Can someone help me understand path diagram and path analysis?

P

I'm currently conducting a study on the factors affecting treatment outcomes of MDR-TB patients. My independent variables include: gender, place of residence, age, civil status, income, occupation, substance abuse and co-morbidities while treatment outcome is my dependent variable. One way to analyze my data is through the use of multiple regression path analysis. In regards to path analysis, how can I identify the exogenous and endogenous variables under my study? Do i need to conduct the study and tabulate the results before I can identify these exogenous and endogenous variables? And is it possible to create a path diagram before the study has been conducted? If you could provide me a good read on path analysis i would appreciate it. Thanks

C

You have already identified your exogenous and endogenous variables - your exogenous ones are the independent variables (e.g., gender, age etc) and your endogenous variable is your dependent variable (e.g., treatment outcome). The terms are used interchangeably. The path model you talk of is just your hypothesised model, so you have already created it. You may use multiple regression analysis to test your model once you have collected data if your dependent variable is 'continuous'. If it is 'categorical/ordinal/binary' (e.g., yes/no or successful/not successful), then logistic regression may be better. You should probably go on a statistics course on multiple regression analyses - I'm not allowed to post the links to the ones I'd recommend, but if you message me I'll tell you

P

and since i want to know the indirect effects of my dependent variable, can some of my independent variables be considered as endogenous variable?

C

If you're looking at indirect effects (or mediators), yes, some of your independent variables may effectively become endogeneous (or dependent) in those analyses. In a path model, it would look like this: Independent variable -> Mediator -> Dependent variable.

26717