Scientists are biased

It may surprise you that I would post a title as provocative as “Scientists are biased,” except that it is the standard belief of people who work in scientific research. Oh no, they would never word it that way, because in today’s parlance it means that the person (or study) being accused of bias is not reliable and not to be trusted. Nevertheless, the whole scientific method and the structure of peer review is based on the idea that those who engage in science are biased, although unwittingly. They key word is that word, “unwittingly.” And it is the loss of that word, “unwittingly,” that is so hurting various types of scientific studies around the world. What do I mean?

Let’s look at the scientific method as it has developed over the last centuries. As science and technology have progressed, one of the earliest principles to develop is that no experimental result was considered valid until it was repeated multiple times by the person engaged in the study. Very quickly, the principle was added that the research should be repeated by other scientists in laboratories independent of the original laboratory. This had, as its corollary, that all scientific research, and its methodology, had to be published and publicly available so that it could be vetted by other researchers.

Why was this so? As modern scientific procedures developed, one of the earliest problems to be noticed was what is called “observational bias.” That is, it is all too tempting to see the result that one wants to see. Recently I was talking to a person who has worked for both the CDC and for a forensics laboratory. She commented laughingly about this one laboratory in another country that her unit at the CDC knew about which published new and startling results on a semi-regular basis. Unfortunately, those results could never be replicated outside of that laboratory. In fact, it became a joke at her unit in the CDC that if you wanted a guaranteed negative result you just had to try to replicate a particular experiment from there. In forensics, that is a problem that must be constantly combated since the results of those laboratories often lead to a prison sentence for someone. And, the rate of innocent prisoners released recently in this country points to the danger of observational bias.

There are additional methodologies, such as the use of double-blind studies in order to minimize the danger of observational bias. In addition to that, there is the danger of racial bias, cultural bias, national bias, bias in favor of those who have hired you, etc. As a result, certain types of studies have ended up needing to be verified by laboratories in other countries or cultures just to make sure that the results do not come from some other type of bias. For instance, do you remember when all the arguments began about “alternative” medicine back in the 1960’s and 1970’s? Some of those debates are still going on, and there is much accusation of cultural bias on both sides. What does appear to be true is that studies done in the place where the alternative medicine originated are very supportive of the technique(s) while studies done in places with a different cultural milieu do not show the same positive result. Thus the debates over those techniques continue.

But, note what I have written above. The standard assumption of the modern scientific method, worldwide, is that no individual study proves or disproves anything. No theory can be assumed to be proven until and unless repeated multiple times and in multiple venues, each independent of the other. If necessary those venues may even need to be in other countries. But, there is a more negative way to phrase this. All scientific research assumes that we are all biased to some degree or another and that this bias needs to be countered in order to try to achieve results that are reliable and reasonably reflective of “truth.” More than that, there is an element of Murphy’s Law in the standard assumption of the modern scientific method. That element is the belief that if something can go wrong in your experiment or your study, it will go wrong in the worst possible way at the worst possible time, and may even go repeatably wrong in the worst possible ways at the worst possible times. Thus, part of the reasoning behind the requirement for something to be repeatable by other laboratories is the presumption that you cannot tell whether Murphy’s Law was at work in your research at the time you were doing your research.

There is an interesting side effect to all this. Because of the assumption of bias and Murphy’s Law, it is not necessary for every laboratory to perfectly agree before some bit of research is considered accurate and “true.” There may even be one or two laboratories, out of a large mass of laboratories, that may not be able to repeat the results. [That is why scientists talk about concepts such as “two standard deviations,” etc.] But, eventually, the proof is in the pudding, even for scientists. You build an experimental plane after doing all the research, and it either flies or it does not fly as per the theoretical data. You do drug trials and either the drug works as expected or it does not. Nevertheless, truth in science has never been a concept of everyone perfectly agreeing, but the vast majority agreeing.

This is troublesome for many non-scientists, despite the fact that they live with the beneficial results of this type of approach to finding out data about the natural world. In fact, you are reading a blog that would have been impossible just in my lifetime because of the results of scientific research. It is troublesome because, especially in today’s cultural climate, we are insisting that something either is or is not. We have trouble with gray. We have trouble with the idea that something is most probably true. We do not like degrees of truth; we want 100% certainty. But 100% certainty is not found within the confines of scientific research.

===MORE TO COME===