Wright, 2009).Although there are various aspects of social networks that have been studied for their influence on couples' relationships, the two most frequently studied variables are social network reactions directed to the relationship (i.e., to what degree do network members support the relationship and perceive the two to be a good pair?) and degree of social network overlap (to what degree do couples have friends and other network members in common?) (Sprecher, Felmlee, Orbuch, & Willetts, 2002).Research has found that relationships with greater social network overlap and relationships that receive more positive reactions from the social network have greater satisfaction, love, and commitment (for a review, see Sprecher et al., 2002).
Social network members also have been found to be good forecasters of whether a relationship will last (Agnew, Loving, & Drigotas, 2001).
Just as members of couples are likely to be aware of lay theories that similarity and relationship satisfaction should be related (Morry, 2005), people (as network members) also should be aware of these lay theories and have corresponding beliefs about the link between similarity and the relationship success of relationships in their network.
Therefore, social network members would likely judge pairs who they perceive to be highly similar to be good matches and would also likely communicate these positive reactions to the couples.In addition, those who share friends in common are likely to be or become similar in other ways as well (e.g., background characteristics, leisure pursuits).The social network literature, therefore, leads to the prediction that the perceptions of social network support and social network overlap will be associated positively with participants' greater perceived similarity.

Purposes to the Present Study
In sum, the purpose of this study was to examine the degree to which couples perceived themselves to be similar across several dimensions, and for two stages of their relationship: retrospectively for the beginning stage of their relationship and for their current stage.A second purpose was to examine relationship correlates of perceived similarity at the two stages.The hypotheses and research questions are: H1: Couples will perceive themselves to be similar across several dimensions (background characteristics, attitudes and beliefs, leisure interests and activities, communication style, personality, physical attributes), at both the initial stage of their relationship and currently.
RQ1: Of the different areas of similarity, which types will be perceived to be greater?H2: Couples will perceive greater similarity at the current stage of the relationship than at the initial stage.Perceived Similarity in Couples Finally, I raise two research questions that can be addressed with data collected from both partners: RQ2: Do partners agree in their perceptions of similarity?RQ3: Do male and female partners differ in their degree of perceived similarity?

Method Sample
The sample consisted of a volunteer sample of 100 romantic couples (200 participants) obtained from a Midwest U.S. university; this was after eliminating one couple that was an outlier on age.Ninety-six were male-female relationships and four were lesbian relationships.The mean age of the participants was 22.59 (SD = 4.04).A majority of the participants (66%) were in a seriously dating relationship.Six percent were in a casual dating relationship, 14% were in a cohabiting relationship, 7% were engaged, and 7% were married.(In some couples, partners did not agree on the stage of their relationship.These couples were assigned the more advanced stage of the two responses.)

Procedure
Couples were recruited through announcements made in classes and signs placed around campus.The recruitment requested couples at any stage of a relationship, including marriage.A time was set up for couples to come to a university setting to complete a questionnaire.Couples were given a gift certificate for their participation.The partners were separated in a large computer classroom, where each partner completed an on-line survey.Typically, multiple couples arrived at the same time to complete the online survey.Physical attributes.These are types of similarity identified in prior literature (e.g., Baxter & West, 2003).Each similarity item was followed by a 7-point response scale anchored with 1 = not at all similar, 4 = moderately similar; and 7 = extremely similar.

Measures
• Degree of Friendship Prior to Romantic Relationship.Participants were presented with the following item: "Some relationships grow out of a long friendship.Others are romantic from the first meeting.How would you characterize your relationship with your partner?"The options presented were 1 = From the very beginning, we defined the relationship as romantic/dating; 2 = We first had a brief friendship and then became romantic partners; and 3 = We first had a long and strong friendship and then became romantic partners.
(Although this item was not modeled after the relationship onset measure used by Barelds and Barelds-Dijkstra [2007], it was similar to theirs.)• Network Overlap.Participants were asked, "Prior to forming a relationship with your partner, to what degree did you have overlapping social networks (e.g., friends and acquaintances in common)?A 7-point response scale followed, anchored by 1 = no overlap, 4 = moderate; and 7 = a great deal over overlap.
• Network Reaction to the Relationship.Participants were asked, "At the time you were first becoming acquainted, to what degree do you think others who were important to you viewed you as a "well-matched" couple, perfect for each other?"Options ranged from 1 = did not view us as well-matched; to 7 = viewed us as very well-matched.
• Perceived Compatibility.Participants were asked, "When you were first becoming acquainted, how compatible did you think the two of you were for a relationship?"Response options ranged from 1 = not very compatible; to 7 = very compatible.
Measures About the Current Stage of the Relationship.The second part of the survey asked the participants about the current stage of their relationship.

