18 Exercise 7: Test your knowledge
In this Exercise, we will work with the pmi data set again. The data set contains 123 observations of 6 variables. There are four variables of interest to us in this data set:
cond: article about sugar shortage was placed at the front (1) or interior (0) page of a newspaper
pmi: belief that others will be prompted to buy sugar as a result of exposure to the article (scale: 1-7)
import: belief that the article is on an important topic (scale: 1-7)
reaction: intention to buy sugar (scale: 1-7)
Let’s load processR
and save the pmi data to a source object called data:
# installing/loading the package:
if(!require(processR)) {
install.packages("processR");
require(processR)
#load / install+load processR
}
<- processR::pmi data
Throughout this exercise, you will test two competing theories:
- Following the first theory, we hypothesize that the effect of placement of the article (front page condition, X = 1; interior page condition, X = 0) on intentions to buy sugar is mediated by both the belief that others will be prompted to buy sugar (pmi) and perceived topic importance (import). We assume that presumed media influence and topic importance are unrelated and that the two mediator variables independently influence the intention to buy sugar. This is a parallel multiple mediator model:
- Following the second theory, we hypothesize that both presumed media influence and perceived topic importance are related. We assume that perceived topic importance influences the belief that others will be prompted to buy sugar because of the article. This means that we don’t assume that topic importance has a direct effect on buying intentions, but only an indirect effect. This is a serial multiple mediator model.
18.1 Task 1
Fit both models in lavaan
and compare their fit.
18.2 Task 2
Create a visual representation of your two models.