MENTAL MEASUREMENTS YEARBOOK (MMY) RELIABILTY: stable & error free o Test-Retest (0.86) Temporal Stability o Alternate Forms (0.89) Counterbalancing Form stability o Internal Reliability (0.81) Item stability o Interrater / interscorer VALIDITY: o Content Validity: concerned -> test construction CVR o Criterion Validity Concurrent Predictive Single-group Validity: predicts for one not for the other Validity Generalization o Construct Validity -> concerned about inferences Correlating scores vs scores from other tests Known-group validity: known to be different o Face Validity Appearance Barnum statements: general statements; true to anyone COST-EFFICIENCY ESTABLISHING USEFULNESS OF SELECTION DEVICE TAYLOR-RUSSELL TABLES (OVERALL) o Who will be successful if using this particular test Criterion Validity Coefficient Base Rate Selection Ratio (# positions/# of applicants) PROPORTION OF CORRECT DECISIONS (OVERALL) o Easier less accurate Scores on criterion Test scores o X axis (scores), Y axis (criterion score) o Quadrant I & III: hit o Quadrant II &IV: miss o Points in Q II & IV / Total Points (percentage of time before we become accurate) o Points in Q I & II / Total Points (check improvement) First is higher than second = chose the new LAWSHE TABLES (APPLICANT) o Applicant will be successful Base rate Applicant score Validity coefficient BROGDEN-CRONBACH-GLESER FORMULA o How much an org will save if it uses this test? DETERMINING FAIRNESS OF A TEST Bias: technical Fairness: Adverse Impact: less than 80% Single Validity Differential Validity: valid to both but more valid to one MAKING THE HIRING DECISION SELECTION VS PLACEMENT NEPOTISM QUALIFIED WORKFORCE MULTIPLE REGRESSION TOP DOWN APPROACH COMPENSATORY APPROACH: low score can be compensated RULE OF THREE PASSING SCORES: Who will be able to perform at acceptable level o Multiple Cut-off: high positions o Multiple Hurdle: entry level o Composite Score BANDING o Compromise bet. Top down and passing o Anyone w/in the band (depends on SE)