Need help in coding (Factorial Survey Method)

[Nawaf Rasheed]

Hello everyone.
This is my first post in this forum. I am having a problem running the code for a replication study” A scenario-based factorial survey approach”. The model is similar to the one in the pic. All my IVs and the moderators are categorical and the DV is continuous. If someone can help me to the write the code it will be appreciated.

Thank you.

[Clyde Schechter]

As you don't provide example data, nor even an explanation of the design of data collection, it is hard to be specific. But at some level it seems you are looking at a regression model with two predictors that are your main interest, and five effect modifiers. So with some appropriate regression command it will boil down to something like this:

regression_command DV i.(perceived_distrib_injustice perceived_proc_injustice) ## ///
    i.(injury_denial victim_denial ledger_metaphor perceived_sanction_severity perceived_sanction_certainty), ///
    ADD APPROPRIATE OPTIONS DEPENDING ON DETAILS OF DATA COLLECTION

If you are new to Stata and unfamiliar with the ## notation, read -help fvvarlist-.

Let me point out that this is a very complicated model and I think you will have considerable difficulty interpreting it. With five dichotomous effect modifiers, you will have 32 distinct effects of each of the injustice variables (and if the effect modifiers are polytomous it will be even worse!). From the labeling of the arrows in the diagram, I get the sense that perhaps you intend only to look at the signs of the baseline effects and effect modifications. While this will reduce the cognitive burden of interpreting the model it does so by sacrificing important details. For example, assuming denial of injury increases the effect of perceptions of procedural injustice and perceived sanction severity decreases it, what happens when both are present? Without quantitating all of the combinations you will have only the most superficial understanding of the complexity here.

[Nawaf Rasheed]

Clyde Schechter, Thank you for your reply. I am new to this forum; I thought I will get a notification email if someone replied to my question.
The data were collected using "Qualtrics". We created 36 scenarios based on our model. Each subjects will got 3 to 4 random scenarios.
Here is how we based our 36 scenarios:
- Three IVs: Perception of Distributive Injustice,

Perception of Procedural Injustice, and no technique of injustice (Reference group).
- Deterrence theory as moderators, 4 levels (Perceived Sanction Severity; high or low and Perceived Sanction Certainty; high or low)
- Neutralization theory as moderators, 4 levels: Denial of injury, Denial of Victim, Metaphor of ledger, and No neutralization technique.
3x3x4 = 36 scenarios.
This is the code I tried to run but it is not giving me the expected result and I am not sure whether it is right or not:

mixed BI Procedural Distributive ProceduralXD2 ProceduralXD3 ProceduralXD4 DistributiveXD2 DistributiveXD3 DistributiveXD4 ProceduralXNInjury ProceduralXNVictim ProceduralXNMLedger DistributiveXNInjury DistributiveXNVictim DistributiveXNMLedger Gen|| Respondant: ProceduralXD2 ProceduralXD3 ProceduralXD4 DistributiveXD2 DistributiveXD3 DistributiveXD4 ProceduralXNInjury ProceduralXNVictim ProceduralXNMLedger DistributiveXNInjury DistributiveXNVictim DistributiveXNMLedger Procedural Distributive
__________________________________________________ __________________________________________________ ________________________________________

BI = DV
D1 = (High Severity , High Certainty)
D2 = (Low Severity , High Certainty)
D3 = (High Severity , Low Certainty)
Gen = Gender
I also included the interaction terms there.

Again, thank you.

[Clyde Schechter]

Well, it is very difficult to keep track of that many interaction terms. Who knows if you left one out somewhere along the line? Or included something that doesn't belong? The way to avoid this problem is to use Stata's factor variable notation. Read -help fvvarlist- and then model your code on what I showed in #2. The code will be much shorter and, more importantly, much clearer: you will be able to see clearly if you have the right code for your model.

What does seem problematic in your code is that you are using -mixed- (which is appropriate given the repeated measures) but you have not specified a person-level in the model. All you have is the bottom level effects. You need to have -|| PersonID:- there (or whatever the name of your variable identifying persons is.)

Added: Try to implement this on your own. If you need help with the coding, when you post back, be sure to show example data, and use the -dataex- command to do that. If you are running version 16 or a fully updated version 15.1 or 14.2, -dataex- is already part of your official Stata installation. If not, run -ssc install dataex- to get it. Either way, run -help dataex- to read the simple instructions for using it. -dataex- will save you time; it is easier and quicker than typing out tables. It includes complete information about aspects of the data that are often critical to answering your question but cannot be seen from tabular displays or screenshots. It also makes it possible for those who want to help you to create a faithful representation of your example to try out their code, which in turn makes it more likely that their answer will actually work in your data.


I sure hope you have enough people in your study to seriously estimate a model this complicated.