For example, several years ago, I asked university students to indicate the degree to which they desired various characteristics in a relational partner (the type of relationship that they were asked to consider was manipulated) (Sprecher & Regan, 2002)
Although not generally referred to as studies on similarity, mate selection studies (in which participants are asked how much they desire various traits in a partner) have, in some cases, included items that refer to similarity. Included in the list, in addition to traits such as physical attractiveness, ambition, warmth and kindness, were four types of similarity: similarity on background characteristics (e.g., race, religion, social class), similarity on attitudes and values, similarity on social skills (e.g., interaction styles), and similarity on interests and leisure activities. Participants expressed preferences for all four types of similarity, as indicated by mean scores to the items that were above the midpoint of the response scales. The order in which the other types of similarity were rated in importance was: similarity in interests, similarity in social skills (interaction styles), and similarity in background characteristics. Although similarity was generally desired across all types of relationships, it was preferred to a greater degree in a marital partner, particularly as compared to in a friend.
Similar results were found in an earlier partner preference study that I and my colleagues conducted (Regan, Levin, Sprecher, Christopher, & Cate, 2000). The participants rated the same four types of similarity to be moderately important in a partner. In addition, similarity in attitudes and values was rated more important than similarity in interests and leisure activities, which was judged to be more important than similarity in social skills (interaction styles) and similarity in background characteristics. This study also demonstrated that a preference for similarity was greater in a long-term romantic partner than in a short-term sexual fling.
In some mate selection studies, participants’ own characteristics are assessed in addition to their preferences for the same characteristics in a partner. For example, Dijkstra and Barelds (2008) had their participants complete measures of the Big Five Personality characteristics (openness, agreeableness, conscientiousness, extraversion, neuroticism) and then indicate how much they would desire the same personal characteristics in a potential mate. Strong correlations were found between the individuals’ own personality characteristics and the degree to which they desired the personality characteristics in a mate.
In their design, they manipulated not only the level of similarity of the other but also the expectation that a relationship could develop
In a second type of research, the bogus stranger paradigm (e.g., Byrne, 1971), participants respond to a hypothetical or phantom other, about whom information is manipulated so that the hypothetical other varies in the level of similarity to the participant. Similarity research using this paradigm has led to the “law of attraction” (Byrne & Rhamey, 1965), which describes a positive linear association between the degree of similarity (e.g., attitudinal similarity) and attraction for another. Although the use of the bogus stranger paradigm to examine the similarity effect is less likely to appear in recent literature (the focus has shifted to the study of ongoing relationships), Aron, Steele, Kashdan, and Perez (2006) used this method to examine the effect of similarity of interests on initial attraction to a same-gender other. Based on self-expansion theory (Aron & Aron, 1986), they predicted that when participants are not led to believe that a relationship was certain, the similarity effect will occur, but when there is certainty of a relationship, the effect of similarity may be reduced or even disappear (because a dissimilar other can be desirable for the self-expansion opportunities offered). The findings supported the predictions, particularly for men. The lack of effect found for women was explained as due to the lesser relevance of activities to the friendships of women.