MIMIC model

Question 1

where is covariance of items??

Answer 1

in the causes, not the indicators. the indivators are locally independent. The causes indicators may covary

Question 2

Explain two sides of mimic model

Answer 2

1. indicators = reflective
2. causes = formative

Question 3

How specialise a variable?

Answer 3

• no residual variance of indicator
• loading = 1

Question 4

what do we expect from predicors

Answer 4

The observed predictors (i.e., covariates) need not be normally distributed…

• the observed predictors can be
  
  • categorical
    
    • e.g., group membership, gender, or education level
  • continuous

Question 5

using the MIMIC model to compare latent means assumes

Answer 5

strict measurement invariance. We can only test intercept of variables

Question 6

the regression paths represent
differences in means between groups
* i.e., for a binary observed predictor this is
equivalent to

Answer 6

the latent intercept 𝛼! in
the second group in the context of
homogeneity testing