Knowledge Check 2- Staffing and Selection
Obtained scores are comprised of:
true scores and error
What does it mean when someone scores in the 95th percentile on a test?
their score was higher than 95 percent of the people taking the test
If the coefficient of determination is .64, then we know that:
64% of the differences in individuals' job performance can be explained by differences in their test scores
Two individuals' scores on a test should not be considered significantly different from one another unless the difference is:
2 times the standard error of measurement of the test
Which of the following represents one approach to conducting a criterion-related validity study?
Both predictive and concurrent validation are criterion-related strategies
Which of the following is the most important source of information on tests for human resource selection?
Buros' Mental Measurements Yearbook
Research conducted by Barrett, Phillips, and Alexander drew which conclusion about predictive and concurrent validation?
Concurrent validation is probably just as useful as predictive validation for determining the validity of ability tests
What does it mean that the relationship between a predictor and a criterion is statistically significant at the .05 level?
If we conducted our study in the same manner 100 times, we would find a similar relationship between the predictor and criterion in 95 out of 100 of those studies
Which of the following is true? Paper and pencil selection tests never yield interval data Nominal data cannot be statistically analyzed Most selection tests represent ratio scales of measurement Interval data permit more powerful statistical analyses to be performed than ordinal data
Interval data permit more powerful statistical analyses to be performed than ordinal data
Test - retest reliability estimates are most useful for:
ability tests
Which of the following can influence whether concurrent validation is a useful strategy?
all three of the other options influence its usefulness
What is a content validity ratio?
an index computed on a panel of experts' ratings regarding the degree of overlap between the content of a selection measure and the content of the job in question
If the reliability of a test is .81, then we know that the validity of the test:
cannot be greater than .90
For a selection test to be "standardized", it is important to assess the:
content, administration, and scoring of the test
Which of the following is not a classical approach that can be used to validate measures used in selection?
face validity
Which of the following measurement scales ranks objects from "high" to "low" on a variable of interest?
ordinal
Which of the following IS NOT an estimate of internal consistency? parallel forms reliability Coefficient Alpha split-half reliability Kuder-Richardson reliability
parallel forms reliability
Which type of validation strategy is most influenced by sample size?
predictive validity
Consistency of measurement is most closely associated with:
reliability
"Level B" tests:
require some formal training of testing concepts to be properly scored and interpreted
Which of the following symbolically represents a validity coefficient?
rxy
Cross validation is necessary because of the effects of ______ on predictive accuracy.
shrinkage
In which type of validation are selection measures validated across several jobs in an organization rather than using only one job?
synthetic
Precision in measurement refers to:
the number of distinct scores permitted by the measure that is being used
What does the correction for attenuation formula show?
what the true variance in test scores would be if their range was not restricted