Consider the following scenario. You have just assumed a new position as Chief Human Resources Officer (CHRO) for a U.S.-based start-up firm that provides sales and service of sophisticated computer-based simulation software. There is growing demand in the marketplace for the products your firm sells and services. As a result, you expect to hire many new associates for a variety of sales and service jobs over the next several years.
In considering alternative selection systems for these jobs, a number of issues immediately present themselves, and you jot down some of the most prominent ones:
- With respect to the validity of any tool or instrument that we use as a basis for choosing among candi-dates, can we rely on evidence developed in similar situations elsewhere, or do we have to develop our own "in-house" data as well? 2. What types of measures (i.e., criterion measures 1 ) should we use to assess the performance of associates on the jobs in question? 3. If we develop "in-house" data on predictors (i.e., assessment tools) and measures of job performance on one sample of individuals, do we need to use a different sample in order to check (that is, cross-validate) the results obtained with the first sample? 4. Is evidence of statistical significance sufficient to draw conclusions about TEST DEVELOPMENT AND USE: NEW TWISTS ON OLD QUESTIONS