A lot of that is because social sciences are often genuinely non-rigorous, hand-wavey, and opinionated.
I was required to take a sociology class once. Some of what was presented was solid, like the observed stages of childhood development with respect to self-awareness and the issues with the poverty cycle, but in other areas a lot of theories were presented as the "right" thing to believe that had some combination of being unverifiable, having insufficient sample sizes in their foundational studies, having glaring methodological problems, contradicting other things also presented as factual, or having reaching conclusions.
There was an assignment to make a food plan based on an average food stamp allocation, that let you eat 2000 calories per day along with some other requirements that I don't remember. The professor claimed only one person had ever been able to do it. You weren't expected to be able to do it, but if you couldn't, you had to write an explanation of it and how the experience made you feel - assumedly an "eye opening experience" was expected. It was conveniently not mentioned that food stamps are intended to be supplemental, not for all of your food. The program even has the word supplemental in the title: https://en.wikipedia.org/wiki/Supplemental_Nutrition_Assista...
Regardless, even with unrealistic restrictions, I was able to do it without any trouble. Taking the high values of what the food prices ranged at the local budget food store, I had a result that was well-balanced and below-budget. Of course, the professor had to make some snide comment about "sodium poisoning" to the class about it, which this diet certainly cannot cause.
Here's a paste of it: http://pastebin.com/QnmnHXbA
Not everybody has access to as good of a budget store as ALDI, so results can vary, but this particular ALDI is located in a low-income area, and there's clearly no epidemic of malnourished low-income individuals nationwide given that poverty is correlated with obesity. This was just one example of the many problems in that curriculum. For another example, the wage gap was simply stated to be indisputably the common 77 cents per dollar figure, which carries certain implications of it being entirely discrimination when stated without any qualifiers; however, when just looking at the Wikipedia article, it can be seen that the problem has more subtleties to it, including which portions are "explained" vs "unexplained," and that widely varying figures have been reported. Motivating more women to participate in male-majority fields is a good thing, along with other proposed solutions to help the problem, but none of that was mentioned in the class.
Aside from this one biased professor, I've read from others that these problems are very common in social science classes. It would be nice if there were some data available on how students rate their professors' bias in this field.