Automated scoring models for the "e-rater"® scoring engine were built and evaluated for the "GRE"® argument and issue-writing tasks. Prompt-specific, generic, and generic with prompt-specific intercept scoring models were built and evaluation statistics such as weighted kappas, Pearson correlations, standardized difference in…

Twelve common errors which can invalidate an otherwise sound evaluation are identified, and ways to avoid them are presented. The hazards are: (1) grade-equivalent scores; (2) inappropriate statistical adjustments with nonequivalent control groups; (3) administering norm-referenced tests at inappropriate times of the school year; (4) inappropriate…