Moderator variables: A clarification of conceptual, analytic, and psychometric issues

https://doi.org/10.1016/0030-5073(82)90254-9Get rights and content

Abstract

A distinction is drawn between the degree of relationship between two variables X and Y and the form of the relationship between the same variables. The correlation coefficient rxy is the index of degree of relationship, while the regression coefficient Byx is the index of the form of relationship. If both the form and degree are constant across values of some third variable Z, the XY relationship is constant or unconditional with regard to Z. If the degree of relationship varies with values of Z, Z is said to “moderate the degree” of the XY relationship. If the form of relationship varies with values of Z, Z is said to “moderate the form” of the relationship. The separate statistical tests which are required to test for each type of moderator variable are outlined. Hierarchical multiple regression is the appropriate method to test for different forms of relationship, but not for different degrees of relationship. The differing substantive implications and interpretation of the two types of moderated relationships are discussed. Finally, the effects of unreliability of measures on the power of tests to detect differing forms and degrees of relationships are outlined and demonstrated.

References (33)

  • H.J. Arnold et al.

    Testing multiplicative models does not require ratio scales

    Organizational Behavior and Human Performance

    (1979)
  • R.P. Abelson

    Sex differences in predictability of college grades

    Educational and Psychological Measurement

    (1952)
  • V.R. Boehm

    Negro—white differences in validity of employment and training selection procedures

    Journal of Applied Psychology

    (1972)
  • V.R. Boehm

    Differential prediction: A methodological artifact?

    Journal of Applied Psychology

    (1977)
  • H.E. Brogden

    On the interpretation of the correlation coefficient as a measure of predictive efficiency

    Journal of Educational Psychology

    (1946)
  • T.A. Cleary

    Test bias: Prediction of grades of Negro and White students in integrated colleges

    Journal of Educational Measurement

    (1968)
  • W.G. Cochran

    Errors of measurement in statistics

    Technometrics

    (1968)
  • J. Cohen

    Multiple regression as a general data analytic system

    Psychological Bulletin

    (1968)
  • J. Cohen

    Statistical power analysis for the behavioral sciences

    (1977)
  • J. Cohen

    Partialled products are interactions: partialled powers are curve components

    Psychological Bulletin

    (1978)
  • J. Cohen et al.

    Applied multiple regression/correlation analysis for the behavioral sciences

    (1975)
  • R.B. Darlington

    Another look at “cultural fairness”

    Journal of Educational Measurement

    (1971)
  • N. Fredericksen et al.

    Differential predictability in the use of test scores

    Educational and Psychological Measurement

    (1954)
  • N. Fredericksen et al.

    Adjustment to college: A study of 10,000 veteran and nonveteran students in sixteen American colleges

    (1951)
  • E.E. Ghiselli

    Differentiation of individuals in terms of their predictability

    Journal of Applied Psychology

    (1956)
  • E.E. Ghiselli

    The prediction of predictability

    Educational and Psychological Measurement

    (1960)
  • Cited by (479)

    • Moderator variables in consumer research: A call for caution

      2023, Journal of Retailing and Consumer Services
    • Media attention and its impact on corporate commitment to climate change action

      2021, Journal of Cleaner Production
      Citation Excerpt :

      We also note in Table 4 that a company's industry membership (that is, being a member of the environmentally sensitive industries) is correlated negatively significantly (at the 1% level) with both its commitment to taking action on climate change (COMMIT) and the volume of media coverage (MEDIA) it receives (see Table 4). A joint consideration of the significant effect of the interaction term with the significant correlation coefficients suggests that industry membership of a firm moderates the form of the relationship between the media's coverage of a firm and its commitment to climate change actions (Arnold, 1982; Hartmann and Moers, 1999; Sharma et al., 1981). In other words, the slope of the association between the media's coverage of a firm and its commitment to climate change actions is steeper for firms in the environmentally sensitive industries relative to those in other industries (Hartmann and Moers, 1999; Sharma et al., 1981).

    View all citing articles on Scopus

    Frank Schmidt provided extremely helpful advice in the development of the manuscript. The suggestions and comments of Martin Evans, Art Jago, Stephan Motowidlo, and several anonymous reviewers are also gratefully acknowledged.

    View full text