16.2.6 Sexual Orientation and Gender Identity

There are no census estimates for Sexual Orientation and Gender Identity (SOGI). Instead, we have based our risk scores on the California Health Interview Survey (CHIS) population estimates, as CHIS is the largest California survey and has publicly reproducible results via AskCHIS and harmonized with the risk scores assigned to the age and race/ethnicity category population thresholds.

AskCHIS 2021 population estimates for SOGI are as follows (Gender Identity population estimate is from pooled years 2019-2021 as per CHIS recommendations):

• Gay/lesbian: 1.2 mil

• Bisexual: 1.5 mil

• Asexual: 400K

• Transgender or non-conforming: 279K

Based on the risk scores for age and race/ethnicity, Sexual Orientation data aggregated to the above CHIS categories would be assigned a +2 score as the smallest group (Asexual) falls within the "300,001 – 4 million" risk category. Gender Identity data aggregated to the above CHIS categories would be assigned a +3 score as "Transgender or non-confirming" falls within the "100,001 – 300,000" risk category.

There are no estimates for more granular categories such as genderqueer and two-spirit. A +5 risk score was assigned based on the “Other Variable” category for 5-9 groups in the DDG.

The option for a variable limited to "Male or Female" has been kept for datasets that are limited to these options, or in cases where other categories aren't displayed due to confidentiality concerns. The risk score remains at +1 as both populations are greater than 4 million.

Table 18: Sex, Sexual Orientation, and Gender Identity Scoring

Variable	Characteristics	Score
Sex	Male or Female	+1
Sexual Orientation	Straight, Gay or Lesbian, Bisexual, Asexual	+2
Gender Identity	Man/Male, Woman/Female, Transgender or Non-Binary	+3
Gender Identity	Man/Male, Woman/Female, disaggregation of Transgender/ Non-Binary category into more specific identities (e.g. Genderqueer, Two-Spirit, etc.)	+5

Distinction Between Sexual Orientation and Men Having Sex with Men (MSM) Activity

While the CDC definition of sexual orientation includes an implicit behavior in "sexual attraction", this guidance defines Sexual Orientation and Gender Identity as identifiable factors (akin to race, ethnicity, and age) to distinguish them from sexual behaviors such as "Men having Sex with Men" (MSM). This is because the former may be a publicly identifiable trait and thus a re-identification risk, while the latter is, generally, not an identification risk. For example, some individuals may engage in MSM activity but not identify as gay, while others may identify as gay but practice abstinence.

We acknowledge that the boundary between the two is not clear cut and programs will need to assess on a case-by-case basis as to whether MSM should be considered a risk factor for redisclosure. An example of this is an incident cited in NIST SP 800-188 where geographic location from app data was used to infer MSM activity for an individual:

According to commercially available records of app signal data obtained by The Pillar, a mobile device correlated to Burrill emitted app data signals from the location-based hookup app Grindr on a near-daily basis during parts of 2018, 2019, and 2020 — at both his USCCB office and his USCCB-owned residence, as well as during USCCB meetings and events in other cities.

This incident demonstrates the possibility that if an individual is known to use an app or frequent certain locations, one may infer MSM activity for the individual. This makes it potentially usable as a quasi-identifier to re-identify an individual. As mentioned by Abowd and Hawes in 21st Century Statistical Disclosure Limitation: Motivations and Challenges:

The release of any statistic derived from a confidential source always carries some incremental risk of disclosure of identifiable information.