Day 9- Selection

Question 1

Reliability

Answer 1

The degree to which interveiws, tests, and other selection procedures yield comparable data over time and alternative measures

If not reliable cannot be valid

Question 2

Validity

Answer 2

“the degree to which available evidence supports inferences made from scores on selection measures. Validity is not an attribute of the measure itself” (it is measuring what it is supposed to measure)

If it is not reliable it cannot be valid

Question 3

Why do Reliability and Validity matter?

Answer 3

Increase the probability of hiring a high performer—in terms of both task and contextual performance

Selection systems are legally defensible

Ethical reasons- selection decisions are high stakes; they have a profound impact on the lives of applicants. (ensure the systems are fair and valid)

Question 4

Reliability in eqution form

Answer 4

Xobtained = Xtrue + Xerror

True score- actual score
error- meaasurement error
How much of the obtained score is made up of error and true

Question 5

Methods of Estimating Reliablity

Answer 5

Test-retest: give test, wait, then retest- similar results?

Internal consistency: Conbachs alpha
Where the questions straight forward? Interpreted the same way?

Inter-rater Reliablity: 
Panel type interviews
-Intraclass correlation

Question 6

What is Validity?

Answer 6

Evidence about the extent to which inferences can be made from scores on an assessment

Appropriateness, meaningfulness, and usefulness of the specific inference made from test scores (APA, 1985)

Evidence that the test is measuring OR predicting what it’s supposed to be measuring or predicting

Will the score actually relate to the job performance? Supposed to predict job performance and does.

Question 7

Types of Validity

Answer 7

Criterion-related validity

Content validity

Question 8

Criterion-related validty

Answer 8

concurrent
-assessing current jobs and comparing it to their current job performance (least time consuming and inexpensive)
Issues: new hires might work harder then current employees, job applicants will respond differently (more motivation and attention to detail) than current employees.

predictive

• trying to predict the job performance (more time consuming) give test then see how they perform after.
• these are the best*

Question 9

Content validty

Answer 9

Does assessment contain a representative sampling of the content domain of the job?
Experts’ systematic judgment
No statistical measure of content validity

Question 10

Approaches to Validation

Answer 10

Cross-validation
Verifying the results obtained from a validation study by administering a test or test battery to a different sample (drawn from the same population) (more certainty that test is valid

Validity Generalization
The extend to which validity coefficients can be generalized across situations (over different populations)

Question 11

Best predictors of Job Performance (tests)

Answer 11

• Work sample tests
• IQ tests and structured interviews
• job tryout procedure
• personality test

Question 12

Selection Decision Models

Answer 12

Compensatory Model
Multiple Cutoff Model
Multiple Hurdle Model

Question 13

Selection Decision Models:

Compensatory Model

Answer 13

Permits a high score in one area to make up for a low score in another area

Question 14

Selection Decision Models:

Multiple Cutoff Model

Answer 14

Requires an applicant to achieve a minimum level of proficiency on all selected dimensions.

Question 15

Selection Decision Models:

Multiple Hurdle Model

Answer 15

Only applicants with sufficiently high scores at each selection stage go on to subsequent stages in the selection process.

Question 16

Selection Ratios and Cutoff Scores

Hits and misses

Answer 16

Predictor cutoff- interview
Criteria- job standards, what level do with find sufficeint

Hired and perform above cutoff= A, Correct Acceptance

Not Hired and would not do well = C, correct rejection

Hired and suck = B, erroneous acceptance

Not hired but would have done great = D, Erroneous rejection