[always_good]

You can't condemn sample size without also condemning the statistical methodology of an experiment.

"Small sample size thus irrelevant" does not make sense despite often being the top comment on Reddit.

[Grieverheart]

This is correct. Sample size should be dependent on the strength of the effect. As a rule of thumb though, a sample of 50 people is very little. If physics used these kind of sample sizes and confidence levels, we would have discovered multitudes of new particles.