A DESCRIPTION IS GIVEN OF ITEM SAMPLING, OR "PSYCHOMETRIC-STATISTICAL INFERENCE," AN APPROACH TO GATHERING AND USING EDUCATIONAL DATA THAT ALLOWS STATISTICAL INFERENCES TO BE MADE SIMULTANEOUSLY WITH PSYCHOMETRIC INFERENCES. THIS IS A PROCEDURE IN WHICH BOTH PEOPLE AND ITEMS ARE SAMPLED AND THE DATA FROM A SAMPLE OF PEOPLE TAKING A SAMPLE OF ITEMS IS USED TO DRAW INFERENCES ABOUT THE PERFORMANCE OF A POPULATION OF SUBJECTS TAKING A POPULATION OF ITEMS. FIVE GENERAL CASES FOR EDUCATIONAL RESEARCH ARE PRESENTED TO INCLUDE MOST POSSIBLE USES OF ITEM SAMPLING--(1) STATISTICAL OR PSYCHOMETRIC INFERENCE (SIMPLE ITEM OR PEOPLE SAMPLING), (2) SIMPLE STATISTICAL-PSYCHOMETRIC INFERENCE (SIMPLE ITEM AND PEOPLE SAMPLING), (3) STATISTICAL-PSYCHOMETRIC INFERENCE WITH TWO SUBJECT SAMPLES AND ONE ITEM SAMPLE (SIMPLE HYPOTHESIS TESTING), (4) STATISTICAL-PSYCHOMETRIC INFERENCE WITH MULTIPLE ITEM SAMPLING AND MULTIPLE SUBJECT SAMPLING, AND (5) STATISTICAL-PSYCHOMETRIC INFERENCES WITH MULTIPLE ITEM SAMPLING AND MULTIPLE SUBJECT SAMPLING IN THE CASE OF TWO SUBJECT POPULATIONS (HYPOTHESIS TESTING). THESE FIVE CASES BEGIN WITH THE MOST SIMPLE SITUATIONS AND BECOME MORE COMPLEX, AND THE FIRST THREE USE ITEM SAMPLING ONLY SUPERFICIALLY. THE AUTHORS NOTE THAT ITEM SAMPLING PROVIDES A TECHNIQUE FOR OBTAINING EVALUATIVE DATA ON CLASS OR GROUP PERFORMANCE, NOT INDIVIDUAL PERFORMANCE, AND CAN BE ESPECIALLY EFFECTIVE WITH CHANGE OR GROWTH STUDIES. (JH)