Psychometrics
Reliability and validity:
Some published tests have
questionable psychometric properties. In some
cases, the reference populations are inadequate
and it is important to look at the ‘N’s before
interpreting scores with confidence. Reliability and
validity are also important considerations when
choosing a test and deciding how much weight to
place on findings. There is no objective cut-off but
the general consensus is that, in relation to reliabil-
ity, a correlation coefficient (
r
)
of greater than 0.6
on test-retest is the accepted minimum for a test
to be judged reliable. The validity of a test is mea-
sured in a number of ways, the most important of
these being construct validity (the degree to which
the test measures what it purports to measure).
Do'stlaringiz bilan baham: |