Prediction

[Amanda Lane]

I have created a path, simplified below:

Awareness -> Knowledge -> Usage

Where Awareness and Knowledge are formative and made up of two indicators each.

In Market Research we would assume advertising has an impact on Awareness, but it may also have a direct impact on Knowledge. If I was to take an uplift in the indicators that make up Awareness and feed them through the path to find the impact on Knowledge and Usage, this is easy. But what if at the same time, I want to increase the two indicators that make up Knowledge. Will there be an additive effect on Knowledge (once from Awareness increasing and once from the indicators that make up Knowledge increasing)?

I know this is partly a philosophical question, but would appreciate your thoughts on this from a technical point of view.

Thanks,
Amanda.

[vonbergh]

Hi Amanda,

I interpret your model to be as follows:

• Advertising -> Awareness
  Advertising -> Knowledge
  Awareness -> Knowledge
  Knowledge -> Usage

As a result awareness would be mediating the relationship Advertising -> Knowledge. You may check the following literature on mediation:

• Baron (1986) The moderator-mediator variable distinction in social psychological research
  Frazier (2004) Testing moderator and mediator effects in counseling psychology research
  MacKinnon (2002) A comparison of methods to test mediation and other intervening variable effects
  McClelland (1993) Statistical difficulties of detecting interactions and moderator effects
  Muller (2005) When Moderation Is Mediated and Mediation Is Moderated
  Shrout (2002) Mediation in Experimental and Nonexperimental Studies- New Procedures and Recommendations

With regards to increasing the number of items for awareness and knowledge, this may increase the reliability of the factors and may contribute to the GoF of the model.

Good luck,
Dennis

[Amanda Lane]

Thanks very much, you have interpreted my path correctly. I will take a look at the literature you suggest.

Amanda.