Simon P. Couch
This is a research compendium for a term paper for SOC 322 (Gender & Work) at Reed College in Fall 2019.
I argue that code documentation in the tidyverse is written more effectively and inclusively than other R code documentation, and that this is meaningful for understanding gender representation in the user community.
Initially, I trace the literature articulating supply-side mechanisms for gendered occupational segregation, showing that the work of gender minorities in STEM is subjected to greater scrutiny than that of their cisgender men counterparts. Further, too, gender minorities in these fields are more likely to incorporate this negative feedback, as well as their awareness of masculine elements of their physical surroundings, into their self-perceptions of their competence and identity alignment with the stereotypical worker. Collectively, these mechanisms coalesce into a system of exclusion discouraging gender minorities from pursuing careers in some STEM fields.
I argue that these mechanisms are at play in the context of statistical computing, which incorporates methodological and cultural elements from many STEM fields and, further, that code documentation in statistical computing settings is capable of both delivering feedback about competence and transmitting norms about the stereotypical worker.
Through interviews with women in these fields, I developed a set of measures to quantify the effectiveness and inclusiveness of code documentation, and analytically showed that, using these measures, code documentation from the tidyverse is more effectively and inclusively written than code outside of the tidyverse. These interviews also allowed me to more thoroughly scrutinize this association between gender minority representation in the tidyverse and the increased quality of documentation arising from the community. The interviewees articulated several confounding factors, as well as supply-side mechanisms, relating these two phenomena—namely, I argue that increased documentation effectiveness and inclusivity are a likely mechanism for the greater representation of gender minorities in tidyverse development communities, and also that, in turn, this greater representation contributes to more effectively and inclusively written documentation. That said, the confounding factors highlighted in the interviews are also plausible mechanisms driving this association, and the evidence provided is not sufficient to fully articulate a causal mechanism in either direction.