Originally Posted by
cspan
Quoting bivariate Pearson correlations, R-squared, F-tests, p-values, etc. [for instance see "notes for nerds" on the backpack survey].
IMO it's a bit much for something that is simply an "informal survey" of those on social media, as you termed it. It's a mismatch - informal survey, formal stats.
The selection bias is more than thru-hikers. It's limited to a) those who are on social media and b) those who respond to survey requests on social media. And it's not clear there's a way to prevent phony responses. I'm not saying that they should have people at Springer and Katahdin with clipboards tracking everyone who is a true thru hiker ... though that could be more accurate ... I'd just lose all the fancy stats.
Reply With Quote