SOLAS Sample Screens

Missing Data Pattern

By right clicking on any cell in the matrix, you can identify the variable and observation details. This feature allows you to study the missing data patterns and helps you choose the most appropriate imputation techniques.

Predictive Model based Multiple Imputation

The predictive information contained in a user specified set of covariates is used to predict the missing values in the variables to be imputed. First, the linear regression model is estimated from the observed data. Using this estimated model, a new linear regression model is randomly drawn from it Bayesian posterior distribution.. This randomly drawn model is used to generate the imputations, which include random deviations from the model's predictions. Drawing the model from its posterior distribution insures that the extra uncertainty about the unknown true model is reflected. The base setup allows the user control over the variables to be used as predictors in the regression model. When the data are categorical, discriminant multiple imputation is used.

Propensity Score based Multiple Imputation

As in the Predictive Model based approach shown above, the Base setup allows the user to decide which variables to impute and the Monotone and Non-monotone setups allow the user control over the variables to be used as predictors in the regression model. In both the Predictive Model based multiple imputation and the Propensity Score based multiple imputation approaches, the Non-monotone portion of missing data is imputed sing the predictive based approach. The user also has substantial control over the donor pool selections and can also choose a refinement variable which will further reduce the donor pool to cases that match on this refinement variable.

Hot Decking-Single Imputation

Imputed values are selected from responders that are similar with respect to a set of auxiliary variables. There are other single imputation methods available in Solas, Predicted Mean Imputation,Last Value Carried Forward,and Group Means.

©2002 Dataxiom Software, Inc.. All rights reserved. e-Mail comments and questions to support@Dataxiom.com