Advances in technology now enable employers to utilize computers to administer online employee selection tests, which result in lower costs, increased efficiency, and fewer transcription errors (Richman, Kiesler, Weisband, & Drasgow, 1999; Tippins et al., 2006). Additionally, online employment testing software can effectively and efficiently assist in identifying individuals best suited to an occupation, reducing poor person-job fit, lowering turnover rates, decreasing training costs, and minimizing errors in hiring (Bingham, Ilg, & Davidson, 2002; Mooney, 2002). This article addresses issues related to online employment testing software including types of tests available, validity and reliability, proctoring, and social desirability. Additional terms are defined and implications and future directions for research are discussed.
TopBackground
For decades, the similarities and differences between written tests and computer-based tests have been assessed (Epstein & Klinkenberg, 2001). Early research investigated how a computer-based medical records keeping and interview system impacted patients (Slack & Van Cura, 1968; Slack, Hicks, Reed, & Van Cura, 1966). Additionally, research in this era investigated the use of computers as data gathering instruments (Evans & Miller, 1969; Vinsonhaler, Molineaux, & Rodgers, 1968). However, it is not until the 1990s that we see a research trend that begins to examine the equivalence of computer-based tests vs. conventional tests in an organizational setting (Donovan, Drasgow, & Probst, 2000; McHenry & Schmitt, 1994).
As technology has evolved, tests previously administered in a paper-and-pencil format have been changed to online versions. These tests include clinical measurements, personality tests, attitude scales, cognitive ability tests and training inventories (Mead & Drasgow, 1993). Further examples of computer-administered assessments include medical admissions data, psychiatric evaluation exams, and consumer preference evaluations (Kiesler, Walsh, & Sproull, 1992; Richman et al., 1999; Synodinos & Brennan, 1988; Synodinos, Papacostas, & Okimoto, 1994).
The most simple and widely used type of computer-based test is computer assisted testing (CAT). These tests display a question on a computer screen and the respondent enters their response (Epstein & Klinkenberg, 2001). Computer assisted tests enable the online format of a test to very closely resemble the paper-and-pencil version and make the testing situation as similar as possible to a written one (Rozensky, Honor, Rasinski, Tovian, & Herz, 1986).
Another type of CAT program uses computer adaptive testing. Adaptive testing settings are different from assisted ones in that adaptive tests allow the computer to “go beyond a simple page turning function” (Epstein & Klinkenberg, 2001, p. 298). Adaptive tests allow a computer to receive a response, score it, and then choose the next appropriate question, either easier or harder based on a respondent’s answer (Green, Bock, Humphreys, Linn, & Reckase, 1984). In adaptive tests, which Epstein & Klinkenberg (2001) assert are similar to most non-computerized intelligence tests, there are multiple types that can “individualize” the testing experience and narrow the number of questions needed to assess the underlying trait (Burke, 1993; Weiss, 1985).