Likert or Not, Survey (In)validation Requires Explicit Theories and True Grit

From the time of Likert (1932) on, attitudes of expediency regarding both theory and methodology became apparent with reference to survey construction and validation practices. In place of theory and more--theoretically minded methods, such as those found in the early work of Thurstone (1928) and Coombs (1964), statistical models and methodological heuristics have come to dominate, with the result of gross simplification of complex techniques. Maul's Focus article, "Rethinking Traditional Methods of Survey Validation" (Andrew Maul), lampoons this atheoretical attitude with a rhetorical flair that is all too rare in this literature. More importantly, it draws stark attention to the scientific and ethical dilemmas that this attitude has created as self-report, survey-based instruments have become ubiquitous in psychological research, education systems, and modern society more generally. In this commentary, the authors emphasize Maul's call for survey validation practitioners (and methodologists in the psychological sciences, more generally, including ourselves) to critically reflect on current practices with the same attitude of openness, optimism, and perseverance toward conceptual and methodological growth that we increasingly expect of students in their own cognitive development as 21st-century learners. Such growth necessitates a spirit and process of critical inquiry, in order that survey validation methods find firmer scientific foundation through careful conceptual, logical, and empirical analysis (Petocz & Newbery, 2010). A primary tool of critical inquiry is logical analysis, which may be used to examine the clarity and coherence of current practices and their inherent (and often implicit) assumptions. In the discussion presented in this article, the authors direct such a logical analysis toward the empirical aspects of Maul's article. In doing so, the authors will somewhat disagree with Maul's empirical means of demonstration to strongly agree with his article's overarching message. No reverse coding will be necessary. [For "Rethinking Traditional Methods of Survey Validation," see EJ1156001.]