>>715042676Depends on your field of study, your topic and what kind of data gathering you do.
Generally the biggest fuckup you see all the time is people making normative statements about - say - an age group or ethnic group or country, while not accounting for major differences inside said group. For example: I had to do a study on peoples' attitudes towards their employers, what they find attractive, how much money or what benefits they want. I only surveyed university students and most of them were economists. The results of my study describe a smaller subsection of the actual labour force - university graduates AND even among them accounts not for their majority.
Thus my study is good for describing that tiny subsection of workers - but that alone is not enough to draw conclusions upon the specificities of their age group or their social status, etc. So when a retard writes an article about some simple correlation, I immediately know that not only are they full of shit, they're probably trying to sell me something.
Generally in medicine, a couple hundred is really good. In social sciences you need many, many thousands.