Selecting bus drivers: Multiple predictors, multiple perspectives on validity, and multiple estimates of utility

Rick R. Jacobs, Jeffrey M. Conte, David V. Day, Jay M. Silva, Russ Harris

Research output: Contribution to journalReview articlepeer-review

32 Scopus citations


For the past two decades, the transportation industry has operated in an environment of diminishing public funding, escalating costs, and declining ridership, all of which contribute to shortfalls in available income for this vital public service. Improved selection techniques that screen out undesirable drivers and select applicants with the potential for lower absenteeism rates and safer driving behaviors should result in more cost-effective operations. This study identified potential predictors and developed meaningful criterion measures to assess the validity of a variety of measures that might aid in the forecasting of job performance for a sample of 864 bus operators representing 9 bus properties from across the United States and Canada. Job analysis information converged on a tripartite theory of performance that can be described as requiring a bus operator to "Be There," "Be Safe," and "Be Courteous." A predictor composite was created that successfully predicted supervisory ratings of performance as well as objective absence and accidents. Utility analyses indicated that the predictor composite could be employed to reduce the overall operating expenses of bus transit agencies by over $500,000. The tremendous benefits that can be achieved through consortium-type arrangements that develop and utilize such selection systems was discussed.

Original languageEnglish (US)
Pages (from-to)199-217
Number of pages19
JournalHuman Performance
Issue number3
StatePublished - 1996

All Science Journal Classification (ASJC) codes

  • Applied Psychology
  • General Psychology
  • Organizational Behavior and Human Resource Management


Dive into the research topics of 'Selecting bus drivers: Multiple predictors, multiple perspectives on validity, and multiple estimates of utility'. Together they form a unique fingerprint.

Cite this