Listwise Deletion: It’s NOT Evil

[Richard Williams]

http://www.statisticalhorizons.com/l...n-its-not-evil

This isn't a Stata-specific topic per se. But, the list does get several posts about using multiple imputation, and I am often left wondering whether (a) MI is really necessary, or (b) whether it is even appropriate in that specific case. Allison gives a nice short discussion of key points concerning these issues.

For those not familiar with it, Allison's blog has a lot of other great short pieces. At least in Sociology, I think Allison is one of the best people in the world at bringing statistics to the masses. See

http://www.statisticalhorizons.com/blog

[Richard Williams]

Here is one quote in particular that really struck me. Maybe the rest of the world already knew this, but I didn't.

"What most researchers don’t know is that listwise deletion may actually be less biased than multiple imputation or ML when data are missing on predictor variables in regression analysis. For example, suppose you’re doing a linear regression in which the dependent variable is number of children and one of the predictors is annual income. Suppose, further, that 30% of respondents did not report their income. Multiple imputation or ML, when done correctly, can produce approximately unbiased estimates of the regression coefficients when income data are MAR. But so will listwise deletion. And, remarkably, listwise deletion will produce unbiased estimates even if the data are not missing at random."