I did a bit of googling, and it really surprised me. I thought the social science IQs would be lower on average than the STEM IQs, but I found a lot of conflicting stuff. Most sources seem to put physics and maths at the top of the ranks, but then there's engineering, social science and biology and I keep seeing those three in different orders. If you split up 'social science' and 'humanities', then humanities stays at the top and social science drops a few places, presumably because law is a very attractive profession for smart people (high prestige and pay) and law is technically a humanity. I'm not very confident in any of my Google results, though - they all looked slightly dodgy - so I'm not linking to any and would love it if someone else could find some better data.

I don't think it's an argument for disregarding social science, even if we did find data that showed all social scientists are stupider than STEM scientists. I mean, education came last for IQ on almost all of the lists I looked up. Education. Nobody is going to say that this means we should scrap education. If education really does attract a lot of stupid people, I think that is cause to try and raise the prestige and pay of education as a profession so that more smart people do it - not to cut funding for schools. (Though the reason education is so lowly ranked for IQ could be that a lot of countries don't require teachers to have education degrees, you get a different degree and then a teaching certificate, so you only take Education as a bachelor's if you want to do Childhood Studies and go into social care/work.)

It's clearly very important that our governments are advised by smart social scientists who can do experiments and tell them whether law X or policy Y will decrease the crime rate or just annoy people, or we're just letting politicians do whatever their ideology tells them to do. So, even though the IQ of people in social sciences is lower on average than the IQ of people in physics, we shouldn't conclude that social science is worthless - I think we should conclude that efforts must be made to get more smart people to consider becoming social scientists.

I also don't think you necessarily need a high IQ to be a successful social scientist. Being a successful mathematician requires a lot of processing power. Being a successful social scientist requires a lot of rationality and a lot of carefulness. If you're trying to do some problems with areas of circles, then you will not be distracted by your religious belief that pi is an evil number and cannot be the answer, nor will you have to worry about the line your circle is drawn with being a sentient line and deliberately mucking up your results. Social scientists don't need as much processing power to throw at problems, but it takes a lot of care and ability to change one's mind to do good social science, because you're doing research on really complicated high-level things with sentient agents who do weird things and you were probably raised with an ideology about it. Without a good amount of rationality, you will just end up repeatedly "proving" whatever your ideology says.

To make physics worthwhile you need high IQ; without that, you'd produce awful physics. To make social science worthwhile, you need to be very very careful and ignore what your ideology is telling you in the back of your mind; without that, you produce awful social science. Unfortunately, our society's ability to test for IQ is much better than our society's ability to test for rationality, which could explain why more people get away with BS social science than they do with BS physics. (The other explanation is that there are both awful social science papers and awful physics papers, but awful physics papers get ignored by everyone, whereas awful social science papers are immediately picked up by whatever group whose ideology they support and linked to on facebook with accompanying comments in all-caps.)