by On the impossibility of objectivity in science Operations Research does not incorporate the arts and humanities largely because of its distorted belief that doing so would reduce its objectivity, a misconception it shares with much of science. The meaning of objectivity is less clear than that of optimality. Nevertheless, most scientists believe it is a good thing. They also believe that objectivity in research requires the exclusion of any ethical-moral values held by the researchers. We need not argue the desirability of objectivity so conceived; it is not possible. Most, if not all, scientific inquiries involve either testing hypotheses or estimating the values of variables. Both of these procedures necessarily involve balancing two types of error.
Topics:
Lars Pålsson Syll considers the following as important: Theory of Science & Methodology
This could be interesting, too:
Lars Pålsson Syll writes Kausalitet — en crash course
Lars Pålsson Syll writes Randomization and causal claims
Lars Pålsson Syll writes Race and sex as causes
Lars Pålsson Syll writes Randomization — a philosophical device gone astray
On the impossibility of objectivity in science
Operations Research does not incorporate the arts and humanities largely because of its distorted belief that doing so would reduce its objectivity, a misconception it shares with much of science. The meaning of objectivity is less clear than that of optimality. Nevertheless, most scientists believe it is a good thing. They also believe that objectivity in research requires the exclusion of any ethical-moral values held by the researchers. We need not argue the desirability of objectivity so conceived; it is not possible.
Most, if not all, scientific inquiries involve either testing hypotheses or estimating the values of variables. Both of these procedures necessarily involve balancing two types of error. Hypotheses-testing procedures require use of a significance level, the significance of which appears to escape most scientists. Their choice of such a level is usually made unconsciously, dictated by convention. This level, as many of you know, is a probability of rejecting a hypothesis when it is true. Naturally, we would like to make this probability as small as possible. Unfortunately, however, the lower we set this probability, the higher is the probability of accepting a hypothesis when it is false. Therefore, choice of a significance level involves a value judgment by the scientist about the relative seriousness of these two types of error. The fact that he usually makes this value judgment unconsciously does not attest to his objectivity, but to his ignorance.
There is a significance level at which any hypothesis is acceptable, and a level at which it is not. Therefore, statistical significance is not a property of data or a hypothesis but is a consequence of an implicit or explicit value judgment applied to them.
The choice of an estimating procedure can also be shown to require the evaluation of the relative importance of negative and positive errors of estimation. The most commonly used procedures are “unbiased”; therefore, they provide best estimates only when errors of equal magnitude but of opposite sign are equally serious — a condition I have never found in the real world.