회귀분석과 구조방정식 모형에서의 상호작용효과 검증 : 이론과 절차 = Testing the Interaction Effects in Regression and Structural Equation Models: Theories and Procedures

다국어 초록 (Multilingual Abstract)

As behavioral science research advances, third variables have increasingly been considered in the relationship between independent and dependent variables in many research. Research has commonly examined the role of mediating and/or moderating variables. A moderating variable indicates a variable that has an interaction effect on a dependent variable, producing a joint effect with an independent variable. Given that mediation (or indirect) effects can be easily tested in structural equation models that are very widely used, mediation effects are more commonly tested than interaction effects. One of the main possible reasons interaction models are underutilized is that testing interaction effects can be complicated and thus many researchers often experience difficulties. In view of this, analysis methods for testing the interaction effects are explained and discussed in detail using regression analysis and structural equation models in the present study. First, mean centering, correction of standardized interaction coefficient, and reliability issues of interaction variables are explained in regression analysis. Next, constrains of parameters and unconstrained methods as well as the abovementioned three issues are emphasized in structural equation models. In addition, regression analysis and structural equation models are applied to a real data set to explain the procedures of the analysis. Finally, several issues that are commonly misunderstood by many researchers are presented and clarified.

참고문헌 (Reference)

1 Little, T. D., "To parcel or not to parcel : Exploring the question, weighing the merits" 9 (9): 151-173, 2002

2 Marsh, H. W., "Structural equation models of latent interactions : Evaluation of alternative estimation strategies and indicator construction" 9 (9): 275-300, 2004

3 Wen, Z., "Structural equation models of latent interactions : An appropriate standardized solution and its scale-free properties" 17 (17): 1-22, 2010

4 Edwards, J. R., "Statistical and methodological myths and urban legends" Routledge 143-164, 2009

5 Çivitci, N, "Self-esteem as mediator and moderator of the relationship between loneliness and life satisfaction in adolescents" 47 (47): 954-958, 2009

6 Bohrnstedt, G. W., "On the exact covariance of products of random variables" 64 (64): 1439-1442, 1969

7 Bandalos, D. L., "New developments and techniques in structural equation modeling" Lawrence Erlbaum Associates 269-296, 2001

8 Aiken, L. S., "Multiple regression : Testing and interpreting interactions" Sage 1991

9 Ping, R. A., "Latent variable interaction and quadratic effect estimation : A two-step technique using structural equation analysis" 119 (119): 166-, 1996

10 Jackman, M. G., "Estimating latent variable interactions with the unconstrained approach : A comparison of methods to form product indicators for large, unequal numbers of items" 18 (18): 274-288, 2011

11 Busemeyer, J. R., "Analysis of multiplicative combination rules when the causal variables are measured with error" 93 (93): 549-562, 1983

12 Algina, J., "A note on estimating the Jöreskog-Yang model for latent variable interaction using LISREL 8. 3" 8 (8): 40-52, 2001