|
Abstract:
|
The inclusion of human sex and gender data in statistical analysis creates a nuanced problem for the modern statistics practitioner. There are multiple, overlapping, considerations for data collection, combination, analysis, and interpretation that involve variables intended to represent sex and/or gender. A few of these considerations include the privacy, protection and ethical treatment of human subjects, avoiding non-response and missingness, and mitigating the potential for classification bias. These considerations are not unique to variables representing sex and gender, nor to decisions that are informed by analyses involving these variables. However, consideration of the relevance of aspects of the ethical practice standards for statistics and data science to sex and gender variables is timely, with results that can be applied to other socio-cultural variable definitions and selection. Historically, human gender and sex have both been categorized according to a binary system. The continuation of this tradition in modern research persists mainly because it is easy, and not because it produces the most valuable scientific information. Utilizing the traditional binary classification of gender enables older and newer data sets to be wrangled, munged, and combined easily. However, this classification system eliminates the individual’s ability to articulate and assert their gender identity and conflates gender and sex. Moreover, defaulting to a binary system for sex obscures potentially important differences by collapsing across many valid and authentic categories (e.g. involving primary and/or secondary sex characteristics). This approach perpetuates historical, inaccurate, simplicity - and bias - while also limiting the information that emerges from analyses of human data. These limitations also violate multiple elements in the American Statistical Association's (ASA) Ethical Guidelines for Statistical Practice. Information that would be captured with a non-binary classification (including the prevalence of differences in sexual development) could be relevant to decisions about analysis methods as well as to decisions based on otherwise expert statistical work. Modern statistical practitioners are increasingly concerned with inconsistent, uninformative, and even unethical data collection and analysis practices. This paper presents a historical introduction to the collection and analysis of human gender and sex data, offers a critique of a few common survey questioning methods based on alignment with the ASA Ethical Guidelines, and considers the scope of ethical considerations for human gender and sex data from the early stages of design through the later stages of analysis.
|
