Hello guys!
My problem is as follows: I am now writing my dissertation which is concerned with the extent to which customers' purchase intention is affected by information provided by anonymous vs familiar person. Independent variables in the study include argument quality, credibility etc. and purchase intention is dependent variable.
So, let's say I have four variables.
SocialMedia Argument Quality which predicts SocialMedia Purchase Intention
Anonymous Argument Quality which predicts Anonymous Purchase Intention.
Now, I would like to compare which effect is stronger. Normally, if there was only one dependent variable, I would simply run a multi regression and check the standardised beta to gain insight into which predictor impacts outcome more... However, here I have two independent variables and two dependent. How to approach this situation?
In the past, a similar study was conducted (Erkan & Evans 2016, the influence of eWOM...) and authors seem to have conducted multi regression - although they measured purchase intention twice for anonymous and familiar sources, they somehow computed a single variable out of it and then use both socialmedia argument quality and anonymous argument quality as independent variables, then simply taking standardised beta. Is that even a valid approach? How to compute a single purchase intention as a dependent variable if there are in fact two variables reflecting it (anonymous and familiar source purchase intention?
I would gladly appreciate help, as its crucial for my dissertation.
All the best!
"Multivariate analysis is needed when there are 2 or more Dependent Variables (DV) are in your research model. Base module of SPSS (i.e. without add-on module) can't handle multivariate analysis. Don't know exactly the research model with how many Independent Variable(s) & how they connected to the 2 DV. Very likely Structural Equation Modeling (SEM) can be used. However, if you intend to use variance-based SEM / Partial Least Squares (PLS) then you might use SmarPLS etc. software. If you intend to use Covariance-based SEM then you might want to use AMOS (add-on module from SPSS), LISREL etc." -copied
Agree with Dewan's point of view
Multivariate analysis (MVA) is a Statistical procedure for analysis of data involving more than one type of measurement or observation. It may also mean solving problems where more than one dependent variable is analyzed simultaneously with other variables. (cited)
Multivariate analysis is proper for your study. You can also track the relevant papers and information by means of google scholar.
Hope the above is helpful. Thanks.
I would like to add more information and hope it can help
Your questions above will be solved by your research model or your research questions.
For example: cited the example you mentioned above:
SocialMedia Argument Quality (SAQ) which predicts SocialMedia Purchase Intention (SPI)
Anonymous Argument Quality (AAQ) which predicts Anonymous Purchase Intention (API).
1, you can do basic regression twice and compare the Beta value and P-value. This is the same with using SEM. You can report the unstd. and std. results and report which IV influences which DV more. It depends on you how to report. U can discuss with your supervisor.
2, You can also use SEM. For example:
I have completed a SEM model for you. Unfortunately, this web page can not allow me to copy it here. Let me explain to you. You can draw a model: SAQ-->SPI-->PI; AAQ-->API-->PI; the common DV is PI, and the IDs are SAQ and AAQ, SPI and API as mediator variables. You need add the common DV (Purchase Intention). Subsequently, you can compare them.
It also depends on you number of sample in a bid to use which software (AMOS, LISREL or Mplus etc or SmartPLS) to analysis your data.
This is just my viewpoints. Hope it can help. Thank you very much. Have a nice day.
Might be a Partial dependence plot (PDP) that helps you to fix it.
Through PDP you can understand and visualize how the model predictions are influenced by each predictor when all other predictors are being controlled. Partial dependence plots are used to interpret the output of complex machine learning models.
I can help you to run PDP in your case.
Let me know if you have any questions.
Thanks
Best Regards
Arfan