- Joined
- Dec 25, 2008
- Messages
- 8,173
- Reaction score
- 7,536
Then it should be presented as exploratory, which this wasn't (apparently). Like I said, this is the textbook case of HARKing: if you are doing post hoc analyses after your originally planned analysis is nonsignificant without describing them as exploratory.
Like I said repeatedly, they should have definitely included that the linear analysis was non-significant. But I'm not even seeing this as a real deviation from the original hypothesis--that bisexual participants would differ from homosexual and heterosexual participants. The question is just how you operationally define "mostly gay" and "mostly straight" participants. Maybe a forced-choice option would have been better for this study, given these questions, honestly, because there's not clear precedent either way.
Also sounds like there is a need for a better validated measure of sexual orientation. Or a general better idea of the construct of sexual orientation (not on my area of knowledge). If these are not well understood in the field then I wouldn't be working on the OMTH or the double jeopardy hypothesis.
Yes and no. We've been using scaled models of sexual orientation since Kinsey--they aren't new. But they are difficult to translate socially--is someone who's "mostly gay" going to be read the same as someone who's "completely gay" socially? Would a "mostly straight" person categorize themselves as gay, straight,, bi, or other if given a standard forced-choice option? Do asexual homoromantic folks count as "gay"? Can you identify as lesbian if you are dating someone who's AFAB non-binary or AMAB non-binary? We don't really know much about any of that, so we mostly avoid the issue by asking about sexuality in simple forced-choice categories and most people honestly just pick a forced-choice category that fits them best even if it's not a perfect fit.[/quote]