WHEN A PERSONALITY INVENTORY IS ADMINISTERED TO AN INDIVIDUAL ON TWO OCCASIONS, SOME ITEMS MAY NOT ELICIT THE SAME RESPONSE ON BOTH OCCASIONS AND THE INDIVIDUAL'S SCALE SCORES MAY DIFFER BETWEEN THE TWO ADMINISTRATIONS. TRADITIONALLY, SUCH INTRAINDIVIDUAL SCORE VARIABILITY HAS BEEN AVERAGED ACROSS A SAMPLE TO FURNISH ONE TYPE OF ESTIMATE OF THE RELIABILITY OF THE SCALE. THIS AND OTHER STANDARD PSYCHOMETRIC PROCEDURES FOR ERROR ESTIMATION REST ON THE IMPLICIT ASSUMPTION THAT INTRAINDIVIDUAL RESPONSE VARIABILITY TO THE SAME SET OF ITEMS ARISES PRIMARILY FROM CHARACTERISTICS OF THE ITEMS THEMSELVES, AND NOT FROM ANY PARTICULAR ATTRIBUTES OF THE SAMPLE. THE AIM OF THE PRESENT STUDY WAS TO DISCOVER WHETHER THE HYPOTHESIS THAT THERE ARE NO DIFFERENCES BETWEEN INDIVIDUALS IN THEIR GENERAL LIKELIHOOD OF ELICITING CONSISTENT VS. INCONSISTENT RESPONSES IS REASONABLE, OR ON THE OTHER HAND, WHETHER INTRAINDIVIDUAL CONSISTENCY IN RESPONSES TO STRUCTURED PERSONALITY INVENTORIES COULD BE USEFULLY CONCEPTUALIZED AS A PERSONALITY TRAIT. THE EXPERIMENTAL DESIGN USED THREE PERSONALITY INVENTORIES--THE CALIFORNIA PSYCHOLOGICAL INVENTORY (CPI), THE EDWARDS PERSONAL PREFERENCE SCHEDULE (EPPS), AND THE MINNESOTA MULTIPHASIC PERSONALITY INVENTORY (MMPI)--WHICH WERE ADMINISTERED ON TWO SEPARATE OCCASIONS. ALL INVENTORIES WERE ADMINISTERED IN CLASS, 1 EACH WEEK, TO 95 MALE AND 108 FEMALE UNIVERSITY OF OREGON UNDERGRADUATE STUDENTS ENROLLED IN AN INTRODUCTORY PSYCHOLOGY COURSE. A TOTAL OF 15 DIFFERENT SUBSETS WERE USED TO OBTAIN MEASURES OF RESPONSE VARIABILITY VS. CONSISTENCY. THE AVERAGE SUBJECT IN BOTH SAMPLES RESPONDED CONSISTENTLY TO ABOUT 80 PERCENT OF THE EPPS DUPLICATED ITEMS, 90 PERCENT OF THE CPI DUPLICATED ITEMS, AND 95 PERCENT OF THE MMPI DUPLICATED ITEMS. DETAILED ANALYSES OF THE FINDINGS ARE REPORTED AND ABUNDANT STATISTICAL DATA ARE PROVIDED. THE STUDY CONCLUDED THAT EVIDENCE REGARDING A TRAIT OF RESPONSE CONSISTENCY IS EQUIVOCAL. WHILE CONSISTENCY MEASURES CAN BE CONSTRUCTED WHICH POSSESS SUBSTANTIAL HOMOGENEITY AND A SIGNIFICANT DEGREE OF CONVERGENT VALIDITY, SUCH MEASURES INEVITABLY ARE CONFOUNDED BY VARIANCE ATTRIBUTABLE TO THE PARTICULAR SET OF STIMULI (ITEMS) USED TO ELICIT THE RESPONSES, A PROBLEM SHARED WITH ALL OTHER PUTATIVE RESPONSE SETS AND STYLES. THE STUDY INCLUDES FOOTNOTES AND REFERENCES. (PRG)