• Perceived Similarity.
Participants were asked about their perceptions of their similarity with their partner on the same six dimensions referred to above, although for the current stage of their relationship.
• Network Overlap.Participants were asked, "At this time, to what degree do you have overlapping social networks (e.g., friends and acquaintances in common)?"A 7-point response scale followed, anchored by 1 = no overlap, 4 = moderate; and 7 = a great deal over overlap.
• Network Reaction to the Relationship.Participants were asked, "At this time, to what degree do you think others who are important to you view you as a "well-matched" couple, perfect for each other?"Options ranged from 1 = do not view us as well-matched; to 7 = view us as very well-matched.
• Perceived Compatibility.Participants were asked, "How compatible do you think the two of you are for a long-term relationship?"Response options ranged from 1 = not very compatible; to 7 = very compatible.
• Relationship Satisfaction.Participants completed the Hendrick (1988) 7-item Relationship assessment scale.Example items include, "How good is your relationship compared to most?" and "In general, how satisfied are you with your relationship?"Although the specific anchors varied across the items, the higher score indicated greater satisfaction.Cronbach's alpha was .86.
• Relationship Commitment.Participants completed the Rusbult, Martz, and Agnew (1998) 7-item commitment scale.Example items include, "How much do you want your relationship to last a very long time?" and "How likely is it that you would date someone other than your partner within the next year?"(reverse scored).Each item was followed by a 1 = not at all; to 7 = very much response scale.Cronbach's alpha was .94.

Overview to Analyses
Because the data were dyadic (i.e., data from both members of the pair), mean couple scores were created for each variable, and these couple mean scores were used in the analyses below.Although it is recognized that there could be other ways of analyzing the data -including separately for each partner and through the Actor-Partner Interdependence Model (e.g., Kashy & Snyder, 1995), the couple means scores approach was used because the focus was not on partner versus actor effects.When reporting on the degree of similarity between themselves and their partner retrospectively for the period when they first became acquainted, the participants perceived, on average, above moderate similarity.Column 1 of Table 1 presents the couple mean scores for the six types of perceived similarity at the initiation stage of the relationship.The means were all above the mid-point of the scale, ranging from 4.88 to 5.23 (on a 7-point scale).
Single-sample t-tests indicated that each mean was significantly greater than the mid-point of the response scale (i.e., moderate similarity).Therefore, in regard to the initiation stage, support was found for H1, which predicted that the couples would perceive themselves to be similar across several dimensions.
To address RQ1 (which asked which areas were perceived to have greater similarity), a repeated measures ANOVA was conducted to compare the scores on the six types of perceived similarity.A significant effect was found (F[5, 495] = 2.87, p = .019,partial η 2 = .131).The types of perceived similarity that received the highest scores were leisure activities, attitudes and beliefs, and personality.Similarity on background characteristics received the lowest score, although the score was still above the midpoint.

Degree of Current Perceived Similarity
Column 2 of Table 1 presents the couple mean scores of perceived similarity assessed concurrently.Once again, all of the means were significantly above the midpoint of the responses, as indicated by single samples t-tests, in support of H1, which predicted that the couples would perceived similarity.A repeated measures ANOVA indicated an overall difference in scores based on type of similarity (F[5,495] = 3.36, p = .005,partial η 2 = .033).Once again, higher scores were found for similarity on leisure activities and interests, similarity on attitudes and beliefs, and similarity on personality; the lowest scores (although still above the midpoint) were for similarity on background characteristics and similarity on communication style.
H2 predicted that couples would perceive greater similarity at the current stage of the relationship than for the initiation stage of the relationship.In support of this hypothesis, the scores on each dimension of similarity were significantly higher for the current period in the relationship than as recalled for the initial stage of the relationship, as indicated by paired t-tests (see Table 1).

