The research I've seen claims that the benefit of diversity is higher performance, not increased collegiality. They may even be inversely correlated. From a Kellogg School of Management (Northwestern) piece (1):
"The mere presence of diversity in a group creates awkwardness, and the need to diffuse this tension leads to better group problem solving... while homogenous groups feel more confident in their performance and group interactions, it is the diverse groups that are more successful in completing their tasks."
Apples and oranges. The issue Erica is concerned about is her social discomfort at work, not her team's performance. Also, the study you cite was measuring gender and length of group membership diversity, not racial/ethnic diversity.
The last team I worked on had 2 Englishmen, 5 Indians, 2 Pakistanis, a Frenchman, a Portuguese, a Greek, an Australian, and a Kazakh. It was chaos. No one understood each other, and so heated arguments were the norm. This was not higher performance.
Presumably any benefits gained from diversity can very easily be lost in the mix of language and true cultural barriers. In the spirit of the article it'd be more relevant if it were a mix of genders, races, and socioeconomic standing at birth than a mix of different nationalities, languages, and workplace customs.
"The mere presence of diversity in a group creates awkwardness, and the need to diffuse this tension leads to better group problem solving... while homogenous groups feel more confident in their performance and group interactions, it is the diverse groups that are more successful in completing their tasks."
(1) http://insight.kellogg.northwestern.edu/article/better_decis...