I am very new to SPSS and I am trying to understand how and what type of statistical test that I should use for moderation analysis. For the first part of analysis, I will be doing Pearson correlation analysis to assess the relationship between the IV and DV. However, Im not too sure about the next step to analyse the effect of the moderator variable in the relationship between IV and DV. Please find the attached conceptual framework of my thesis.Can someone guide me on the exact next step to be taken ?
Dear Villasini Velautham ,
I have a number of videos on Mediation/Moderation analysis using Smart PLS. Dont Forget to "Subscribe" my channel.
PLS Estimations/Prediction and Hypothesis Testing including validity and Reliability
The following video explain how the measurement model is development using zero order for one independent and one dependent variable.
https://youtu.be/7uKIBrX71Cc
The following video explain the mediation and moderation analysis using smart pls software at zero order level.
https://youtu.be/4FAq4_8QIxM
The following video explains the Smart PLS procedure to be adopted for the estimation of simple model using second order construct.
https://youtu.be/Cuh8gOsIf6g
The following video explain the steps and procedure adopted for the estimation and reporting of 2nd order model with the inclusion of mediation using Smart PLS.
https://youtu.be/Xsf4Q24KzIo
The following video explain the procedure of structural equation modeling using 2nd order moderation with the help of Smart PLS software.
https://youtu.be/zOnrIeNhurY
@Villasini Looking at your model, it seems that your independent variable (effective project management practices) and dependent variable (project success) both are second order constructs. While your Moderator (a first order latent construct) has 24 observed items which too many and would definitely lead to identification problems while running structural model. I suggest you to reduce the number observed variables for the moderator and use Smart-PLS for testing the model because your model is based on second order latent constructs and AMOS won't work with latent constructs while testing moderation. Wishing you best with your work.
Dear Villasini Velautham it is very normal for a novice researcher to feel confused regarding basic academic research issues. Looking at your conceptual model, first of all I think you need to understand about the variable types i.e., Observed Vs Latent, Unidimensional Vs Multidimensional, and, if latent, Reflective Vs Formative. The way you have drawn your model conveys that your IV and DV both are multidimensional Reflective-Formative Type 2 constructs, which may not be the case in real and just drawn without the understanding about Reflective or Formative constructs. Anyways, though performing moderation would be a simple process if all three variables are observed, in your case you must have to justify the structure of your variables first by doing Reliability and Validity process (Called Measurement model analysis or CFA) in a SEM package. I agree with Imran Anwar that SmartPLS can easily help you out in your case. Tq.
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude