Risky Predictions and Damn Strange Coincidences in Theory Appraisal

The elucidation and empirical testing of theories are
vital components of research in any field. However,
the extent to which theories are supported or
contradicted by the results of empirical research
remains ill defined. Commonly, support or
contradiction is based solely on the reject or fail to reject decisions that result from null
hypothesis tests derived from aspects of theory.
Decisions and recommendations based on this forced
and often artificial dichotomy have been scrutinized
in the past. This overly simplified approach to
theory testing has been challenged on logical grounds
(Meehl, 1997, 1990, 1978; Serlin & Lapsley, 1985).
Theories differ in the extent to which they provide
precise predictions about observations and the
precision derived from theories is proportional to
the strength of support that may be provided by
empirical evidence congruent with the prediction.
The notion of linking precision to strength of
support is surprisingly absent from many discussions
in this realm. Using Monte Carlo methods, this study
examined the performance of a multivariate extension
of Meehl s Corroboration Index. Implications for
theory and practice are discussed.

Autorentext

Kristine Y. Hogarty is the Director of Assessment for the Collegeof Education at the University of South Florida. Her experiencesinclude instrument development,research design and data analysisin Education and Industry. Her research interests are focused onapplied statistics analysis. Her work has been published in ahost of scholarly journals.

Klappentext

The elucidation and empirical testing of theories arevital components of research in any field. However,the extent to which theories are supported orcontradicted by the results of empirical researchremains ill defined. Commonly, support orcontradiction is based solely on the "reject" or"fail to reject" decisions that result from nullhypothesis tests derived from aspects of theory. Decisions and recommendations based on this forcedand often artificial dichotomy have been scrutinizedin the past. This overly simplified approach totheory testing has been challenged on logical grounds(Meehl, 1997, 1990, 1978; Serlin & Lapsley, 1985). Theories differ in the extent to which they provideprecise predictions about observations and theprecision derived from theories is proportional tothe strength of support that may be provided byempirical evidence congruent with the prediction. The notion of linking precision to strength ofsupport is surprisingly absent from many discussionsin this realm. Using Monte Carlo methods, this studyexamined the performance of a multivariate extensionof Meehl's Corroboration Index. Implications fortheory and practice are discussed.