I'm pretty sure the age and gender in that table is just showing the frequency of the ages in the sample, not a crosstab of age or gender with personification/anthropomorphism.

So that's saying their autistic population skewed younger than their non-autistic population. Which isn't unsurprising, it's a lot easier to get a diagnosis as a child, and generally easier to get diagnosed now compared to a few decades ago. So people over 35 or so are going to just be less likely to have had the opportunity for diagnosis. The authors do address differences in gender representation between the samples but I don't see age addressed specifically. It could just be that younger people tend to personify/anthropomorphize more, so since the sample of people with autism skewed pretty heavily towards the 16-24 group the results could instead be displaying differences by age. I don't think they quite have the sample size to delve into age too much. I think they'd only be able to get away with doing two groups at 34 & under and 35+. That would be a good start though.

This is also a heavily self-selected population, apparently largely from social media. I'm always automatically skeptical of social media sampling.

I would've liked to see a little more detail about exactly which tests and assumptions they were using. The gender difference looks like they did a t-test, but it's left to the reader to assume they ran a two-tailed t-test. They could easily have bolstered their numbers by reporting the one-tailed test.