Useful Research?

In response to this post, NW asks:

1. Before deciding what research projects to pursue (and, thus, before knowing what we will discover when we finish the projects), how do we know which projects are useful and which ones are not?
   
   
2. Consider math and physics, two fields that are often seen as esoteric to most people, and two fields that many often question for their relevance or usefulness to everyday life. ("Why should we spend money on sub-atomic physics when we haven't cured cancer?) Much basic research in these fields have led to greater understanding about the world that then inspired the applied research to give us great things like MRI, PET, and CAT. In biology, the principles of natural selection that Darwin discovered have had a tremendous impact on drug development (e.g. Superbugs), and Darwin likely didn't think of such wide-ranging applications when did his research.

Thus, when you say useful, do you mean applied research, or basic research that people would still have some reason to care about? Either way, I think that the above examples show that basic research, unrestrained by boundaries of what is considered to be useful, can give us greater understanding of things that we otherwise would not have.

What do you think of this position? (By the way, by no means am I ignoring that intellectual masturbation exists, and I do think that it better to do basic research that people care about than basic research is esoteric and irrelevant. However, I would appreciate your views on:

• how usefulness is defined
• whether there is a case to be built for unrestrained basic research

These are important questions.

First, there's a difference between research in the "hard" and "social" sciences -- let alone the humanities!. Much social science research is is contentious, controversial, opinionated and beyond replication. Research in the hard sciences is much more incremental. Although there are conspicuous exceptions, quite a bit of hard research is still valid many years later, serving as useful building blocks for later research. A lot of social science is dead-end, contradicted by other work elsewhere, and/or too flawed to use.

Second, a lot of "basic" social science research (e.g,. game theory, macroeconomics) cannot be used anywhere in the real or imagined world. It tends to be "mathurbation" build upon a series of increasingly tenuous assumptions. Although Friedman said that assumptions don't matter as long as the predictions turn out to be true, the trouble with that line is that we don't KNOW when the predictions will be true or false if we cannot replicate the "weird" assumptions. That's a troubling "detail" when one is making predictions or policy for the stock market, drug testing, etc.!

Bottom Line: The best research can be falsified, replicated and applied. Other "research" just keeps academics from getting into trouble while they jet between conferences, pretend to teach (the incentives are for RESEARCH, remember), blog -- as if they know something :), and continue to be paid by YOU.