PROC MIXED Method=ML and missing data (FIML?)

[halladje]

Hello, 

I came across a paper that stated they used maximum likelihood to handle missing data using PROC MIXED. I am wondering:

(1) How does maximum likelihood estimation account for missing data in the analysis (documentation states: "A favorable theoretical property of ML and REML is that they accommodate data that are missing at random (Rubin 1976; Little 1995)." but I am not sure why/how. 

(2) What is the difference between the ML in SAS versus FIML in Mplus? Does ML use all available data (even when there are missing values within an observation) to predict the outcome similar to FIML? Does ML retain all observations that have something observed or does it listwise delete individuals with missing predictors but retain observations with complete predictors and missing outcome s(similar to default FIML in Mplus)? 

Thanks in advance, 

Jillian 

[seeff]

1. There are a lot of resources around the web to explain how ML handles missing data, but here is a nice paper from a SAS forum: http://www.statisticalhorizons.com/wp-content/uploads/MissingDataByML.pdf

2. FIML and ML are different terms for the same thing. "Missingness" in the sense which you are using it is referring to the outcome, not the predictor variables. So, yes, PROC MIXED deletes an observation if a predictor variable is missing, but retains an observation if predictors are valid but outcome is missing.