Which is the same issue with many physcology and political survey-based studies. Take a survey that was horribly implemented, run some lightweight analysis (cause what can you really do with 20 respondents anyways), slap a funny name on it, and boom another piece of terrible research.
This + confounding factors + a garbage can regression + a not so random small sample -> their findings are rubbish.
Which is the same issue with many physcology and political survey-based studies. Take a survey that was horribly implemented, run some lightweight analysis (cause what can you really do with 20 respondents anyways), slap a funny name on it, and boom another piece of terrible research.