I have read part of this silly article. I did not read the whole thing because it became clear almost immediately that this was one of those articles portraying studies as scientific experiments. They are not for a number of reasons one being the fact that there are two many uncontrollable variables, two being that there can be no random assignment of the variable in question and three being that there is no way to fully verfiy compliance. I am really frustrated with social scientists pretending that they are real scientists. They are not. They only play at being scientists.

Therefore I am going to spank anyone I see misbehaving under the age of ten. My kid, your kid, the Mayor's kid; doesn't matter. Spankings for everybody.

Damn, I was hoping you really meant *everybody*. How come grown ups get off so easy?

I like your come-back. Nice job!

[Therefore I am going to spank anyone I see misbehaving under the age of ten. My kid, your kid, the Mayor's kid; doesn't matter. Spankings for everybody.]

why stop at spanking? considering your cavalier attitude toward statistics, you could up the ante to sodomizing. i'm sure that will stop the little buggers...

On the contrary I don't have a cavalier attitude towards statistics which is why I dislike social science.

Do you think that some social scientists might be more conservative than other social scientists when it comes to the use of statistics? Or are all social scientists equally cavalier....I guess you would argue that we can't answer that question with any degree of certainty so we should just assume the answer is NO. Well.... what about other scientists then??? Are other scientists really less cavilier with their use of statistics than social scientists....Crap! Without social science we have no idea....

BTW: I personally try to stay away from social science (except game theory and parts of evolutionary psychology) because much of it appears to be related to aesthetics that I do not understand or properly appreciate. But it is far from my expertise to provide a scientifically grounded critique of social science as a whole.

BPKowal,

Well you are correct in a sense. I wrote this because of my frustration because I see many methodological problems in much of social science. But there is social science I enjoy and find useful such as the work of Daniel Kahneman in psychology. One does not need social science to determine if a particular individual is cavalier. One can simply examine their methodology. My problem is less the scope that social scientists try to cover with a particular statistic (although that is a huge problem) than the fact that they are trying to find a "scientific" answer to a question that is not scientific.

Inquisitor,

I get it - I think (Kahneman has become a personal hero of mine over the last few years as well). But I am still not sure that there is a good (universal?) guideline to what makes a scientific question and what does not. Again, this feels more like aesthetics than logic. Just think (and I can sympathize with you on this one ----aesthetically) ---why does it seem more scientific to study weather patterns with 15-20 known uncontrolled variables than patterns of social human behavior?

bpkowal

bpkowal:

BTW: I personally try to stay away from social science (except game theory and parts of evolutionary psychology)

Okay, I was totally fine with the whole "Social science isn't really science" spiel until this point. Evolutionary psychology? Really?

Well we know whats on your mind.

Bpkowal,

I think I understand your point. Let me try and address it. Let me start with weather. In the short term we have gotten pretty good at predicting weather. Say within two days. But in the long term we are still completely lost. However I believe that the reasons we are completely off in predicting what happens in physical systems over the long term are different than the reasons we are off about social systems. First I think we have an exceedingly good understanding of the way physical systems work and the factors involved and what there impacts are. However we are horrible about prediciting long term results because any tiny change in the system can have a huge unanticipated impact (Nicholas Nassim Taleb has a great passage on this in The Black Swan or Fooled by randomness). Thus the problem is that we essentially don't and perhaps can't have enough data. I don't think this is the core problem with bad social science. I think that much of social science especially this kind which is in a highly politicized field doesn't even understand the system. This is clear from the article. The author is trying to prove that a particular input has a particular impact. A weather forecast does not do this because there is no disagreement about what the impact of inputs are there is simply ignorance about what the imputs are. So what you see in bad social science is a lot of begging the question and a priori assumptions. So I don't think that it is simply an aesthetic question. I have similar problems with some branches of physical science.

As a social scientist, it is my experience that most social scientists do abuse statistics to some degree. The level to which this is intentional deception is unclear (probably low). But it is wise to be wary of statistical studies from a highly politicized discipline like sociology, where much of the published research is used to validate already preconceived notions of the ideal society. Most social scientists incorrectly interpret statistics beyond what the particular analysis allow them to do. This leads us to a whole lot of bad information circulating in society and a deterioration of the discipline (which is really now more quasi-scientific).