Shitposter while I tend to two babies. Maybe when I have my life back, I’ll help us get a few more niche communities back?

  • 0 Posts
  • 590 Comments
Joined 2 years ago
cake
Cake day: July 8th, 2023

help-circle



















  • Yup, although technically you can still run z-tests and t-tests for statistical validity because samples are always normally distributed around a population mean (as long as it’s a large enough sample), even for skewed or uniform distributions. Some tests work better than others, though. The graphs z might be a sample, which might make sense is population data is available for these phrases, but otherwise I’m a little confused why that’d use it, lol.

    To clarify for the commenter, though, Z-scores themselves can be applied to individuals; it’s just a simple conversation setting mean to 0 and standard deviation to 1. In that case, it’s strictly a relative case, that 2.5% isn’t a probability so much as the percentile – % cases that fall below that score. It’s a nifty metric to compare things that aren’t on the same scale, for instance.