Posted by Jan den Breejen (213.75.36.2) on October 18, 2002 at 04:24:18:
Often I meet with people who reject doing psychological tests 'because they are not valid'. Of course there are very bad tests, but when using a well 'tested' test, its still something to pay attention to and reflect on.
When putting things into perspective:
++ correlation coefficient of relation between smoking an lung cancer:
r = 0,08
++ predictive power of a job interview r = 0,25
++ predictive power of a assessment center (predicting job performance)
r = 0,37
++ predictive power of (good) job aptitude tests r= 0,30
++ predictive power of weather forecast r= 0,4-0,65
++ predictive power of MMPI test towards psych. disorders 0,37
++ predictive power of astological 'forecasts' r = 0
Re-test correlation when using good psychological tests: r= 0,8-0,9
Jan