Page:Sm all cc.pdf/168

 • consciously recognize the discrepancy and conclude that either the hypothesis or the evidence is wrong;

• consciously recognize the discrepancy, then deliberately revise the hypothesis to make it more compatible with the evidence;

• reduce the inconsistency by biased interpretation of the evidence;

• subconsciously revise the hypothesis to make it more compatible with the evidence.

All four strategies also are employed by scientists, but only the first two are valid. The first three have been discussed already and are also familiar in daily experience. Subconscious revision of a hypothesis, in contrast, is a surprising pitfall. Kuhn et al. [1988] found that subjects usually modified the hypothesis before consciously recognizing the relationship of the evidence to the hypothesis. They seldom realized that they were changing the hypothesis, so they failed to notice when their theory modification was implausible and created more problems than it solved. Fortunately for science but unfortunately for the scientist who succumbs to the pitfall of subconscious hypothesis modification, someone usually detects the error.

Kuhn et al. [1988] found that hypotheses of causal relationships between variables are particularly resistant to overthrow by new data. The new data must overcome the expectation of a correlation; even if the data set as a whole does so, nonrepresentative subsets may still appear to confirm the correlation. Furthermore, the original proposal of a causal relationship probably also included a plausible explanation. To discard the correlation is also to reject this explanation, but the new data do not even address that argument directly.

The hidden influence of accepted hypotheses on evidence evaluation harms scientists as well as science. A scientist’s beliefs may fossilize, leading to gradual decrease in creative output (though not in productivity) throughout a professional career.

As we saw in the previous section on paradigms, hidden influence of prior theory has other manifestations: (1) ignoring data inconsistent with the dominant paradigm; (2) persistence of theories in spite of disproof by data; and (3) failure to test long-held theories.

Incremental Hypotheses and Discoveries
Because the dominant paradigm molds one’s concepts, it largely controls one’s expectations. Hypotheses and discoveries, therefore, tend to be incremental changes and elaborations of the existing theories, rather than revolutionary new perspectives. Mannoia [1980] says that “the answers one obtains are shaped by the questions one asks.”

‘Fight or Flight’ Reaction to New Ideas
The expression ‘fight or flight’ describes the instinctive reaction of many animal species to anything new and therefore potentially threatening. Beveridge [1955] pointed out that ‘fight or flight’ is also a scientific pitfall. When presented with new ideas, some individuals fight: the theory is immediately rejected, and they only listen to pick out flaws. Their biased attitude should not be confused with the scientifically healthy demand, ‘show me’, suspending judgment until the evidence is heard. Other scientists flee, ignoring any new idea until more conclusive, confirming evidence can be provided. A scientist who rejects relevant evidence, on the grounds that it leaves