I wish I knew the date and context of this. With words like "gay" and "woman" being censored out by the U.S. administration, I imagine that official government surveys might end up looking this stupid now.

I also imagine the possibilty that some biased group is attempting to inaccurately count the number of participants that are/aren't straight. If they put choices like "gay" or "bisexual" or "asexual" (if it's even included) on a second page (after clicking "More options"), they're creating friction for pretty much anyone who isn't "straight." That is, not everyone will bother going through another option page - they may settle for "Prefer not to answer" because it's already right there in front of them. Then when totaling up the answers at the end, whoever's using these data can simply omit the respondents that didn't answer. Boom - just like that, it will appear as if a greater percentage of the group are straight. Such easy LGBTQ-erasure.