Item analysis is the process of assessing items either qualitatively (based on the criterion above) or quantitatively. In quantitative analysis the individual item is related to some criterion. The criterion may be people known to possess the psychological characteristic being measured, it may be to total score of the test, or the change as a result of some treatment (change score).
Sample data generated for this example (fake data) using the following instrument is used in this example.
Item Analysis
[Now you really are going to need syntax files. Computing the mean using the "click procedure" just doesn't work very well.]
The results follow:
This is the first set of correlations of each item with the overall score of all of the items of the subtest (SDPRESS). The the higher the correlation the better the item.
The next subtest results are for the items of the (SANXIOUS) subtest.
Then the schizophrenia subtest (SSCHIZ).
Finally the borderline subtest (SBORDERL).
The variables sdepress, sanxious, sborderl, sschiz are subtest scores (or total scores). And are obtained in the statements below like "compute sdepress =Mean(sad, hurtself, useless, worthr)." That results in a subtest score or subtest total score. In Item Analysis you correlate the item with the total (subtest) score. That is what the following syntax file accomplishes.
The item analysis scores for the respective items are:
SAD = .837, HURTSELF = .811, USELESS = .826, and WORTHR = .804.
SAD = .837, HURTSELF = .811, USELESS = .826, and WORTHR = .804.