|
After the data are converted to a common statistic, reliabilities and range
departure information are collected. If all the studies include reliability
estimates, or range departure data, then the effect size can be corrected
at the individual study level. However, most studies do not provide this
information. Fortunately Hunter and Schmidt (1990) and Hunter et al (1982)
provide procedures for estimating the corrections for reliability and range
departures by constructing distributions for the independent and dependent
variables.
When the literature base is assembled, and reliability and range information
are collected, the next step is to eliminate the bias caused by sampling
error. Sampling Error refers to the random variation in study results
due to sample size. Smaller sample sizes tend to vary more widely from the
true relationship within the population than do studies that have large
samples. Accordingly by weighting the study effect size by its sample size,
will give a more accurate approximation to the relationship within the population,
unaffected by the size of the sample.
|
