Hierarchy of evidence
About
The hierarchy of evidence is a ranking of different types of evidence, to be used as a guideline for determining which evidence should be considered more credible when more than one type is available.
The following list is a first pass, and should not be considered complete, definitive, or certain – from most to least reliable:
- 1. demonstrably sound reasoning from agreed-upon facts (see rationality detection)
- 2. repeatedly verified hypothesis
- 3. verified hypothesis
- 4. physical evidence
- 5. circumstantial evidence
- 6. experience (argument from authority)
- 7. intuition
- 8. hearsay
Evidence higher on the list (lower numbers) generally trumps evidence lower on the list (higher numbers).
Each type of evidence may have a range of reliability, and it's entirely possible that more-reliable varieties of a given evidence-type should trump less-reliable varieties of a normally more-reliable evidence-type – e.g. the opinion of a professional with years of experience in a given field would normally trump circumstantial evidence that is not overwhelming, unless there is reason to suspect that the professional is biased.
Science
Scientific papers generally rank pretty high on the list -- it's fair to put a given paper at #3 by default, but bad science does exist. Science papers are fair game for criticism, like anything else, and that may be cause to downrank them. Among the many reasons that a science paper's findings may be downranked are:
- biased selection of subjects (observed trends may have been caused by how the subjects were selected and not reflect overall trends)
- small sample-size (too few data-points; results may reflect statistical fluctuation)
- bad logic (the conclusions stated are not supported by the evidence presented)
It's also important to distinguish between science papers and science reporting. The latter often exaggerates the significance of findings (in order to attract readers) or even blatantly misunderstands the subject-matter.
Further, it should be noted that there are a handful of subjects over which the science is very settled (coming in at #2) and yet over which there is or has been political controversy, often claimed to be scientific. For example:
- global warming
- evolution by natural selection
- the shape of the earth
Flat Earth theories are an especially good example of how "controversy" can be generated despite the most direct and inarguable evidence plus literal millennia of accumulated observations and less-direct evidence all clearly pointing towards the same conclusion.