Employment Regulation

Variable X is called the criterion variable and represents the performance criterion.

Predictive validation is a means of establishing criterion-related validity.

Content validity is demonstrated by empirical data proving a selection procedure is predictive of the important elements of the job in question.

It is possible that the results of a concurrent validation may be biased because of the sample.

An overriding factor in deciding to use construct validation is the availability of an individual with psychometric skills and expertise needed for this level of analysis.

Both credit and background checks are important in avoiding negligent hiring charges.

A strong negative correlation means that a change in the score of X indicates commensurate change in the score Y in the opposite direction.

The EEOC advocates the use of tests of significance.

Under predictive validation the correlation coefficients between test and performance are determined by administering the test to current employees.

Criminal background checks could possibly have a disparate impact on ethnic groups.

The EEOC considers all criteria used in making employment decisions to be "tests."

Requiring a secretary to type a letter on a personal computer using a specific word-processing package would be a selection criterion with construct validity.

Performance appraisals known as graphic rating scales may be particularly susceptible to a high degree of subjectivity.

Language proficiency tests can only be used to test English.

While often confused, cognitive tests are distinctly different from cognitive ability tests.

A problem with integrity tests is their propensity to being manipulated by the test-taker.

Pearson product-moment correlation coefficient is a parametric statistic.

The complaining party can offer evidence that there are other employment practices that would accomplish the same objective as a challenged employment practice, but would have less disparate impact on a protected class.

Of the different validation methods, construct validity is probably the most complicated.

Tests of significance help reduce the likelihood that the correlations are merely a random error.