Correlates of Perceived Similarity at the Initial Stage of the Relationship
H3 predicted that perceptions of similarity at the beginning stage of the relationship would be associated positively with being friends before having a romantic relationship, having overlapping social networks prior to the relationship, perceiving positive social network reactions, and perceiving that one's relationship was compatible at the time.
Table 2 presents the correlations of perceived similarity on the six dimensions at the get-acquainted stage of the relationship with scores on these attributes of the relationship also assessed in regard to that stage of the relationship.Neither degree of friendship prior to romance nor overlap in social network prior to meeting was associated with any type of perceived similarity.Participants' perception of positive support from the social network, however, was associated positively with each type of perceived similarity, with the strongest association found for perceived similarity on attitudes and beliefs.In addition, participants' perception (at the time that the relationship began) that their relationship had the potential to be compatible was associated with greater perceived similarity on each of the six dimensions, with the highest correlation found for perceived similarity on attitudes and beliefs, and the lowest correlations (but still significant) found for perceived similarity on background characteristics and perceived similarity on physical attributes.Therefore, partial support was found for the predictions of H3.

Correlates of Perceived Similarity at the Current Relationship Stage
H4 predicted that perceptions of similarity currently would be associated with being friends first, current overlapping social networks, positive reaction from the network for their relationship, perceptions that the relationship was compatible, satisfaction, and commitment.Degree of friendship at the time of relationship initiation was not associated with any type of perceived similarity (rs ranged from -.03 to .15).Table 3 presents the correlations of the six types of perceived similarity at the current relationship stage with the other characteristics of the current relationship, assessed in regard to the current stage.Although perceived network overlap recalled for the initial stage was not associated with perceived similarity at that stage (as reported above), current perceived network overlap was associated positively with current perceptions of three types of similarity: similarity in background characteristics, similarity in attitudes and values, and similarity in leisure pursuits and interest.Furthermore, participants' perceptions that their social network perceived them as a well-matched couple (i.e., a positive social network reaction) was associated positively with each type of perceived similarity, and was most highly associated with similarity in attitudes and beliefs.

Perceived Similarity in Couples
Participants' reports of indicators of the quality of the current relationship were also associated with perceived similarity.First, participants' perception of the current compatibility of their relationship was associated with perceived similarity for five of the areas; the exception was for similarity on physical attributes.The associations were highest for similarity in attitudes and beliefs and similarity in communication styles.Scores on the satisfaction scale were correlated positively with all six types of perceived similarity, with the strongest correlations found for similarity in attitudes and beliefs and similarity in communication styles.Scores on commitment were associated positively with perceived similarity on attitudes and beliefs (strongest correlation) and also with perceived similarity on three other dimensions: communication style, personality, background characteristics.

Two Other Issues
Two additional research questions were examined, made possible by data collected from both partners: RQ2: Were the relationship partners similar in their perceptions of similarity?; RQ3: Are there gender differences in the degree to which similarity was perceived?
Relationship partners' perceptions of similarity were positively correlated, for each type of similarity, both at the beginning stage and the current stage of the relationship.The correlations ranged from r = .21(communication style) to r = .52(background characteristics) for the beginning stage, all significant.For the current period, the correlations ranged from r = .31(for both personality and physical attributes) to r = .65(for background characteristics), all significant.
To examine gender differences in perceived similarity, I included only the male-female relationships (eliminating the four lesbian couples), and compared the male partners' scores with the female partners' scores on each type of similarity, via paired t-tests.No differences were found between the male partners and the female partners on their scores on the six types of perceived similarity for the initiation stage of the relationship.For the current stage of the relationship, female partners perceived more similarity on attitudes and values (t[95] = 2.70, p = .008)than male partners.Overall, though, male and female partners were very similar in their perceptions of similarity.

