This paper reviews the methods and findings of published research on the validity of police selection procedures. As a preface to the review, the typical police officer selection process is briefly described. Several common methodological deficiencies of the validation research are identified and discussed in detail: (1) use of past-selection research designs; (2) inappropriate comparison groups; (3) non-meaningful outcome variables; (4) alpha-inflated analysis; (5) over-emphasis of beta weights; and (6) the search for moderator variables. Validity evidence for several types of selection variables is discussed including biodata, measures of intellect, personality measures, interviews, interest inventories, and subjective background ratings. Of the 14 biodata categories researched only 5 were validated as predictors of poor police performance (prior involuntary termination, criminal and vehicle code convictions, having been married more than once, and short duration of prior jobs). Measures of intellect, subjective background ratings, and personality measures provided mixed evidence of validity. Some scales of the Minnesota Multiphasic Personality Inventory were found to have post-selection validity in more than one study. There was no meaningful evidence of validity for interest inventories or interviews as police selection procedures. (MCF)