Unproctored Internet Testing: Practical, Legal, and Ethical Concerns

The thesis of this brief comment is that unproctored Internet testing (UIT), when used to make important decisions about individuals, is subject to serious ethical and legal challenges, aswell as anumber of organizationally harmful consequences thatmay outweigh any real or perceived benefits. This situation is both alarming and troubling, as the deployment of UIT has become widespread in the United States in recent years, to the point where it is an almost standard part of the application process in some economic sectors. For example, the UIT-based administration of some type of personality test immediately following completion of an online employment application—with no attempt to ensure test security, authenticate applicant identity, or control (or even discourage) cheating and with no followup (or ‘‘verification’’) testing (as discussed below)—is now probably the modal application protocol for nonexempt jobs in nationwide retail chains. Although definitive usage data are lacking, it is quite possible, even probable, that UITaccounts for the majority of individual employment test administrations that currently take place in the U.S. private sector. In conjunction with the ethical and legal challenges just mentioned, on which I elaborate below, this means there is a strong likelihood that most contemporary U.S. employment testing practice is in violation of both ethical standards of psychological practice and accepted professional testing standards. This discussion is limited to UIT settings, that is, situations where there is no (either direct or technologyenabled) humanor electronic monitoring of individuals while they are taking a test. When discussing such situations, the most critical distinction is that between low-stakes testing applications (LSTAs) and high-stakes testing applications (HSTAs). LSTAs are situations inwhich testing consequences or outcomes are of relevance or value only to the individual tested and have limited or no direct effects on or consequences for other individuals or institutions (e.g., the provision of sample or practice tests or self-diagnostic tests). HSTAs are situations in which test results are used to make evaluations, judgments, or decisions that arehighly consequential to an individual and also have consequences or outcomes beyond the individual tested (e.g., as in the cases of employee selection tests or professional credentialing tests). In contrast to the LSTA context,whereUIT can be highly beneficial to both individuals and organizations, as I have described elsewhere (Tippins et al., 2006), UIT is never appropriate for HSTAs, for two major reasons. First, the fundamental, widely acknowledged (cf. Bartram, 2006; Naglieri et al., 2003), and interrelated problems of candidate cheating, candidate identity authentication, and test and item security are of significant concern only in the HSTA context. (For one thing, these problems are much likelier to arise in HSTAs than LSTAs, Correspondence concerning this article should be addressed to Kenneth Pearlman. E-mail: kenpearlman @comcast.net Address: Creative PersonnelManagement Consulting, 7848 Camminare Drive, Sarasota, FL 34238 Industrial and Organizational Psychology, 2 (2009), 14–19. Copyright a 2009 Society for Industrial and Organizational Psychology. 1754-9426/09