3 ms·

Conclusion drawn from experiment on 35 samples? Must be kidding...

• always_good 8y ago
  Taking issue with small sample sizes is a popular way for the layperson to chime in with criticism, but it is rather empty and uneducated criticism. In statistics, sample size does not tell you if correlations are significant or not. Which is why it's amusing when someone thinks they are being a statistics pedant by pointing out a small sample size.
  • projektir 8y ago
    I think this is worth challenging. A small sample means there's a higher likelihood that the sample is not very random. It's often pulled from some easily available source, like a group of college students. An idealized perfect random small sample might be just fine for statistics, but there's little reason to believe that's the sample we're getting. Small samples should be a cause for suspicion.