A Tutorial on Testing, Visualizing, and Probing an Interaction Involving a Multicategorical Variable in Linear Regression Analysis

Andrew F. Hayes,Amanda K. Montoya

DOI: https://doi.org/10.1080/19312458.2016.1271116

IF: 8.044

2017-01-02

Communication Methods and Measures

Abstract:Empirical communication scholars and scientists in other fields regularly use regression models to test moderation hypotheses. When the independent variable X and moderator M are dichotomous or continuous, the practice of testing a linear moderation hypothesis using regression analysis by including the product of X and M in a model of dependent variable Y is widespread. However, many research designs include multicategorical independent variables or moderators, such as in an experiment with three or more versions of a stimulus where participants are randomly assigned to one of them. Researchers are less likely to receive training about how to properly test a moderation hypothesis using regression analysis in such a situation. In this tutorial, we explain how to test, visualize, and probe interactions involving a multicategorical variable using linear regression analysis. While presenting and discussing the fundamentals—fundamentals that are not software specific—we emphasize the use of the PROCESS macro for SPSS and SAS, as it greatly simplifies the computations and potential for error that exists when doing computations by hand or using spreadsheets based on formulas in existing books on this topic. We also introduce an iterative computational implementation of the Johnson-Neyman technique for finding regions of significance of the effect of a multicategorical independent variable when the moderator is continuous.

communication

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to test, visualize, and explore interactions involving multi - category variables in linear regression analysis. Specifically, when the research design includes independent variables or moderator variables with multiple categories, how to use linear regression analysis to test moderation hypotheses. Many research designs involve variables with multiple categories. For example, in an experiment, there are three or more versions of stimuli, and participants are randomly assigned to one of them. However, researchers are usually not trained on how to correctly use regression analysis to test moderation hypotheses in such situations. Therefore, this tutorial aims to explain how to use linear regression analysis to test, visualize, and explore interactions when multi - category variables are involved, especially when the moderator variable is continuous.

### Main problem points:
1. **Treatment of multi - category variables**: When the independent variable or moderator variable in the study is multi - category (at least three categories), how to include it in the regression model for analysis.
2. **Testing of interactions**: How to test the interactions between multi - category variables and continuous variables through regression analysis.
3. **Visualization and interpretation of results**: How to effectively visualize and interpret the results of these interactions in order to better understand the effect differences under different conditions.
4. **Application of statistical methods**: Introduce how to use the PROCESS macro in SPSS and SAS to simplify the calculation process and reduce possible errors when doing manual calculations or using spreadsheets.

### Solutions:
- **Encoding of multi - category variables**: It introduces how to use indicator variables (dummy coding), sequential coding, and Helmert coding to represent multi - category variables.
- **Testing of interactions**: Through the method of model comparison, compare the goodness - of - fit of the model with interaction terms and the model without interaction terms, thereby testing the existence of interactions.
- **Visualization of results**: It provides how to use graphic tools to show the interactions between multi - category variables and continuous variables, helping researchers understand the data more intuitively.
- **In - depth exploration of interactions**: It introduces how to use Pick - a - Point and Johnson - Neyman techniques to further explore the specific areas and conditions of interactions.

### Practical application examples:
The paper uses a fictional video game research data set as an example to show how to apply the above methods in actual research. In the study, participants were randomly assigned to three different types of video games, which contained different levels of violence and sexist content respectively. The researchers explored the influence of game type (a multi - category variable) and participant age (a continuous variable) on masculinity beliefs through regression analysis, and showed how to test and visualize these interactions.

Through these methods, researchers can more accurately understand and interpret the role of multi - category variables in regression analysis, thereby improving the scientific nature and reliability of the research.