Of those who had met in the two years prior to the study, 23% of heterosexual couples and 61% of gay couples had met through the Internet
In a recent published study, Carter and Buckwalter (2009), scientists at eHarmony, compared a group of 157 couples who had been matched through eHarmony with a comparable group of couples (matched on marriage length and age of spouses) who were recruited through a pling procedure. On a number of variables that referred to personality, affect, and values, the couples who had been matched through the eHarmony site were more similar than those who met in other ways. In addition, the online matched couples had higher satisfaction scores, as indicated by their scores on Dyadic Adjustment scale. Similar results had been found in an earlier study reported by Carter and Snow (2004), using a similar sampling procedure.
In a recent study not funded by a dating website, Sociologist Rosenfeld (2010) analyzed a new data set, Wave 1 of the “How Couples Meet and Stay Together.” The sample consisted of a U.S. national representative sample of 4002 individuals, 3009 who were partnered. Rosenfeld found that the Internet has clearly gained in importance as a way to meet partners. Rosenfeld compared couples, based on how 
they met, on a relationship quality score, and found no significant differences. In additional analyses that controlled for a large number of variables, including relationship duration, how the couple met continued to be unrelated to relationship satisfaction. Although Rosenfeld’s study was based on a representative sample, it is also limited in what it can tell us about scientific-based matches at Internet sites because: 1) the analyses did not distinguish between those meeting through dating services versus those meeting in other on-line ways (e.g., chatrooms); and 2) the satisfaction measure was only one item. Therefore, more research is needed on this issue of whether scientific matching can create more compatible matches. In fact, OnlineDatingMagazine has advertised a survey that assesses whether online relationships are more successful than offline relationships. XXIII Regardless of the findings, however, this study will also be limited by self-selection biases of the couples who respond.
This research could be considered as supportive evidence that scientific matching leads to compatible relationships, although it is difficult to draw valid conclusions based on a comparison of two convenient samples that each have unknown selection biases
Speculation also exists on how the attraction process differs between relationships that meet in traditional face-to-face contexts versus those that meet on-line, regardless of the specific on-line venue (Cooper & Sportolari, 1997; Merkle & Richardson, 2000; Sprecher, 2009; Sprecher, Schwartz, Harvey & Hatfield, 2009). The speculations offer suggestions as to how components of or pathways to compatibility can differ as a function of how the relationship begins. For example, the process of attraction in a face-to-face romantic relationship is likely to involve first the influence of proximity and physical attractiveness, and then the discovery of similarity, followed by the rewards of self-disclosure (Merkle & Richardson, 2000). In contrast, Internet-initiated relationships have been described as involving “an inverted developmental sequence,” (Merkle & Richardson, 2000) which first often involves a high level of mutual and sometimes intense self-disclosure, and an initial minimal role for physical attractiveness and proximity. Although there may be an exchange of photographs between potential matches, physical attractiveness and other “chemistry” factors generally play less of a role initially. Furthermore, once two people meet, the impact of physical attractiveness can be reduced because it follows learning other information about each other. As Cooper and Sportolari speculate, “the felt intensity and meaning of any unappealing physical traits are then more likely to be mitigated by the overall attraction that exists” (p. 9).