Previous Page | Next Page

The MI Procedure

Regression Method for Monotone Missing Data

The regression method is the default imputation method for continuous variables in a data set with a monotone missing pattern.

In the regression method, a regression model is fitted for a continuous variable with the covariates constructed from a set of effects. Based on the fitted regression model, a new regression model is simulated from the posterior predictive distribution of the parameters and is used to impute the missing values for each variable (Rubin 1987, pp. 166–167). That is, for a continuous variable with missing values, a model

     

is fitted using observations with observed values for the variable and its covariates , , ..., .

The fitted model includes the regression parameter estimates and the associated covariance matrix , where is the usual inverse matrix derived from the intercept and covariates , , ..., .

The following steps are used to generate imputed values for each imputation:

  1. New parameters and are drawn from the posterior predictive distribution of the parameters. That is, they are simulated from , , and . The variance is drawn as

         

    where is a random variate and is the number of nonmissing observations for . The regression coefficients are drawn as

         

    where is the upper triangular matrix in the Cholesky decomposition, , and is a vector of independent random normal variates.

  2. The missing values are then replaced by

         

    where are the values of the covariates and is a simulated normal deviate.

Previous Page | Next Page | Top of Page