Symbolic Self-Completion, Attempted Influence, and Self-Deprecation (1981) [pdf]

Sorry. The sample size is not actually "minimal", and it's this is a common criticism where the critic, surprise surprise, never says what a non-minimal sample size is. "NCHS also typically requires at least 30 observations": https://en.wikipedia.org/wiki/Standard_error.

Here are some key concepts to keep in mind.

The central limit theorem states: regardless of the shape of the parent population, the sampling distribution of the mean approaches a normal distribution as N increases. This means we can generalize what the size of n means to many many things, generally including all sorts of traits in people, and including flipping a coin. So for an intuition, flip a coin 30 times, and 95% of the time, you will be within 5.5 of the mean (15) and it follows the standard bell curve, 99% I will be within 8.22 of the mean, etc. So, if say "there is a greater than 20% chance a coin flips will end up with heads" and you say, "from a statistics point of view, your sample size is too small", I can say, "not true at all", with consideration to my sample size, I'm 99.9999...% likely to be right. And just glimpsing, I'm confident the study did their math right.

There are other very important things to be critical of, and it is certainly reasonable to be skeptical of the study. Much more likely to cause a false result: getting unbiased, representative sample, publishing bias, causation vs correlation, etc, etc.

And when you think about sample size, the accuracy with respect to the population follows 1/sqrt(n). So n=200 is twice as accurate as n=50.