Moderation Analysis with Categorical Variables using SMART-PLS
Moderation Hypothesis
- A moderating variable is the one that modifies the existing relationship between the independent and dependent variable I-e it holds a strong contingent effect on the association of IV and DV. Although, MV is not influenced by IV but affects the strength and direction of the relationship between IV and DV.
- A categorical moderator is a moderating variable that is divided into categories. for example Gender, Job Rank, Type of Bank.
Categorical Moderation Hypothesis
In this example i test the following hypothesis with the categorical variable Type of Bank moderating the two relationships. The proposed hypotheses are
H1: Type of Bank Moderates the Relationship between Organizational Commitment and Organizational Culture
H2: Type of Bank Moderates the Relationship between Perceived Organizational Support and Organizational Culture
Types of Interaction Effects
Following option are available when running moderation analysis (Bootstrapping)
Product Indicator Approach
When the IV or the Moderator is measured reflectively, use Product Indicator Approach
Two Stage Approach
When the IV or the Moderator is measured formatively, use Two Stage Approach
Product Term Generation
Use Unstandardized if the moderator is Categorical
R-Square
- In Moderation R-Square Change is important. One shall run the model before the interaction effect is added and after the interaction effect is added to assess the R-Square.
- F-Square Effect Size
- Effect Size (f-Square) = Included (.570) – Excluded (.542)
- .02 – Small
- .15 – Medium
- 035 – Large
Interpretation
- In order to clearly identify how the interaction differs in terms of the groups, including the size and precise nature of the effect, we shall draw an interaction plot.
- Review the Steepness of the Line.
- The group where steepness is higher, the impact if stronger.
- Download Link (Shared in Description)
- http://www.jeremydawson.co.uk/2-way_with_binary_moderator.xls
Sample Interpretation - H1
Here is the interpretation of H1. For Detail on how to run Moderation, please watch the video at the end of the tutorial
The results revealed a significant moderating role of Type of Bank on the relationship between OC and Collaborative Culture. The plot shows a steeper and positive gradient for private sector banks as compared to public sector banks. Thus, this shows that the impact of OC in fostering collaborative culture is stronger in private banks as compared to public sector banks.
Sample Interpretation - H2
Here is the interpretation of H2. For Detail on how to run Moderation, please watch the video at the end of the tutorial
The results revealed a significant moderating role of Type of Bank on the relationship between POS and Collaborative Culture. The plot shows a steeper and positive gradient for public sector banks as compared to private sector banks. Thus, this shows that the impact of POS in fostering collaborative culture is stronger in public sector banks as compared to private sector banks.
For a Step by Step Guide on How to Perform Moderation with Categorical Variables, Watch the Video
Reference:
Ramayah, T., Cheah, J., Chuah, F., Ting, H., & Memon, M. A. (2018). Partial least squares structural equation modeling (PLS-SEM) using smartPLS 3.0.