THE OBJECTIVES OF THIS STUDY WERE (1) TO DEVELOP A GENERAL CLASSIFICATION SYSTEM FOR LIFE HISTORY ITEMS, (2) TO DETERMINE TEST-RETEST RELIABILITY ESTIMATES, AND (3) TO ESTIMATE RESISTANCE TO EXAMINEE FAKING, FOR REPRESENTATIVE BIOGRAPHICAL QUESTIONNAIRES. TWO 100-ITEM QUESTIONNAIRES WERE CONSTRUCTED THROUGH RANDOM ASSIGNMENT BY CONTENT AREA OF 200 SELECTED ITEMS TAKEN FROM SEVERAL BIOGRAPHICAL QUESTIONNAIRES. THESE ITEMS WERE THEN RANDOMLY REARRANGED INTO NEW PRESENTATION ORDERS TO FORM TWO ADDITIONAL QUESTIONNAIRES FOR RETEST PURPOSES. EACH STEM-ALTERNATIVE COMBINATION OF THE 200 TOTAL ITEMS (MULTIPLE-CHOICE) WAS RATED ON A 9-INTERVAL SOCIAL DESIRABILITY SCALE. USING RATING VARIATION COEFFICIENTS WITH REGARD TO SOCIAL DESIRABILITY, THE INVESTIGATOR ATTEMPTED TO ESTABLISH A SINGLE ITEM CLASSIFICATION SYSTEM WITH THE GREATEST CRITERION PREDICTION. TEST AND RETEST ADMINISTRATIONS WERE SEPARATED BY A TIME PERIOD OF 10 TO 14 WEEKS. SUBJECTS WERE UNDERGRADUATE MALES IN EDUCATION, PSYCHOLOGY, AND SOCIOLOGY. BOTH NORMAL AND ARTIFICIAL INSTRUCTIONS WERE USED WHEN THE TESTS WERE ADMINISTERED, IN ORDER TO ELICIT STUDENT RESPONSES WHICH COULD BE USED TO OBTAIN ACCURATE RELIABILITY COEFFICIENTS AND "FAKEABILITY" SCORES. RESULTS SHOWED TEST-RETEST RELIABILITY COEFFICIENTS NOT TO BE HIGH, BUT MUCH LARGER THAN CORRESPONDING COEFFICIENTS OBTAINED BETWEEN NORMAL AND ARTIFICIAL (FAKE) SET CONDITIONS. (JH)