Posted by Stephanie A. Smith, Ph.D., on Mon, Dec 2, 2013 @ 2:00 PM
That’s the question I asked readers to consider in my last blog, Simple Samples Part I. I hope you didn’t lose sleep or skip any football games pondering this question, because it is deeply flawed. Forgive me for setting you up with all that mumbo jumbo about how scientists agonize over sampling. That was all true, and I needed readers to understand that, before asking them to consider this question which begs, practically insists, that you not think like a scientist.
The reason it’s so flawed is because the easiest answer is “However you want.” In other words, it invites you to use bias in your sampling, to get whatever answer you want. We all know how this works. It’s political. If you don’t want to find toxins in your lake, your answer might be “sample off the end of the dock, in January, from the surface of the water, right after a fresh snow.” If you want to find toxins, your answer could just as easily be “In August, sample in that cove where there is a bluish-green scum every August, and skim some of the thickest scum off the top.” Bias is the bane of science, and all scientists acknowledge this, and even admit that we are not immune to the forces of biased sampling.
In fact, this scientist is going to make the risky suggestion that invoking bias when sampling your favorite swimming hole is OK! Just recognize that you’re not really doing scientific sampling when you allow for bias. The reason I’ve entitled this series of blogs “Simple Samples” is because I’m of the opinion that if you want to know whether you might be exposed to toxins in your favorite swimming hole, the sampling can be pretty simple. Sample what you will be exposed to. If there is a scum on the surface and you’re going to plunge in, you will be exposed to it. If the water is a consistent pea-green throughout, you will be exposed to it. If the scum is in a cove but your swimming hole is 20 m away, sample the swimming hole because that’s where you’ll be exposed. It’s not scientific, but it’s probably all that can be reasonably achieved within the constraints of time, space, and money.
I have heard of people arguing over whether a swimming hole is safe or not, because they don’t agree with how the sampling was done, or believe that the sampling was biased. My opinion is that people are arguing over the wrong thing. Of course the sampling was biased. We must recognize that biased samples are not science, but that doesn’t mean they are meaningless. They show what is possible. They allow people to think about risks. Once we accept this, we can move on to Simple Samples Part III: the danger of extrapolating the results of biased sampling.