have 3 IV likert scale variables and 1 DV which is also likert scale data. I want to do moderation analysis. can i use regression for this data. And suggest ways to do.
@Yogesh You can use model 1 in Process Macro for SPSS to run Moderation analysis however, it would work only with one IV and DV at a time. You can also use JAMOVI or JASP for Moderation analysis.
Let me first start by setting a dentition for the moderator analysis. It is used to determine whether the relationship between two variables depends on (is moderated by) the value of a third variable.
For instance, you have 2 independent variables i.e. (Performace Expectancy denoted by PE, and Effort Expectancy denoted by EE ) with a dependent variable i.e. (Intention to accept driverless vehicles denoted by INT).
Keep in mind that mediators can be both categorical and continuous
In the following example and diagram, they are based on categorical moderators because it is easier to interpret.
you also have a set of social-demographic and you have splited already into two representative and meaningful groups while dummy coded (0,1) usually the statistical software treats the 0 group as the reference group, however, this is not always the case.
Let's assume you have the age of your respondents as one of your social-demographic variables and after thoughtful checking of your sample, you decided to divide them into two classes young respondents, and elderly respondents.
Bear in mind that there is no true/false for which class among the two (young respondents, and elderly respondents) is set to be the reference for the analysis. However, it is all about your research questions and the approach you are interested in investigating them.
In this example, I will choose to consider the young respondents as my reference, and assuming that my software takes 0 as the reference thus I will code the young respondents with 0, and the elderly respondents with 1.
To reflect that into a path model or a structural model if you are using structural equation modeling, then it should look like what you can see in the Moderation figure.
If the effect of PE * Age -> INT is found to be significant then,
if the sign is positive, it means that elderly respondents are more likely to perceive the benefits of autonomous vehicles, and ultimately more likely to intent to accept the autonomous vehicles. And it would that younger respondents are more likely to accept autonomous vehicles, in case the sign would be negative.
The same logic applies to EE * Age -> INT.
Regarding how to implement this moderation within software environment,
In case you are applying structural equation modeling, I can recommend you to use the variance-based approach then you can use SmartPLS, which is very easy to use, and equipped with highly sophisticated features and functions.
If you are familiar with R software, then there is this link that I can recommend where it present a brief theory and shows a walk-through example of how to implement moderation using R software. https://advstats.psychstat.org/book/moderation/index.php
If you are not applying structural equation modeling and not familiar with programming languages such as R and would like to have a simple to use software then I can recommend you using jamovi, you can add jAMM, which stands for jamovi Advanced Mediation Models, this function would allow testing each path at once similar to what Process Macro would do in SPSS, however, the jamovi environment can be easier to use.