[twirlip]

... From a pool of respondents who follow this internet personality. She acknowledges that only 10% of her followers identify as women and may respond differently than their cohorts.

[wongarsu]

However getting some paid responses and comparing them to the other respondents is surprisingly good methodology for a blog post. And I think it does show that while the results weren't identical they weren't wildly skewed either

[NavinF]

That is incorrect. She got the same results from both her followers and from spending $1057 on Positly surveying people who are not her followers.

Sometimes people are right on the internet.

[crowcroft]

Yea, this is how I interpret the results.

'Men who watch porn self-identify as "knowing what women like"'. Quite a different insight.

[robotresearcher]

The article compares men's predictions of women's preferences to women's stated preferences. They didn't ask men if they were good at predicting.

How could this support your interpretation?