(1) Over-reliance on unique data.
The forensic evaluator must remember to look at all areas relevant to their decision (i.e., demographic information, personality, social components, environmental concerns). However, one must remember not to get stuck on one key area that they consider highly correlated with criminal behavior, etc. While there is some 'red flags', no single piece of data should be used to make a specific conclusion.
(2) Confusing facts and statistical artifacts
We must remember that simply by adding a bunch of extra tests to an evaluation we are not necessarily learning more or providing more support for our decisions. If the tests that are added are highly correlated to each other they will clearly provide a similar response. The new information does not 'double' the amount of knowledge it simply reiterates what we already were aware of. The new information can simply be considered a statistical artifact.
(3) Confirmatory bias.
When working with specific populations like sex offenders it is easy to go in assuming that this person is a high risk offender. While they may very well be, if we go in with expectations of the persons dynamic we will be biased towards finding information that proves what we believe prior to the assessment. To avoid a confirmatory bias one should always take the opposite hypothesis of their expectations and try to refute it.
(4) Over confidence.
Research has shown that there is no positive relationship between confidence in ones evaluation conclusions and their accuracy. In fact, there is evidence to suggest that those who are exceptionally confident are at higher risk of being incorrect. The reason is not that we are invalid evaluators but rather a artifact of polarization. When a person is not very confident with their evaluation they may look for more and more reasons why they personally should believe in the utility of their evaluation, and thereby overconfidence occurs.
(5) Potential limitation in complex analysis.
When using a complex statistical and psychometric evaluation it can be difficult to summarize the pieces of individual data that we have. Small improper generalizations based on this large amount of data can occur. And with a long evaluation these overgeneralizations can add up.
Go To Page: 1 2
