Comparing Height Preferences to Expected Baseline - A Simulation vs. Surveyed Data

Quiet day today, so let’s talk numbers. As discussed elsewhere, I wish to see higher quality of numbers discussion here, so why not be part of the change.

There’s a lot of talk about height preferences in women wanting taller guys. Yet, there is also a sizable voice around here that many women don’t mind dating shorter guys at all. One thing to note though is that men on average are also taller (70 inches, or 5’10”, compared to 64.5 inches, or 5’4.5” for USA) than women anyways. So all else being equal, we would typically expect men to be taller in any given man-woman pairing. Before any discussion/debate about preferences mattering, we would want to establish what a baseline would be if no such preferences existed. Enter simulation data!

I ran a simulation with 10,000 couples, with randomized height pairings (this only took like 5 minutes btw if you want to recreate yourself. Specs details at the end). There’s 3 groupings: man is taller, woman is taller, and equal height (defined here as within 1 inch of each other…gotta draw cutoff somewhere). Results:

So, ~17/20 of couples will have a taller guy anyways, regardless of preference, and the average difference in height is roughly 6.5 inches (this seems high, but remember, all the people within 1 inch are 'equal', so it’s almost the exact same as the average height difference + 1 inch). ~2/10 will be man-woman of roughly height, and 1/20 would be a pairing where the woman is taller, with an average height difference of 2.6 inches. Thus, based purely on raw demographics alone, we only expect to see maybe 4/100 couples where the man is shorter, and often the height difference is not nearly so notable as a taller man-shorter woman pairing.

Is there a further effect of tallness preference on top of that? For example, if we expect to see 88% of couples with a taller man, but the actual population figure is like 98%, then yes, I think one can make a good case that there’s a main effect driven by preference as well. Looks like there’s some figures for us to compare our default benchmarks to: this 538 article cited another study that in 5,000 man-woman couples surveyed, 92.2% had a taller man. So yes: there seems to be a height preference effect on top of baseline chance alone. Is it stark? When your benchmark is already near the edge at 88%, there’s only so much wiggle room to move further up. I’d bet my degree that it is statistically different enough to say height preference has an effect, but I'm sure this comes as 0% surprise to anyone.

This is also of course a simplified premise in that height difference is abstracted from everything else, and that it's all treated linearly. The snapshot leaves out interesting questions: even for women who don’t mind dating shorter men, there’s probably a cutoff minimum height where that changes. As an extreme example, Peter Dinklage is 4’5”, so like the 0.001th percentile of men’s height. While he’s married to a fairly normal height woman, I seriously doubt most of the women who say they don’t mind dating shorter dudes would consider guys at his height as viable prospects. Likewise, a woman who is 5'11" and willing to date a shorter guy is probably thinking like 5'8"-5'10" shorter, and not someone shorter like 5'2". We could of course simulate these conditions as well, but I have a day job to get back to.

Work was done in Excel using benchmarks of: Men height: 70 in, std. dev. 3 in, Women height: 64.5 in, std. dev. 2.5 in (source for figures is google search “average height and standard deviation by gender US” top result). Formula is ROUND(NORMINV(RAND(),$B$2,$B$3),1), where B2 is height, and B3 is std. dev. Round is used to trim the value to the nearest tenth. Groupings generated with if statements. Generate pivot table and pull numbers. I did 10k couples, but you can do 1k, 5k, 100k, etc. and the outcome should be quite similar. Law of large numbers and all that jazz.

TL;DR: I’m here to support your biases, but with numbers.