Issues and non-issues in the fidelity–bandwidth trade-off

✍ Scribed by JOYCE HOGAN; BRENT W. ROBERTS

• Publisher: John Wiley and Sons
• Year: 1996
• Tongue: English
• Weight: 867 KB
• Volume: 17
• Category: Article
• ISSN: 0894-3796
• DOI: 10.1002/(sici)1099-1379(199611)17:6<627::aid-job2828>3.0.co;2-f

✦ Synopsis

This paper makes seven points in response to certain claims made by Ones and Viswesvaran (1996, this issue). First, we see no evidence that the fidelity-bandwidth trade-off has become a crisis in the empirical literature. Moreover, we seen no evidence that anyone prefers narrow band personality measures over broad bandwidth scales. In addition, because job performance is complex and multidimensional, broad bandwidth predictors are normally required in personnel selection. Finally, our conclusion is imple-the nature of the criterion dictates the choice of predictors and matching predictors with criteria always enhances validity. Personality measures are increasingly being used to predict job performance and new questions about old principles arise. One question concerns what to measure. A second concerns how to assess work performance. A third concerns how to align attributes of people with job requirements and a fourth concerns the appropriate level of measurement to maximize performance prediction. The concepts of fidelity and bandwidth are relevant to all these questions.

Shannon and Weaver's (1949) The Mathematical Theory of Communication introduces these concepts to social science. Shannon argued that any communication system involves a trade-off between fidelity-quality of information-and bandwidth-complexity of information obtained. Greater fidelity is achieved at the loss of bandwidth and increased bandwidth comes at the price of fidelity. Cronbach (1960) suggested that we subscribe to the psychometric ideal of high fidelity, narrow bandwidth assessments. This ideal may apply to educational assessments, but it may not apply to assessments in the workplace.