I love OKCupid's stats

[andrewducker]

Men: 90% straight, 6% gay, 4% bi.

Women: 85% straight, 5% gay, 10% bi.

(from)



And highlighting the difference between identity and behaviour, when they asked over 250,000 straight people:


The analysis of personality traits by orientation is also fascinating.

From

Comments

[momentsmusicaux.livejournal.com]

I wonder why there's that funny bend at 3.

[sigmonster.livejournal.com]

It looks lumpy because it is lumpy - it's discrete valued data. Expressing it as percentiles has the side-effect of smoothing out the larger values. If it was a plot of number of partners against number of respondents reporting that value, all the values would look equally lumpy except maybe for a few right around the median.

[drdoug.livejournal.com]

I love their stats too, but they do still jump from this:

Men: 90% straight, 6% gay, 4% bi.
Women: 85% straight, 5% gay, 10% bi.

... to then only producing graphs and charts for straight and gay. Given 55% male to 45% female (from TFA linked), that gives a total of 5.6% gay and 7.2% bi. Not sure I can see the justification in the data for ignoring the larger subpopulation.

And it's not as if it wouldn't be interesting. Ok, so I'm biased to be more interested than average, but I think most people would be interested to see how the bi graphs compare on all those questions.

They're better than most, to be fair - they have reported on bi stuff interestingly in the past, and touching on just these questions. Which is why it seems odd to ignore it here.