In the first section of this article, I discuss how relationship compatibility is presented in relationship science. The second section focuses more specifically on compatible matches, also from the perspective of relationship science. The final section discusses the compatibility matching procedures used at the Internet matching sites. Although the focus of Internet matching services is on compatibility in romantic relationships, compatibility is a concept relevant to all types of relationships, including friendships, parent-child relationships, and co-workerspatibility is also a relationship state that is good and desirable (Berscheid, 1985; Berscheid & Regan, 2005).
Regardless of whether compatibility is assessed directly or is assessed indirectly through its traveling companions, different observers and actors may vary in their perceptions of a relationship’s compatibility. For example, research by Chris Agnew and his colleagues suggests that social networks’ opinions of the compatibility of relationships may be more realistic, predictive, and negative than that of the insiders (Agnew, Loving, & Drigotas, 2001; Etcheverry, Le, & Charania, 2008; Loving, 2006; MacDonald & Ross, 1999).
The second category of causal conditions is the combination or intersection of P’s and O’s characteristics (including their personality characteristics, attitudes and values, interests), which can refer specifically to being a compatible match. This causal factor is a focus on static personal characteristics of both partners and represents the crux of what is being considered by Internet matching sites in their efforts to create matches based on relationship science. The three “principles” from the close relationships field that refer to the intersection of partners’ characteristics are: (1) similarity (or “birds of a feather flock together”), (2) complementary (or “opposites attract”); and 3) matching on socially desirable characteristics. As will be discussed further in the next section, considerable research indicates that similarity contributes to compatibility. In fact, some writers have equated similarity with compatibility. For example, Houts et al. (1996) wrote, “the standard paradigm for studying the role of compatibility in courtship focuses on whether people who marry are more similar than would be expected by chance” (pp. 7-8).
In addition, outside observers (family and friends) may have different opinions of the compatibility of a relationship than do the insiders
Similarity. The similarity effect, referring to similarity leading to attraction and satisfaction, has been described as one of the most well-established findings in the study of interpersonal attraction (Berscheid & Reis, 1998) and, indeed, “one of the most robust relationships in all of behavioral sciences” (Berger, 1975, p. 281). The importance of similarity has been demonstrated in many types of research, including mate selection studies, bogus stranger paradigm studies, brief interaction studies, and assessments of existing couples.
In a fourth type of research, based on surveys with actual ongoing couples, degree of actual similarity is assessed. One issue that is examined is whether there is greater than chance similarity in existing couples, which is referred to as positive assortative mating (Buss, 1984). The correlations between partners are generally strong for age, degree of education, physical attributes, overall physical attractiveness, leisure pursuits, and role preferences; somewhat moderate for political and religious attitudes; and weak or inconsistent for personality characteristics and attitudes (Buss, 1984; Barelds, 2005; Feingold, 1988; Gonzaga, Campos, & Bradbury, 2007; Houts et al., 1996; Luo & Klohnen, 2005; Rammstedt & Schupp, 2008). Barelds and Barelds-Dijkstra (2007) found that couples who were friends before their relationship had transitioned to a romantic one had greater personality similarity than those who had rapid onset to a romantic relationship, presumably because those who were friends first had the opportunity to learn more about each other and therefore more effectively engage in 
positive assortative mating. Research has also yielded evidence that active assortative mating (preference for similarity) that occurs at greater than chance cannot be explained away by social homogamy (shared background leading to similarity) or convergence (couples becoming more similar over time) (Houts et al., 1996; Luo & Klohnen, 2005).
For example, one person may view the relationship to be very compatible, whereas his or her partner may view it as less compatible, differences that occur because the two may come to the relationship with different comparison levels or general expectations (Rusbult, 1983)
The matching principle refers to the notion that individuals tend to pair up with others who have about the same level of socially desirable characteristics, regardless of whether the socially desirable traits are the same or different between partners (Hatfield & Sprecher, 2009). Many years ago, Sociologist Erving Goffman (1952) observed that a proposal of marriage occurs when a man calculates his own social worth and suggests to a woman that her assets are not so much better as to “preclude a merger.” Influenced by Kurt Lewin’s (Lewin, Dembo, Festinger, & Sears, 1944) Level of Aspiration theory, Walster, Hatfield, Aronson, Abrahams, and Rottman (1966) proposed that in making dating and mate choices, people will choose someone of their own level of social desirability and they will do so because of being influenced by both the desirability of the other’s traits and the chances of obtaining the other (Walster et al., 1966).
While Match (including Match.International which operates in approximately 25 countries, including Mexico) and other similar sites offer primarily a “searching” venue through electronic personal advertisements, a “scientific” Internet matching service was launched with eHarmony in 2000, followed by Perfectmatch in 2002 and Chemistry (part of Match) in 2005. (True also claims to provide scientific, compatibility matching.) These sites distinguish themselves from others by offering a “scientific approach” to matching (e.g., Gottlieb, 2006). Members who seek matches at these sites complete a lengthy questionnaire, which the sites state have “science” behind their construction (e.g., “PhD designed”). “Matching algorithms,” also claimed to be guided by scientific principles, are used to sift through the data and match pairs (Orenstein, 2003). Users pay more for the scientific matching sites than for the sites based on posting profiles.
The Duet Total Compatibility System at Perfectmatch is based on both similarity and complementarity, and Pepper Schwartz has argued that both are necessary for romantic compatibility (see interview reported in Gottlieb, 2006). The test is described as being based on the same theory behind the famous Myers Briggs Type indicator. In a relatively brief questionnaire (with items that are dichotomous yes/no questions), eight personality characteristics are measured: romantic impulsivity, personal energy, outlook, predictability, flexibility, decision-making style, emotionality, and self-nurturing style. Schwartz has stated (see Gottlieb, 2006) that similarity operates for the first four factors, and either similarity or differences for the final four. The Perfectmatch website also refers to matching “not only with people who are similar to you but also people who complement you.” XI
Another difference is that social networks ent of Internet-based relationships. In traditional ways of forming relationships, people are often introduced by friends or two people meet based on friend-created social settings that bring people together (e.g., Parks, 2007). Relationships formed through the Internet need to overcome barriers of geographic distance (in some cases) and lack of integration of the couple in a larger social network in order to become compatible for the long-term.
