Interaction effect in SmartPLS

[nielsvanbruggen]

Dear Smart-PLS users,

For my M.Sc study in Marketing, my thesis is about the effect of social media messages (eWOM) on (attitudinal/behavioural) loyalty of consumers. My research design is based on Experimental Vignette Method. I've spread two surveys, each with a pre-test, an intervention (positive/negative) and a post-test. N=279 (140/139). The structural model is attached in a JPEG-file.

The most important hypothesis in this research is that positive eWOM has a significantly different effect on consumers' (level of loyalty) than negative eWOM (so it's a moderator variable).

My question:
Does anyone know how I can calculate the interaction effect of eWOM in SmartPLS? I do have comparison of the means on each pre- and post-test, which give me a first indication that this hypothesis could be supported, but now I want to know if the differences are significant. I thought about doing two-way ANOVA tests in SPSS, but I rather do the analysis in Smart-PLS.

Can anyone help me with this?

Thank you in advance. :)

Greets,
Niels

[cringle]

Thanks

You can compute the interaction effect using the moderator option in SmartPLS. Right-click on the target construct, select the option and choose your settings from the menu.

Please take a look at the moderation chapter in this book: https://www.smartpls.com/documentation/ ... on-pls-sem

Kind regards
Christian Ringle