Search

For 1,610 classes, mutually exclusive contributions of faculty and course to over-all evaluation of instruction were 45% and 15%, respectively, the remainder of the response variation being due to interaction and error.

The overall objectives of the subproject are to develop career paths that enable junior officers to make long-term career plans and to assist the Navy in developing assignment strategies that increase career retention of quality Naval officers. The study was undertaken to identify patterns in the duty assignment system that are associated with retention. If patterns are identified that are controllable through the assignment system alternative strategies that increase retention may be developed. Configural (or pattern) analysis models were evaluated for accuracy and stability in providing estimates of retention proportions for a source-to-assignment matrix with several hundred possible assignment patterns. The major finding was that one of the SPA models provided more stable data than did the calculations based on the actual outcomes. This finding suggests that stable estimates of personnel retention proportions are possible for use with a source-to-assignment matrix in algorithms for optimizing the assignment of personnel.

In the context of both the Fisher and the Neyman-Pearson formulations, this article develops a method of testing superiority of fit to determine which of two different numerical predictions is more nearly accurate. Based on the assumption that the two predictions are not equally nearly accurate, the method is sequential continuing until a decision is reached in favor of one prediction or the other. The error probability of this decision is specifiable in advance of the test.

Attempting to take a nonadversarial approach to the dispute of Slack and Porter with Jackson over the predictive validities of the Scholastic Aptitude Test and the high school record, this report examines both the effect of the selection process on these validities and what these validities might be if all colleges were to use a common standard of academic achievement. The results indicate that for this criterion the Scholastic Aptitude Test may have a predictive validity of .62 and an incremental validity over the high school record of .17 and for the usual criterion the selection process may have a profound effect on the relative predictive validities of the Scholastic Aptitude Test and the high school record. The conclusion thus reached is that the Scholastic Aptitude Test is, as intended by its developers, a valid instrument of selection.

In contrast to the Cleary-McNemar view affirmed by Cole in the October 1981 issue of the American Psychologist on testing--questions of bias are fundamentally question of validity--this report shows that freedom from statutory test bias, as interpreted by the courts, is different from predictive validity. Use of a score-adjustment formula developed here to correct for statutory test bias shows in typical cases not only that the correction tends only negligibly to reduce predictive validity but also that the enhancement of predictive validity without regard to statutory test bias can add a sizable criterion-independent decrement selectively to the already low test scores of low-scoring demographic groups. (Author)

In contrast to the Cleary-McNemar view affirmed by Cole in the October 1981 issue of the American Psychologist on testing--questions of bias are fundamentally question of validity--this report shows that freedom from statutory test bias, as interpreted by the courts, is different from predictive validity. Use of a score-adjustment formula developed here to correct for statutory test bias shows in typical cases not only that the correction tends only negligibly to reduce predictive validity but also that the enhancement of predictive validity without regard to statutory test bias can add a sizable criterion-independent decrement selectively to the already low test scores of low-scoring demographic groups. (Author), http://archive.org/details/racialbiaspredic00weit

In sequential testing for selection, an applicant for school or work responds via a computer terminal to one item at a time until an acceptance or rejection decision can be made with a present probability of error. The test statistic, as a function of item difficulties for standardization subgroups scoring within successive quantities of the criterion, is an approximation of a Waldian probability ratio that should improve as the number of quantities increases. Monte carlo simulation of 1,000 first-year college students under 96 different testing conditions indicated that a quantile number as low as foud could yield observed error rates that are close to their nominal values with mean test lengths between 5 and 47. Application to real data, for which interpolative estimation of the quantile item difficulties was necessary, produced, with quantile numbers of four and five, even more accurate observed error rates than the monte carlo studies did. Truncation at 70 items narrowed the range of mean test lengths for the real data to between 5 and 19, Important for use in selection, the critical values of the test statistics are functions not only of the nominal error rates but also, alternatively, of the selection ratio, the base-rate success probability, and the success probability among selectees, which a test user is free to choose.

This thesis investigates the type of duty assigned--sea or shore--at the point of separation for those non-prior-service (NPS) males who left the Navy before expiration of their initial enlistment contract, using a fiscal year's cohort of accessions drawn from the Navy Enlisted Tracking File (STF). Demographic and organizational attributes of cohort members are described, as well as characterizations of service and reasons for loss of those who left prematurely. Previous research using situational variables suggested the possibility that initial assignment to shore duty after training might be a higher attrition risk than assignment to sea duty for NPS males. The data for this cohort indicate that the predominance of shore processing activities as separating commands obscure the question of whether a premature loss should be attributed to a sea or shore command, and that assumptions about initial assignment to shore duty based on type-duty variable are suspect because of the probability that processing, holding and corrections commands are included under the designation of shore commands., http://archive.org/details/adescriptivenaly1094519725, Lieutenant Commander, United States Navy, Approved for public release; distribution is unlimited.