Scribd
Upload
95 views3 pages

Moderating

A moderating variable alters the impact of an independent variable X on a dependent variable Y. Moderating variables allow for testing the notion of contingent effects. Researchers benefit from focusing on contingent effects that only occur in the absence of offsetting, blocking, or cross-cutting causal processes.

Uploaded by

nursyahidaz
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
95 views3 pages

Moderating

A moderating variable alters the impact of an independent variable X on a dependent variable Y. Moderating variables allow for testing the notion of contingent effects. Researchers benefit from focusing on contingent effects that only occur in the absence of offsetting, blocking, or cross-cutting causal processes.

Uploaded by

nursyahidaz
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

The SAGE Encyclopedia of Social Science Research Methods

Moderating Variable

Contributors: Michael S. Lewis-BeckAlan BrymanTim Futing Liao Print Pub. Date: Print ISBN: 9780761923633 Online ISBN: 9781412950589 DOI: Print page: 660 This PDF has been generated from SAGE Research Methods Online. Please note that the pagination of the online version will vary from the pagination of the print book.

Universiti Teknologi Mara Copyright 2011

Sage Publications, Inc.

A moderating variable represents a process or a factor that alters the impact of an independent variable X on a dependent variable Y. The moderating variable may take the form of an indicator variable (0/1 values), a categorical variable, or a continuous variable. Moderation of the effects of X on Y is usually of interest in the context of a well-theorized area of research, such as the provision of labor by individuals in which the human capital and labor market segmentation theories are both well established. Variables reflecting these two theories can be put together in a multiple regression equation, and a path analysis involving two stages (labor force participation in the first stage and wage determination in the second) can be developed. Paths through the model involve the multiplication of the coefficients from two regressions together, giving the moderated effect of X on Y through Z, and comparing that effect with the direct effect of X on Z (which is assumed to be additively separable) (Bryman & Cramer, 1997). Furthermore, background factors can be tested for a moderating effect on the main impact of each X on the final outcome Y. Moderating variables allow for testing the notion of contingent effects. Realists distinguish necessary effects from contingent effects (Sayer, 2000, p. 94). Researchers benefit from focusing on contingent effects that only occur in the absence of offsetting, blocking, or cross-cutting causal processes. The moderating variables may act as obstacles to the action of X in causing Y. Thus, divorce only results from submitting divorce papers if the case is well constructed, uncontested, or both; it is contingently caused by divorce papers. Two main approaches are used for estimating the impact of moderating variables: path analysis and INTERACTION EFFECTS. The path analysis approach allows for two or more equations to simultaneously represent the set of associations. These equations may involve independent variables Xij, intermediate variables Zij, and final outcomes Yij. In each equation j, variables may be recursively affected by factors in other linked equations. Such models are structural equation models.

Page 2 of 3

The SAGE Encyclopedia of Social Science Research Methods: Moderating Variable Sage Research Methods Online

Universiti Teknologi Mara Copyright 2011

Sage Publications, Inc.

The alternative method of using interaction effects (Jaccard, Turrisi, & Wan, 1990) introduces an extra term that is the product of two existing variables into a single regression equation. The interaction terms may be numerous for categorical variables. The degrees of freedom of the new estimate are reduced. For example, in researching the factors associated with eating out frequently, the interaction effect of marital status with having young children in the household was measured using dummy variables in a logistic regression context (Olsen, Warde, & Martens, 2000). The impact of [Divorced Kids < 5 in Hhold] was a dramatic increase in the odds of eating out in a hotel or of eating an unusual ethnic meal. This interaction term offset one of the underlying associations: Divorced people in the United Kingdom were generally less likely than married/cohabiting people to eat in ethnic restaurants. A simpler alternative is to use dummy variables to control for the impact of moderating factors. Another alternative is to use multilevel modeling to represent contextual moderating influences (Jones, 1997). Olsen, W. K. , Warde, A. and Martens, L. Social differentiation and the market for eating out in the UK. International Journal of Hospitality Management vol. 19 (2) pp. 173190 (2000).

Page 3 of 3

The SAGE Encyclopedia of Social Science Research Methods: Moderating Variable Sage Research Methods Online

You might also like