Project A was the name applied by the U.S. Army to its contribution to the Joint-Service Job Performance Measurement/Enlistment Standards (JPM) Project sponsored by the Department of Defense (DoD) in 1982. Lasting until 1989, Project A—which now also comprises the follow-on Career Force project (examining performance during soldiers’ second tours of duty and spanning 1990-1994)—is arguably the largest selection and classification project ever conducted. Results from Project A continue to shape the way the Army selects, assigns, trains, and promotes its soldiers. Project A data remain a rich storehouse of information about individual differences and job performance. Project results and data have also had an impact on the field of industrial/organization (I/O) psychology at large, having sparked theoretical developments regarding models of job performance (contextual performance, determinants of relevant variance), a resurgence in interest regarding personality tests as selection tools, and development of models for setting recruit enlistment standards.
The Joint-Service Job Performance Measurement/Enlistment Standards Project
The JPM Project was a congressionally mandated multimillion-dollar effort that spanned 1982 to 1994. The impetus for JPM was the discovery of a scoring error that inflated the scores of lower-aptitude recruits on the Armed Services Vocational Aptitude Battery (ASVAB). This miscalibration of ASVAB scores led to the enlistment of approximately 250,000 individuals who otherwise would not have qualified for entrance, and increased congressional concerns about recruit quality.
In response to the miscalibration, the DoD initiated the JPM Project. A primary goal of JPM was to determine if hands-on job performance could be measured. If so, the DoD could set enlistment standards on the basis of that job performance information. Previous standards were tied to training success rather than job performance, and the ASVAB was validated against training performance but not against performance on the job in the field.
The DoD encouraged each branch of the armed forces to conduct its own research for JPM. The U.S. Army Research Institute (ARI) sponsored the Army’s effort. The Army’s approach was ambitious, expanding the predictor and criterion domains by developing new entry-level selection and performance measures. The selection project was named Project A to distinguish it from the classification research effort Project B.
Project A included both concurrent and longitudinal validation samples. The concurrent validation (CV) cohort included soldiers who enlisted during the 1983 and 1984 fiscal years. These soldiers completed the new selection and performance measures at the same time, approximately two years into their first tours of duty (CVI sample). Those soldiers who reenlisted were eligible to complete measures of second-tour (supervisory) performance (CVII sample). Soldiers in the longitudinal validation (LV) cohort enlisted during the 1985 and 1986 fiscal years. They received the Experimental Battery during their first two days in the Army (LVP sample), measures of training performance at the end of technical training for their jobs (LVT sample), and measures of job performance once in their units approximately two years after enlistment (LVI) and during their second tours (LVII).
The project collected data on soldiers from 21 military occupational specialties (MOS). The full complement of performance measures (written test of job knowledge and hands-on tests, in particular) was developed for 10 of these, which were deemed Batch A MOS. They tended to be high-density jobs of central importance to Army functioning such as infantryman, light wheel mechanic, and medical specialist. The other 11 Batch Z MOS reflected more specialized, lower-density occupations, including ammunition specialist, utility helicopter repairer, and intelligence analyst; they did not have hands-on measures developed for them.
To determine whether ASVAB scores predicted job performance, the Army developed numerous performance measures to serve as criteria for its validation and performance modeling studies. Selection studies are only as meaningful as the criteria used, and Project A addressed the traditional criterion problem head-on. This approach of obtaining multiple measures of multiple job behaviors broke with conventional notions that viewed job performance as unitary, hypothesizing instead that job performance was not a single entity but was instead a complex variable to study.
Two broad categories of criterion content were considered: performance elements that are specific to a particular job (assessed by MOS-specific measures), and performance elements that are relevant to all jobs (assessed by Army-wide measures). Criteria were also categorized as can-do measures (assessing how well soldiers are able to perform) and will-do measures (assessing how well soldiers typically perform from day to day). The criteria included written tests of MOS-specific job knowledge, hands-on performance tests (also known as work samples), various anchored rating scales (MOS-specific performance for Batch A and performance on Army-wide dimensions for all MOS), and data from administrative files such as letters of commendation and counseling statements.
In addition to evaluating whether ASVAB scores predicted job performance, the Project A research team investigated the degree to which measures of other individual differences could increase the predictive power of this test battery. Measures of spatial ability, perceptual speed and accuracy, psychomotor ability, temperament, vocational interests, and work values— collectively denoted the Trial Battery (CV sample) and Experimental Battery (LV sample)—were developed, administered to thousands of soldiers, and correlated with the various criterion measures. Results were obtained for the Army as a whole and by subgroups of interest (MOS, race/ethnicity, gender).
Results from Project A research and their implications for personnel psychology have filled journals and books. Nevertheless, some of the major findings from the project include the following:
- ASVAB is a valid predictor of performance across Army jobs and across subgroups (race and ethnicity, gender). Soldiers with higher aptitude perform better than lower-aptitude soldiers on many types of performance measures.
- Measures of other ability constructs such as spatial ability, perceptual speed or accuracy, and psychomotor ability also predict performance across jobs but provide little incremental validity to the ASVAB.
- Performance is not unidimensional but is instead a complex multidimensional construct.
- ASVAB scores predict maximal performance (can-do criteria) better than they predict typical performance (will-do criteria).
- Measures of noncognitive constructs, such as temperament, collected under research conditions predict will-do criteria better than ASVAB, and they provide substantial incremental validity to ASVAB.
- Criterion-related validity of ASVAB scores assessed longitudinally was very similar for first- (LVI) and second-tour (LVII) soldiers. Noncognitive measures, however, tended to show declining correlations over time.
- The elements of performance are similar as the soldier gains experience, but leadership emerges as a performance element in the second tour.
- First-tour performance provides more incremental validity than ASVAB when predicting leadership and effort during the second tour, but it provides less incremental validity for can-do performance criteria.
- Campbell, J. P., & Knapp, D. J. (Eds.). Exploring the limits in personnel selection and classification. Mahwah, NJ: Lawrence Erlbaum.
- Project A: The U.S. Army selection and classification project. (1990). Personnel Psychology, 43(2), 231-378.
- Zook, L. M. (1996). Soldier selection: Past, present, and future. Alexandria, VA: U.S. Army Research Institute for the Behavioral and Social Sciences.