Reason Hit and Run touches on a subject that really needs to be given more public exposure, namely the pseudo-scientific nature of "social science." They mention that these self-proclaimed scientists have a problem with predicting outcomes based on their research, which is at the core of what is wrong with the "social sciences." A major requirement of science and derivative fields such as the engineering and medical professions is the sort of repeatability that eludes social scientists.
I see the biggest problem that these fields face is that they face an open-ended, almost completely non-deterministic system for observation. All other systems that scientists study and engineers work on are for the most part discretely-defined systems ranging from experimental physics studies to computer systems. A software engineer can in most situations reproduce any given bug provided that they have sufficient criteria to narrow down the behavior of the defect. "Social engineers cannot" because in general they cannot agree on the variables that might be the cause, let alone agree on how and why they cause problems. This is essentially a scientific mask for ideology.
The biggest grievance that should be held against the social sciences is that their practitioners generally just don't learn from history or their mistakes, nor consider multiple outcomes as part of causality. People who work in actual engineering positions know that changing one variable may have a cascading reaction on an entire system; this is often a foreign concept to the social sciences. Tightening up gun control in a city might save some lives, but cost other lives where gun ownership might have lead to successful self-defense, increase the lack of respect toward the public from less professional police officers and contribute to a crime increase that has a damaging impact on the economy. Three outcomes from own variable change in crime policy, each of which may in turn spark many more changes.
There are those who would defend them as being still useful for statistical study of current problems, but even that is dubious for a few reasons. First, with surveys the methods used and questions asked must be extremely good if they are to have any value as usable data on complex issues. Second, statistics themselves are not necessarily going to yield worthwhile information about why a social phenomenon is occurring. Finally, when you look at the first two issues, it is highly unlikely that the data yielded will be reproducible to the same degree, which essentially negates the possibility of making informative, verifiable theories that explain complex social phenomena.
The more I see of social "science" and survey techniques, the less I respect them as anything useful at all.
Even in my own field, surveys and opinion polling are becoming useless because people LIE on polls and about their opinions. In interviews, people game the responses and say what they think you want to hear or what will make them feel good about themselves.
Basically, if a trained observer cannot measure a response through direct observation, you should question whether the response truely exists. Just like global warming.
Yep. If the hard sciences used their techniques, we would still be living in squalor and misery. If engineers used them, people would be dying by the hundreds of thousands every good year.
How do we get people to stop calling them sciences?
I think the first step is to start trying to use scare quotes around the science part. Come to think of it, I should probably update this post to set that example, shouldn't I?
This reminds me of the surveys that college people have to take to help them decide what they will be when they grow up. (briggs-meyers?)
My results confirmed that I would be best at being an Army officer. I took the test at West Point. I intentionally answered questions like "I prefer working outdoors" and "I like being in charge of other people" instead of "I prefer working in a basement cubicle" and "just give me a task and leave me alone."
2nd choice: Marine corps officer.
Third Choice: Naval officer.
Such a valid test. It is like they could see into my future.