Session Ready
Exercise

Construct validity & reliability

If semTools includes a measure of Cronbach's alpha, why didn't we just use that to check for it and these other measures earlier?

It's wisest to check for reliability before validity because in practice, reliability is a necessary but insufficient condition for validity. Composite reliability, while an alternative measure of reliability, is a more robust measure because it accounts for differences in variances between manifest and latent variables, unlike Cronbach's alpha. Let's dig further into these relationships between measures of reliability and validity, and when to do what.

Our CFA model definition is stored in brand_rep_CFA_model. psych, lavaan and semTools have been loaded into your environment.

Instructions
100 XP
  • Print the defined CFA model as a refresher. Using cat() from base R will add the original line breaks to the model.
  • Find the measures of construct validity for this model as measured by the appropriate semTools function.
  • Print the coefficient alpha of the original brand_rep_9 survey using the appropriate psych function.