Message/Author 


How to set up the syntax for the model (path analysis) which has both Exogenous categorical variable and Exogenous continuous variables in path analysis for continuous outcome variable (endogenous variable? There are two types of Exogenous categorical variable: Binary and nominal (1= BA 2= MA C=PHD 4=OTHER) 


The model is estimated conditioned on the observed exogenous variables in regression and path analysis. They can be binary or continuous. In both cases, they are treated as continuous. You should create a set of dummy variables for the nominal variable. 


Hello Dr, Muhen, You mean the way to build syntax is not different between continuous and continuous exogenous variable. For example, MODEL=NOMEANSTRUCTURE; INFORMATION=EXPECTED; model: model: y3 on x1 x2 y1 y2; y1 on x1 x2; y2 on x1 x2; This is normal way to build syntax in a path model. In other words, regardless of the type of scale of exogenous variables (binary or continuous), the way to write syntax is not changed... Am I on the track? Thanks 


How to interpret the covariance/correlation between two binary exogenous variables in a path analysis? Is this a pearson correlation? 


Yes, you don't distinguish between exogenous variables due to their scale. A Pearson correlation. 


Can you explain why Mplus doesn't distinguish between binary or continuous for exogenous variables. I would like to report a correlation table of my path model but am not sure if Pearson's is the appropriate correlation coefficient for the exogenous binary variable included. Would it be ok to report Peason's in a correlation table even though one variable is not continuous? Thank you! 


Mplus doesn't distinguish between binary or continuous for exogenous variables because that is not done in regular regression. You can if you like put a binary exog vble on the Categorical list, but that makes an assumption of an underlying normal distribution for this variable which is not necessary. You can present regular Pearson correlations for the binary variables, just being clear that that is what they are. It isn't a perfect summary of the data with categorical variables (raw data is) but it's still informative. 

Back to top 