|
|
Regression with multiple dependents |
|
Message/Author |
|
Sandra Lyons posted on Wednesday, February 10, 2010 - 2:09 pm
|
|
|
I want regress a 13dependent variables (5 continuous and 8 dichotomous)on a set of 24 dependents with missing data. What is the advantage of using Mplus simultaneous regression over OLS/Logit in some other software? For example: (a) adjusted p for multiple tests, (b) smaller standard errors because correlated error terms are taken into account, (c)other? |
|
|
Those are relevant - and with missing data on the DVs, for any given regression the simultaneous analysis helps by drawing on information from the other DVs not missing for that subject (under MAR). |
|
Sandra Lyons posted on Wednesday, February 10, 2010 - 9:21 pm
|
|
|
Thank you. I'm looking forward to the Mplus short course in March. |
|
|
Dear mplus team I have a quick query How is missing data handled in a straight forward regression, such as in path analysis with continuous variables, if the LISTWISE option in DATA is not used? Many thanks for your help |
|
|
Mplus provides maximum likelihood estimation under MCAR (missing completely at random) and MAR (missing at random; Little & Rubin, 2002) for continuous, censored, binary, ordered categorical (ordinal), unordered categorical (nominal), counts, or combinations of these variable types. |
|
|
Thanks! |
|
Back to top |
|
|