It's knot science - wolves in furry porn

[Content warning: the following features discussion of furry pornography and gaping confidence intervals.]

After my recent post that found foxes are (statistically significantly) more likely to bottom than top in two-character furry porn, the natural next step was to look at another species for comparison. Wolves were the obvious choice, repeatedly appearing at the top of species lists in various furry surveys, so I set my expert assistant the task of constructing a similar dataset to that used for the fox analysis. A (highly unscientific) Twitter poll suggested three-quarters of you think wolves top more often than foxes, will the data back this up?

The inclusion/exclusion criteria were nearly identical to those detailed on the original post, so I won’t go into too much detail here. We continue to only include original characters (so Wolf O’Donnell’s obsession with Fox McCloud’s bottom remains sadly unquantified), and limit ourselves to images of wolves having sex with a character of a distinct species (so wolf hybrids are excluded - apparently identifying wuskies caused a lot of difficulty here). Werewolves presented a small complication, and I ultimately decided to treat them as distinct from wolves (spoilers: this may have been primarily because werewolves are kinda scary, and if there’s one thing that seems likely to draw their ire, it’s questionable statistical decisions). However, as detailed below, this made no appreciable difference to the conclusions.

Images were again sourced from, over the same 12 month timescale used for the fox dataset to maximize comparability. The final sample size was 267, about 30% larger than the fox dataset. This seems to be (very) roughly in line with species population estimates from past surveys (see links above), so I was happy to assume this increased sample size wasn’t indicative of some sampling issue.

Here are the main results, which should continue to be interpreted in terms of the underlying datasets: images from over the last year, featuring a character of the given species (wolf, fox, or werewolf) engaged in penetrative intercourse with a character of another species, both of whom have a penis. As always, I’m not making any claims about furries or furry characters in general - these conclusions are limited to the context of the criteria I’ve previously set out!

[For those not super-statsy, a quick explanation of the terminology below: a confidence interval can be thought of as a ‘plausible range’ of values for the thing we’re trying to estimate. The p-value gives a measure of ‘surprise’ (the smaller the value, the more surprising) at how one-sided the observed data are if the associated hypothesis (wolves are equally likely to top as bottom, or equally likely to top as foxes are) is true.]

  • In 267 images featuring a wolf, the wolf was on top in 167, or about 63%. (In contrast, foxes topped in their dataset about 32% of the time.)
  • Wolves are therefore significantly more likely to top than bottom (with a 95% confidence interval of 56%-68%, p-value < 0.0001).
  • Wolves are also significantly more likely top top than foxes are (a 95% confidence interval for the difference between how often foxes top and how often wolves top is 22%-40%, p « 0.0001).
  • In addition to the wolf data, 21 images featured werewolves having sex with another species, with the werewolf on top in 19.
  • Werewolves are therefore significantly more likely to top than bottom (with a 95% confidence interval of 68%-98%, p = 0.0004).
  • Treating werewolves as wolves does not change the results of any of these hypothesis tests (as you can probably guess, it just provides slightly stronger evidence that wolves top).

So there you have it - as the poll suggested, wolves do indeed top more often than foxes in this context (and are more likely to top than bottom in general). Please use this knowledge responsibly, and watch this space for more questionable uses of my time!

Written on January 7, 2017