Wednesday, May 16, 2012

Knowing and Uncertain

NPR did a piece this week about uncertainty, and I loved it.  Why?  Why would a scientist want uncertainty; don't we do all that we can to limit confounding variables, measure, collect data, and reduce uncertainty?  Don't scientists want to know things and answer questions?  Here's the thing- those are all different issues.  Researchers most definitely design experiments, measure, collect data, and try to learn.  That doesn't necessarily answer questions, and it most certainly doesn't prove anything.  We collect data, and use that new data, previously collected and interpreted data that has been disseminated to the public sphere, add our own interpretation to the new data, and disseminate our data and interpretation.  We can provide support for a hypothesis or theory, or we can refute these things.  We almost never "prove" something.  That's not the way science works, and it's not the goal of science.

To have "proof," one would have to assume static conditions, that don't change, and omniscience.  However, Heisenberg's uncertainty principle asserts that whenever we reduce uncertainty somewhere, we increase uncertainty somewhere else.  Essentially, as we learn, we learn how much we don't know.  If we answer one question, we raise five more.  As much as I may disagree with him on policy, Rumsfeld got it right when he talked about "known knowns, known unknowns, and unknown unknowns."  He wasn't talking about research or science, but his description works here very well. Science is about rejecting hypotheses that the data do not support, not finding truth.  Science disproves, it doesn't prove.  It's the perfect career for a self-esteem challenged contrarian like yours truly, because I don't have to say "I am right," I can simply say "You are wrong, and here's the reason."

Something else that science does do is change.  As we learn new techniques, as we invent new technology, and as we answer some questions, we find new questions to ask, we can measure previously unmeasurable quantities, and we can perform experiments differently, not to mention re-interpret data with our new insights.  At the same time as we create new ways to conduct science, the world around us is changing as well, so that we can only apply new techniques, principles, or technology to the current world, not the past, revealing uncertainty that we might not have been aware of previously.  We can see this frequently in health science, where what was "good for you" last week may be declared "bad for you" in another month.  That doesn't necessarily mean the science was bad, just that the studies came to different conclusions, for any number of reasons. 

This lack of certainty and certainty of change is hard for many humans, because a lot of us desperately want some kind of certainty and stability.  We want something to hang on to, like a board on the waves as the world tosses us about.  The thought is appealing, I know, but that's not what science provides.  Many turn to religion to serve this purpose, and that's a valid choice (provided they don't try to make policy based on it or force it on others).  Maybe that's why Americans tend to distrust science?  I don't know.  For some people, that uncertainty is filled by our faith in other humans and the natural world.  Others embrace uncertainty, or do something completely different.  Our personal experiences, nature, and nurture shape which path we take.  The one thing that is certain, aside from a lack of funds, there will always be career opportunities in science, thanks in part to this uncertainty.   Obviously, right now, lack of funding is a very real problem, but that's another issue.

All of this is to say that if someone tells you that something has been scientifically proven, ask questions of them.  Be skeptical.  Consider what role the person fills, what they stand to gain, and where their personal bias may lie.  At the same time, if you hear a scientist use "may", "could", "potentially", "might", and other "hedging" words, don't discount them.  They're not trying to sound equivocal; they're being a good scientist.  We just deal in data, and leave the "truth" to other people. 

And don't think that all this means we won't defend the data and interpretations passionately, but we'll do it based on logic and the data, although we are human and have human failings just like everyone else.  For the times our failings pop up, please consider forgiving us, just like you would anyone else, and we'll all try to learn from our mistakes.