Investigated and conceptualized by Clore and Byrne (1974), the similarity-attraction paradigm states that sharing certain characteristics has a positive effect on relationships. In other words, people are more attracted to people who are similar, rather than dissimilar, to themselves. This could mean that sharing a demographic characteristic such as gender, could affect LMX relationships positively. Indeed, Wayne, Shore, and Liden (1997) suggest that gender significantly and positively impacts the feeling of similarity, and thus subordinate’s evaluation of supervisors. Vice versa, supervisors who see a higher degree of similarity between them and a subordinate, are more likely to categorize this subordinate into an in group. Consistent with these results, Varma and Stroh (2001) found that same-sex dyads reported higher ratings of LMX, and that in both in and out groups, the majority seems to consist of same-sex members. They found that being in a same-sex dyadic relationship significantly predicted a leader’s interpersonal affect. Moreover, female subordinates with female leaders received higher LMX scores than male subordinates with male leaders. This supports the notion that female-female dyads are of highest LMX quality. This could be because of the fact that they invest more time and attention to relationships, making communication between females more smooth than in female-male or male-male dyads.

Supporting these results, in a 1978 study by Larwood and Blackmore, students were instructed to recruit participants for a research project. The results of the study showed that female students tended to recruit more female participants, and male students tended to recruit more male participants, suggesting a same-sex preference. Research also shows that males tend to believe successful leaders have more “male” characteristics and personalities, and females tend to believe successful leaders can be both “masculine” and “feminine” (Schein, Mueller, Lituchy, & Liu, 1996). This could indicate that male-male dyads are of higher LMX than female-female dyads. However, in other studies, this positive relationship between demographic similarity and LMX has not been found (Liden, Wayne, & Stilwell, 1993; Basu & Green, 1995; Goertzer & Fritz, 2004).