Repeated Measures of Information Processing

Abstract

Two information processing tasks were considered for inclusion in a test battery which was being developed for repeated measures investigations of adverse environmental effects. The tests, adapted from the Rose Battery were: Baron's Graphemic and Phonemic Analysis, and Posner's Letter Classification. Alternate forms of the the tests were administered on 15 consecutive workdays (Monday to Friday) to 21 Navy enlisted men. A total of 23 measures were taken. These scores were examined to determine when in practice they obtained unchanging or linearly changing means, homogeneous variances, and constant (differentially stable) intertrial correlations of an acceptably high level (task definition). In general, correlations of the basic measures tended to become stable with sufficient practice, but derived measures such as difference, slope, and ratio scores did not attain stability. The Baron and Posner tasks had high reliabilities and were highly correlated with each other. Preliminary analysis indicated that both tests may be measuring the same thing. Either the Baron SN or Posner Name score, test may be recommended for repeated measures experimentation.