imo porn preferences are mostly a function of the porn sites' UIs and algorithms
for instance, TikTok used to promote dance content, now it's vlog content. people might say their viewing preferences have shifted, but it was more a platform decision than a collective one
3 of the biggest takeaways I got from my college undergraduate stats class is that a) limited data is always better than no data, b) when doing surveys, survey design is way more important than the size of your dataset, and c) randomness isn't everything. You seem to get it.
Assuming you have some demographic info in the survey, it would probably be pretty easy to do MRP on it and it would likely improve the description. An example with R code: