1
Q
z scores
A
- will have a mean of 0 and a SD of 1
- more positive or negative scores are farther aware from the mean and are more extreme
- z-score = (score-mean)/standard deviation
- part of distribution model
2
Q
why are measurement and statistics important to staffing?
A
- understanding how staffing and talent contribute to growth or change is one of the most important goals of talent analytics
- predicting trends can give insight into where a firm can generate returns
3
Q
4 scales of measurement
A
- nominal
- ordinal
- interval
- ratio
4
Q
how are correlations used in selection to estimate criterion-related validity?
A
- the empirical relationship between predictor and outcome variables is usually estimated using correlation
- can use either concurrent or predictive relationship
- concurrent is between predictor and outcome variables when both types of data are collected on job incumbents
predictive is between predictor and outcome variables when data is collected on job applicants and outcome data is collected after some of these applicants are hired and become job incumbents
5
Q
what is selection?
A
- refers to the strategies, processes, methods, and practices used to assess and hire applicants on job and organization related KSAOs
- strategically valuable
- goal is to identify those applicants that will fit with the job, org., and national culture
- using selection practices/methods that best identify talent and are unique to one’s firm and culture can create talent resources that generate competitive advantage
6
Q
basic logic of selection
A
- we make a hiring decision because the KSAOs assessed with the predictor are linked to effective performance on critical tasks
- essentially make a bet on the future
7
Q
validity
A
- the degree to which the inferences we make from predictor scores are appropriate, accurate, meaningful, and useful
- predictor: measures of KSAOs
- criterion: the thing we are trying to predict (ex: job performance, sales, turnover, etc.)
- reliability sets the limit on validity
8
Q
selection process
A
9
Q
predictor KSAOS and methods
A
10
Q
the best predictor of future behavior is _______ ______
A
11
Q
biodata predictor
A
- biographical data
- best predictor of future behavior is past behavior
- good reliability and moderate validity
- identify past history items that maximize performance on the criterion
- more verifiable items reduce faking
- requiring elaboration reduces faking (ex: “how many languages do you speak? list them”)
12
Q
personality predictor
A
- consistency in behavior
- five factor model (big five)
- neuroticism (emotional stability)
- extraversion
- agreeableness
- openness to experience
- conscientiousness
13
Q
cognitive ability predictor
A
14
Q
interview predictor
A
- most commonly used selection method
- many are done poorly
- interview can be effective is properly developed and administered
- effective interviews are structured (ask the same questions of everyone, minimize use of prior info, use behavioral rating scale, minimize deviations from questions, train interveiwers)
- structured interviews are the most valid predictor method
15
Q
simulation predictor
A
16
Q
cognitive ability
A
17
Q
interviews
A
18
Q
simulations
A
19
Q
selection decision making
A
20
Q
nominal
A
- nominal - when scores represent simple categories by provide no additional info (ex: gender, race, country, etc.)
t
21
Q
ordinal
A
- ordinal - provide slightly more info and tell us about category membership and rank order (ex: forced ranking performance assessment programs, school grades, top ten lists, etc.)
22
Q
interval
A
- interval - takes into account both rank order and ensures differences between each rank order is the same (ex: 5 point or 7 point scales used in surveys/performance management systems)
23
Q
ratio
A
- ratio - has all the elements of the ordinal scales but includes a true zero point (ex: money, temperature, or weight)
24
Q
why are scales of measurement that provide more info better for staffing?
A
- the more info provided by a measure, the more insight we have into the conceptt
- the more fine distinctions we can make, the more sophisticated analytical models can be used to make predictions
25
4 types of validity
1. criterion related
2. content
3. construct
4. face
26
criterion related validity
- empirically correlate predictor and criterion scores
- concurrent (one point in time; collect predictor scores from job incumbents)
- predictive (two points in time; collect predictor scores from applicants, hire some of the, and collect criterion scores from those hired)
- differences are negligible but important
- must have large samples for criterion-related validation
27
content validity
- there is no criterion measure
- compare the content of the test to the content of the job
- requires the use of expert judgment
- most common type of validity because it is not dependent on sample size
28
construct validity
- ultimately what we are interested in (does the test tell us info about the underlying attribute)
- is especially relevant when we change the format (convert from paper to internet)
- convergent (a test should correlate with tests measuring similar attributes)
- discriminant (a test should not correlate with tests measuring different attributes)
29
face validity
- does a test look like it measures the attribute
- is not a true form of validity, a test may or may not look job-related but still predicts performance
- face validity helps reduce complaints
- try to make valid tests that are also face valid
30
type of predictor KSAOs
- cognitive (quantitative, verbal, analytical; knowledge and skill)
- noncognitive (experience and history; resumes, biodata, accomplishment records, references; personality, values, motivation, employability)
- physical
Staffing Exam 3
flashcards
Decks in class (1)
# Cards
