Test #1 - Predictors and psycho assessments

why

• improving the fit between people and their job

• includes changing the work (job redesign, org) or employees (training, selection)

predictors

• any variable used to forecast a criterion

• provide info that can be used in making HR decisions

• relationship between predictor and criterion is often represented by a correlation coef ( r )

how to use predictors

• job analysis (certain attributes are important for a job)

• group of applicants → predictor test to measure their ‘something’

• measuring → intelligence, interpersonal skills, physical ability, etc.

assessing quality of predictors (RELIABILITY)

TWO PSYCHOMETRICS CRITERIA

• reliability

• consitency, stability, or equivalence of a measure

• test retest

• equivalent form reliability

• internal consistency reliability

• inter-rater reliability

test retest

part of reliability

• measuring something at two different points in time and compare the scores

• length of time can impact, should be considered

equivalent form reliability

• develop two forms of a test to measure the same and give forms to the same group of people

• then you correlate the two scores for each person

• least popular

internal consistency reliability

• extent to which items on a test are similar (correlate to each other)

• items are similar - high internal consistency

• items cover wide variety of topics - low internal consistency

• .7-.8 is acceptable

inter-rater reliability

• degree of agreement among the assessments provided by two or more rather

• e.g multiple raters analyzing a job and multiple interviers evaluating job candidates

• evidence of high inter-rater reliability establishes a basis to conclude that the behavior was reliably observed and is accurate

assessing quality of predictors (VALIDITY)

• accuracy

• to validate a test - measn to evalute how appropriate the test is for drawing inferences about a certain criterion (job perform)

• a test is valid if it can accurately predict a criterion

• ideally, we should onyl use highly valid tests

criterion validity

• the degree to which a test measures what it claims to be measuring

• we claim this test is measuring intelligence, is it tho

• if our test really does measure intelligence, then we have high construct validity

content validity

• does the test approx. cover the content of the concept?

• SME’s evaluate your test and judge how content valid your test is

• primarily relevant in achievement testing (mastery or knowlege)

construct validity

• if the test is measuring what it intents, then people’s scores on the test should be similar to their scores (CONVERGENT VALIDITY)

• must have

  • correlate with other measures that it should be related to

  • not correlate with measure that it should be relation to

concurrent criterion

give current employees a test that is supposed to predict sales performance

face validity

does the test look like its measuring properly