> In general, hard sciences are much more reliable than social sciences because standards are higher and topics are less emotional.
Having read lots of both, I'm not sure that's true. There's no way to prove it because nobody has clear definitions for what the words hard, social, science, standards or reliable mean. But the extreme political bias doesn't go away, researcher degrees of freedom are just as large, the topics are sometimes much more emotional, and a lot of fields you'd expect to be hard are methodologically no different to any social science.
For example, a guy in Wales recently claimed a big payout from a fraud suit he won against the Dana Farber cancer center at Harvard. It'd been publishing fraudulent papers for years, yet either nobody noticed or nobody cared. Climatology tells people that the end is nigh, a message sufficiently distressing to make some psychologically vulnerable people commit suicide. Do any social sciences have an emotional effect that extreme? A lot of the COVID pseudo-science was specifically designed to manipulate people's emotions (e.g. journals rejecting correct papers because fewer people might take vaccines as a result [1]). And epidemiology isn't based on any empirical understanding of viruses or disease. It's just modelling no different to the type described in the article.
Unfortunately, ideology and bad incentives are the same no matter what field you look at. There is a hard/soft distinction to be made, but it's more about how close the field is to engineering. Engineering fields have a lot of movement between public and private sector, which keeps the universities a bit more honest. Maybe other fields like law and finance are the same, I don't know, I never read papers in those.