First, as a physicist, I do have to point out that this article concerns mainly softer sciences, e.g. psychology, medicine, etc.
It seems to me that biology and medicine are “softer” sciences than chemistry or physics, not due to the subject matter, but mainly because the scientists behave in softer, less rigorous ways—such as the explanations suggested for this article’s findings.
Personally, I see them as being softer because they’re about less fundamental, i.e. higher order, systems. Experiments and theories in biology and medicine treat high-level concepts (e.g. organs, tissues, symptoms of diseases) as basic. This is somewhat of a necessity, because the tools to model such things as complex collections of and interactions among the component particles don’t exist yet.
However, when they treat conclusions drawn from those high-level approximations as universal, problems inevitably arise. We might all have organs that seem the same, and are constructed by the same cellular machinery in large part, but there’s still a lot of room for variation. Even subtle variations might have effects that wouldn’t be obvious from theories based on high-level approximations. This is where the lack of rigor is a factor: even if scientists understand that they’re using high-level approximations, they may not want to admit a lack of universality in a result. (I have noticed this to be especially true in psychology.)
I think the focus on universal theories is a shame; something can be perfectly worthwhile even if it’s only useful in some circumstances. Many theories in physics are like this; the usual example is, of course, Newtonian mechanics vs. Einsteinian relativity. A description of some biological or psychological phenomenon can still be useful even if it doesn’t apply to every possible organism/brain. There is, however, the problem that determining the set of circumstances in which such a description might be useful can also be more difficult due to the high-level, complex nature of the subjects.
An example of a theory in psychology I find to be useful even if not universal is Kohlberg’s stages of moral development. Kohlberg himself was quite convinced the stages were universal, and made a few contortions to try to keep it that way. Given that the original theory was formulated based solely on interviews with males in the USA, a limited range of applicability – based largely on culture and to a smaller degree gender, especially as one looks at the higher stages – is more likely. However, the theory is still a good way to understand how people in the USA think about moral issues; it just might not be as good a way to understand how people in other cultures and social settings think.
Maybe scientists in soft fields behave differently, but this correlation with other notions of “soft” needs to be explained. It could be a founder effect, but that doesn’t seem plausible to me.
“If your experiment needs statistics, you ought to have done a better experiment.”—Rutherford
One problem is that soft sciences need statistics or at least make them more tempting. I don’t think it’s useful to phrase it in terms of blame, that the soft scientists were less able to resist this temptation.
(I agree with everything the Dreaded Anomaly said, too.)
It seems to me that biology and medicine are “softer” sciences than chemistry or physics, not due to the subject matter, but mainly because the scientists behave in softer, less rigorous ways—such as the explanations suggested for this article’s findings.
What do you think?
Personally, I see them as being softer because they’re about less fundamental, i.e. higher order, systems. Experiments and theories in biology and medicine treat high-level concepts (e.g. organs, tissues, symptoms of diseases) as basic. This is somewhat of a necessity, because the tools to model such things as complex collections of and interactions among the component particles don’t exist yet.
However, when they treat conclusions drawn from those high-level approximations as universal, problems inevitably arise. We might all have organs that seem the same, and are constructed by the same cellular machinery in large part, but there’s still a lot of room for variation. Even subtle variations might have effects that wouldn’t be obvious from theories based on high-level approximations. This is where the lack of rigor is a factor: even if scientists understand that they’re using high-level approximations, they may not want to admit a lack of universality in a result. (I have noticed this to be especially true in psychology.)
I think the focus on universal theories is a shame; something can be perfectly worthwhile even if it’s only useful in some circumstances. Many theories in physics are like this; the usual example is, of course, Newtonian mechanics vs. Einsteinian relativity. A description of some biological or psychological phenomenon can still be useful even if it doesn’t apply to every possible organism/brain. There is, however, the problem that determining the set of circumstances in which such a description might be useful can also be more difficult due to the high-level, complex nature of the subjects.
An example of a theory in psychology I find to be useful even if not universal is Kohlberg’s stages of moral development. Kohlberg himself was quite convinced the stages were universal, and made a few contortions to try to keep it that way. Given that the original theory was formulated based solely on interviews with males in the USA, a limited range of applicability – based largely on culture and to a smaller degree gender, especially as one looks at the higher stages – is more likely. However, the theory is still a good way to understand how people in the USA think about moral issues; it just might not be as good a way to understand how people in other cultures and social settings think.
Maybe scientists in soft fields behave differently, but this correlation with other notions of “soft” needs to be explained. It could be a founder effect, but that doesn’t seem plausible to me.
“If your experiment needs statistics, you ought to have done a better experiment.”—Rutherford
One problem is that soft sciences need statistics or at least make them more tempting. I don’t think it’s useful to phrase it in terms of blame, that the soft scientists were less able to resist this temptation.
(I agree with everything the Dreaded Anomaly said, too.)