Discussion
The role of similarity in relationships has been investigated for decades (e.g, Montoya et al., 2008;Newcomb, 1961).This study was conducted to examine couples' perceptions of various type of similarity at two points in their relationship, and examine factors associated with the couples' perceived similarity.Consistent with prior research (e.g., Baxter & West, 2003;Byrne & Blaylock, 1963;Montoya et al., 2008) and Hypothesis 1, couples in the present study perceived themselves to be similar across several dimensions -attitudes and beliefs, personality, communication style, background characteristics, physical attributes, and leisure interests -both at the current stage and initially in the relationship.Furthermore, couples perceived greater similarity on each of the dimensions in regard to the current stage of their relationship than for the stage when they first entered the relationship (viewed retrospectively), in support of Hypothesis 2. Whether the couples actually became more similar over the course of the relationship could not be determined in this study.However, it is likely that the higher perceived similarity at the current stage than for the initiation stage occurred due to a combination of convergence (partners becoming more similar over time) and the increasing tendency to project own traits and interests onto a partner as a relationship develops (e.g., Morry, 2005).Furthermore, as they were completing the survey, the participants may have been influenced by lay theories that relational partners should be similar, and presumably more similar than when they first met (e.g., Morry, 2005).Not all types of similarity were perceived to the same degree, however.The couples perceived greater similarity for leisure interests/activities, attitudes and beliefs, and personality than for the other types of similarity (e.g., background characteristics), at both times of the relationship.The type of similarity that had the lowest score (although still above the midpoint) was for background characteristics.The differential degree to which types of similarity were perceived could be due to actual differences in similarity across these dimensions, but also could have occurred to the degree that it is easier to project certain attributes (activity interests, attitudes) than others (e.g., background characteristics, physical attributes) onto a partner.This study also examined correlates of perceived similarity, both characteristics that are internal to the relationship (e.g., satisfaction, perceived compatibility) and characteristics that are external to the relationship (e.g., social network overlap).Of the factors considered in this study in regard to the beginning stage of the relationship, the assessment that the social network perceived the partners as well-matched (i.e., positive social network reaction) and the assessment of being a compatible match were the two variables found to be associated (positively) with perceptions of similarity.Degree of friendship prior to the relationship and overlap in social network prior to the relationship were not associated with similarity recalled for the initial stage.Thus, partial support was found for Hypothesis 3.
For the current stage, all of the factors measured were found to be associated with perceived similarity, in support of Hypothesis 4.More specifically, perceived overlap in social network, perceived network reaction, perceiving one's relationship as compatible, satisfaction, and commitment were associated with concurrent perceived similarity.I had also examined whether the type of relationship onset (i.e., being friends first, measured for the initiation stage) was associated with perceived similarity at the current stage, and it was not.Not surprising in these results is the high association between perceived similarity and measures of relationship quality, both at the beginning stage (i.e., perception of compatibility) and at the current stage (i.e., perception of compatibility, satisfaction, commitment).These positive associations between perceived similarity and relationship quality would be predicted by both the similarity-attraction hypothesis (e.g., Byrne, 1971) and the attraction-similarity model (Morry, 2005).That is, not only is it likely that similarity leads to satisfaction, but also that satisfaction leads to perceptions of similarity.More surprising are some of the null associations found.Although there would

Conclusions
The results of this study showed a strong link between perceived similarity and variables that reflect the quality of the relationship -satisfaction, commitment, and perception of relationship compatibility.Regardless of the degree to which these associations can be attributed to either causal direction (similarity leading to satisfaction vs. satisfaction leading to perceived similarity), the results suggest that relationship enhancement programs could focus on relationship partners' attention to and development of their similarities, which likely enhance other positive outcomes in the relationship (e.g., Stanley, Amato, Johnson, & Markman, 2006).Indeed, the results have implications beyond the study of dyads also to intergroup relations, including reactions to outgroup members.If people can be encouraged to emphasize their similarities, along with increased contact with others, the result will likely be increased liking of outgroup members (Pettigrew & Tropp, 2006).In addition, the results particularly point to the importance of similarity in attitudes and beliefs.

Table 1
Perceptions of Similarity on Six Dimensions: At the Initiation Stage (Assessed Retrospectively) Versus Currently

t Currently Mean (SD) Beginning Stage Mean (SD) Type of Perceived Similarity
The above means represent couple scores, the average of two partners.Perceived similarity scores ranged from 1 = not at all similar; **p < .01. *** p < .001.

Table 2
Correlates of Perceived Similarity at the Initiation Stage of the Relationship

Perceived Compatibility Perceived Network Reaction to the Relationship Overlap in Social Network Prior to Meeting Degree of Friendship Prior to Romance Type of Perceived Similarity
Note.The above correlations are between couple scores, the average of two partners.

Table 3
Correlates of Perceived Similarity for the Current Stage of the RelationshipThe above correlations are between couple scores, the average of two partners.*p < .05. **p < .01. ***p < .001.