I only skimmed the wikipedia article, but it seems like it is missing emphasis on the biggest problem in soft sciences like psychology - the fact that a huge chunk of their data comes from self-reporting by subjects.
It's the equivalent of basing nutrition science on a Pew poll where people self-report their favorite food.
Sure, it's useful to know people's general preferences sometimes, but for science that data is junk.
Self-reported data is subjective, but when the very thing you are studying is the self-reported subjective experiences of people then it is actually the only data you should care about. Yes “I don’t feel depressed anymore” is a subjective statement but so is “I feel depressed, can you help me?” That specific example is a caricature of course, the data are usually much more specific than “do you feel depressed?”
It's the equivalent of basing nutrition science on a Pew poll where people self-report their favorite food.
Sure, it's useful to know people's general preferences sometimes, but for science that data is junk.