The construction and validation of a simplified measurement analysis model for item selection and interpretation of scores

The purpose of the study was to construct and validate a simplified Measurement Analysis Model that could be used by a teacher in analyzing teacher-made tests in the classroom. The model was developed in an attempt to give teachers the item analysis tools of the measurement specialist without requiring a high degree of computational or statistical skill. A systematic review of the literature was made in three areas: (1) main constructs in measurement theory, (2) major components used in standardized item analysis procedures, and (3) procedures currently available for estimating major item analysis components. Systematic sampling at specific percentile values, and the centroid of the area sampled from the extremes in terms of the normal distribution, formed the foundation for developing the Measurement Analysis Model. Eighteen sampling patterns were developed and analyzed. These sampling patterns involved sampling from the extremes of distributions which had been rank ordered, stratified sampling at specific points in a rank ordered distribution, and combinations of the above. The sampling patterns were used with skewed populations and class sizes ranging from 20 to 60 with absolute deviations calculated between sample estimates and actual values of major item analysis components obtained by a standardized item analysis procedure. [...]