Using a Repeated-Measures Approach to Validating Personality Tests in Small Samples: A Feasibility Study with Implications for Small Businesses

Article excerpt

In the last 20 years or so, as large businesses have been reducing their workforces, small businesses have contributed substantially to job growth in the United States (Gatewood and Feild, 2001). Additionally, small businesses are becoming increasingly important participants in foreign markets (Darling et al, 1997). The Small Business Administration (2006) noted that small businesses employed roughly half of private sector employees in 2005, and employed individuals and demographic groups who might otherwise have not been employed. Thus, small businesses are making more and more personnel selection decisions, and they are making such decisions without the support and guidance of a large human resources (HR) department.

However, we would assert that small businesses are most in need of ensuring that every HR decision is a good one as they have less "room to maneuver" in response to labor issues in comparison to larger firms (Popkin, 1999: 3). Specifically, in small businesses there is simply less capital, fewer assets, fewer employees, and thus less room for decision errors as the limited resources available are more easily drained in comparison to larger businesses (Klaas et al., 2002). Human resource support in terms of selection, staffing, training, and compensation have all been noted as often lacking in small businesses, and yet are very necessary for reduced employee turnover and small business success (Carlson et al., 2006; Heneman et al., 2000; Hope and Mackin, 2007; Minkus, 2007). For example, unlike large businesses, small businesses cannot afford extensive training programs to provide employees with job-relevant knowledge, skills, and abilities (KSAs), and thus HR selection is all the more critical to the success of the small business because it needs to hire applicants who already possess job-relevant KSAs. Indeed, small businesses have faced difficulties hiring employees with needed job-specific technical skills and good work habits (Popkin, 1999). Staffing in particular has been identified as an understudied area among small businesses, and researchers have called for additional studies to improve selection practices for small business owners (Heneman et al., 2000). Further, given many small businesses produce limited profits, poor business decisions may have a large effect on the small business's financial bottom line. Thus, in the small business one or two poor hiring decisions may lead to low-quality service and products, resulting in subsequent customer dissatisfaction that could potentially reduce profits to the point at which the small business cannot survive.

Also important is the fact that small businesses with as few as 15 employees are subject to the major equal employment opportunity (EEO) laws (e.g., Title VII of the Civil Rights Act of 1964). Given the limited resources of small businesses, a single EEO lawsuit regarding an unlawful selection decision could be sufficient to bankrupt the business. As small businesses are overrepresented in business closures (Small Business Administration, 2005), making accurate and legal selection decisions is one of the most important HR activities for small businesses.

Small businesses also tend to have limited options in terms of practical and defensible selection procedures. Training and experience evaluations, structured interviews, and work sample tests, which rely primarily on content validation strategies, may be feasible for many small businesses (Gatewood and Feild, 2001), but other tests that require large sample sizes for criterion-related validation are often impractical. Small businesses may use existing test batteries or rely on strategies such as validity generalization (Ployhart et al., 2006), but local validation efforts can still be necessary for legal defensibility. Also, small businesses are unlikely to obtain statistical significance via criterion-related validation because of their small sample sizes and the resulting loss of statistical power. …