Nine hundred andWfty-nineWrms in twenty diVerent countries responded to their
survey on the manner in which hiring decisions are made. Extensive use of
interviews was reported in all countries ranging from an average of two per
applicant to nearly four per applicant in France. Across all countries, theseWrms
reported using just underWve test types to evaluate job applicants. Of those who
used tests, work samples, medical screens, and cognitive ability tests were most
frequently used. Physical ability tests, integrity/honesty tests, video-based tests,
projective tests, drug tests, and graphology were infrequently or never used. There
were some relatively large diVerences across countries also; respondents in Japan
and Malaysia reported no test use. Firms in all parts of the world report that they
often or always use application blanks, educational qualiWcations, references from
previous employers, and, to a somewhat lesser extent, personal references as a
means to make decisions about prospective employees. While the Ryan et al. study
does provide descriptive data on the methods used and the extent of their use, it
does not inform us as to the manner or sequence in which such data are collected.
In US companies, it is probably the case that educational qualiWcations, application
letters, and letters of reference are used as initial screens followed by interviews and
more formal, quantiWable data collection using tests. However, the use of tests is by
no means universal. On average, approximately 30 percent of US organizations
indicated using these devices, with a slightly greater use reported across organiza-
tions in all countries.
The constructs measured by these various methods are often categorized into
‘can do’ measures indicating the ability to perform important work tasks and ‘will
do’ measures that reXect a person’s motivation or willingness to perform work well.
Measures of both sets of constructs prove to be valid predictors of subsequent job
performance, as described in the next section.
- 1 Validity of Methods Used
One common way in which test use is justiWed is to correlate test scores or other
methods of evaluating applicant potential with subsequent measures of job
performance. A recent summary of these criterion-related validation studies
conducted over an eighty-Wve-year period has been provided by Schmidt and
Hunter ( 1998 ). They report average correlations above. 50 for general mental
ability tests, work sample tests, and structured interviews. The average validity
reported for job knowledge tests is. 48. Somewhat lower validities are reported for
measures of personality constructs such as conscientiousness and integrity and
for methods such as job experience measures, unstructured interviews, and
reference checks. Assessment centers which include multiple methods of data
collection display validities of. 37. The manner in which organizations assess
criterion-related validity (and the type of study that is the source of the Schmidt
selection decision-making 301