SEM Flashcards
What is SEM?
A confirmatory approach looking at the fit between observed data and a predefined model.
What is a latent variable?
An unobserved theoretical construct which is indicated by observed variables that load onto it.
What is an endogenous variable?
A latent variable influenced by another latent variable, or a DV.
What is an exogenous variable?
A latent variable that predicts fluctuations in another latent variable, an IV.
What is the focus of a measurement model?
The relationship between a latent variable and its measured variables.
What can CFA help us do?
Check that items contribute meaningfully to a factor, refine models and compare theoretical models.
What does Chi square assess?
The probability that the model fits the population.
What CFI indicates good fit?
> .95
What CFI indicates adequate fit?
> .90
What RMSEA indicates good fit?
< 0.05.
What RMSEA indicates adequate fit?
.05 - .08
What SRMR indicates good fit?
< 0.08.
What is the key extra step in a full latent variable model?
The researcher specifies the impact of one latent variable on another.
When interpreting latent variable relationships, what do standardised path coefficients convey?
A change in X number of SDs of endogenous variable is predicted by a change of 1 standard deviation in the exogenous variable.
For standardised path coefficients, what is a small effect size?
.05