Mediation Analysis with Categorical IV with 4 levels

#1
Hello all,
I would like to do a mediation analysis using the following 3 variables:

1. Integration Strategy: This is the independent variable. It is categorical and has 4 levels
2. Perceived Pressure to Assimilate: This is the mediator variable. It is continuous.
3. Readjustment Difficulty: This is the dependent variable. It is continuous.

Is it possible to run a mediation analysis with a categorical IV with more than two levels at all?
Do I use dummy coding for the IV?

I am familiar with running mediation analysis with continuous variables, but I am completely confused as to how to go about wih these variables. :confused:
Please help me! I'd really appreciate it!
 
E

EliteResearch

Guest
#2
Go to: http://www.afhayes.com/spss-sas-and-mplus-macros-and-code.html

Download the PROCESS plug for SPSS [Hayes, A. F. (2012). PROCESS: A versatile computational tool for observed variable moderation, mediation, and conditional process modeling.]

It is easy to use- everything is entered in a single dialogue box and I think you can use categorical IVs with more than two levels.

This program uses bootstrapping methods (in fact, the Baron/Kenny steps to mediation are starting to be replaced by Preacher & Hayes bootstrapping mediation).

According to the website, "PROCESS uses an ordinary least squares or logistic regression-based path analytical framework for estimating direct and indirect effects in simple and multiple mediator models, two and three way interactions in moderation models along with simple slopes and regions of significance for probing interactions, conditional indirect effects in moderated mediation models with a single or multiple mediators and moderators, and indirect effects of interactions in mediated moderation models also with a single or multiple mediators. Bootstrap methods are implemented for inference about indirect effects in both unmoderated as well as moderated mediation models. PROCESS extends what my existing macros can do by vastly expanding the number and complexity of models that can be estimated and pieced together in the form of a conditional process model. It also has the ability to estimate moderated mediation and mediated moderation models with multiple mediators, multiple moderators of individual paths, interactive effects of moderators on individual paths, and models with dichotomous outcomes."

-Amanda
1925 E. Beltline Rd. Suite 200
Carrollton, TX 75006
(800) 806-5661 (972) 538-1374
consulting@eliteresearch.com
 
#3
Hi,

Can someone confirm that this is ok? I am trying to do a mediation analysis with a categorical IV (6 levels), a continuous M and a continuous DV. Is the PROCESS macro ok to run this as default?

Any help is GREATLY appreciated.

Neil
 
#4
yes this is correct. See the Youtube video by Andy Field where you shows a quick example with a moderator (still applicable) with multiple levels.
 
#5
I have a similar problem, I am using PROCESS with SPSS and I want to run a moderation with an IV with 3 levels (a,b,c). In Andy Field's tutorial he shows how to do so with an IV with 2 levels, but I am not sure how to proceed when having 3 levels. Hayes's website gives a an explanation on how to do (see below). So far the best solution I came up with is to create 2 data sets, one data set having IV with 2 levels (a,b) and then run a moderation; and another data set having IV with 2 levels (a,c) and run another moderation. I am using "a" as my reference group. Am I doing it completely wrong? Thanks for your help.

Hayes' explanation:

It is possible with PROCESS or INDIRECT but not with SOBEL. Using PROCESS or INDIRECT, the best approach is to dummy code the categorical IV. If your IV has k categories, construct k-1 dummy variables and then run INDIRECT or PROCESS k-1 times. With each run, make one dummy variable the IV and the other one(s) the covariate(s). You will not get a single test of the indirect effect, but you will get indirect effects for each category relative to the reference category in the dummy coding scheme. See the documentation.