Field Driving Test Situations for Ml Tank Drivers

Abstract

The presentation deals with the testing of Ml tank drivers and focuses on the requirement to establish reliable measures of nonprocedural portions of the tank driver's job. Tests of tank driving skills have been difficult to develop because of situational reliability problems. Actual conduct of the tests with 60 ton tanks change the driving terrain for each subsequent subject. Further, problems of rater reliability in the use of observational ratings and checklists have been encountered. The purposes of the present study were (a) to develop test methodologies for tank driving that would enable these problems to be overcome and (b) to evaluate these methodologies by developing and assessing a battery of on–tank tests of nonprocedural, tactical Ml tank driving skills.

Eleven driving tasks for which tests were to be developed were derived from an ARI criticality survey. Analysis of the tasks resulted in decisions to test nine of the tasks, but only those aspects that related to the driver and that were feasible for testing. Five of the tests were Obstacle/Judgment tests, and four were Tactical tests. The tests were tried out on 77 soldiers in two Ml OSUT classes (none of the soldiers took all the tests). The data were used to assess scorer agreement and internal consistency, to estimate utility based on reliability and variability, and to direct revisions and recommendations for future testing.

For each of the nine tests, the data indicated that driver performance could be measured reliably. Both the Obstacle/Judgment tests and Tactical tests had been designed so that usable quantitative data could be obtained, and for every test, refinements were suggested based on data and on informal observations. For two of the Tactical tests, an innovative scoring technique using an Ml tank profile overlay was explored. Despite high ratings in the criticality survey, one Obstacle/Judgment test was recommended for deletion. While more replications and developmental refinements of the tests are needed, the analysis and development performed have produced tests that are already minimally reliable. The tests were designed for use in measuring on–tank driving performance against which to assess driver simulator training. They may also be useful in general field applications for training.