How accurately deltas (statistics measuring the difficulty of items) established by pre-test populations reflect deltas obtained from final form populations, and the consequent utility of pre-test deltas for constructing final (operational test) forms to meet developed statistical specifications were studied. Data were examined from five subject areas of the Admissions Testing Program (ATP) Achievement Tests (ACH): (1) humanities; (2) mathematics; (3) sciences; (4) social sciences; and (5) foreign languages. The stability of pre-test statistics is evaluated by examining: summary statistics of deltas obtained from pre-test and final form populations; bivariate plots comparing deltas derived from the two populations; the values of the two deltas per item via difference plots; item position shifts; the composition of population samples; and deviation from linearity in the data studied. The research shows that despite the problems inherent in pre-testing on college samples (e.g., the sampling method, motivational factors, relevance of items to course content, method of presentation, item position shifts, and in some cases, inadequate sample sizes), pre-test equated deltas are relatively successful predictors of final form equated deltas. Ten tables and 10 graphs are appended. (SLD)