IVEware: Imputation and Variance Estimation Software, Version 0.3
IVEware developed by the Researchers at the Survey Methodology Program, Survey Research Center, Institute for Social Research, University of Michigan performs:
- Imputations of missing values using the Sequential Regression (also known as Chained Equations) Method;
- Multiple imputation analyses for both descriptive and model-based analysis;
- Analysis that accounts for complex design features, weighting, clustering and stratification;
- Other applications of multiple imputation such as disclosure limitation, combining information from multiple data sources, Bayesian analysis through prediction, causal inference and measurement error.
IVEware is organized into seven modules to perform various tasks. The six core modules are IMPUTE, BBDESIGN, DESCRIBE, REGRESS, SYNTHESIZE and COMBINE and the seventh module, SASMOD, is specific to SAS.
- IMPUTE uses a multivariate sequential regression approach (Raghunathan et al (2001), Raghunathan (2015)). This approach is also called Chained Equations, (Van Buuren and Oudshoorn (1999)) and Fully Conditional Specification (Van Buuren (2012)) and is used to impute item missing values or unit non-response. IMPUTE can create multiply imputed data sets and can handle continuous, categorical, count and semicontinuous variables.
- BBDESIGN implements the weighted finite population Bayesian Bootstrap approach to generate synthetic populations from complex survey data. The primary goal is to incorporate weighting, clustering and stratification in a nonparametric approach for generating the non-sampled portion of the population from the posterior predictive distribution, conditional on the observed data and the design information. For more details see Zhou, Elliott and Raghunathan (2015, 2016a, 2016b)
- DESCRIBE estimates population means, proportions, subgroup differences, contrasts and linear combinations of means and proportions. A Taylor Series Linearization approach is used to obtain variance estimates appropriate for a user-specified complex sample design. Multiple imputation analysis can also be performed when there is missing data.
- REGRESS fits linear, logistic, polytomous, Poisson, Tobit and proportional hazard regression models. For data resulting from a complex sample design, the Jackknife Repeated Replication technique is used to obtain variance estimates. As in other IVEware commands, a multiple imputation analysis can be performed when there are missing values.
- SYNTHESIZE uses the multivariate sequential regression approach to create full or partial synthetic data sets to limit statistical disclosure (See Raghunathan, Reiter and Rubin (2003), Reiter (2002) and Little, Liu and Raghunathan (2004) for more details.) All item missing values are also imputed when creating synthetic data sets. However, DESCRIBE, REGRESS and SASMOD modules cannot be used to analyze synthetic data sets as they DO NOT implement the appropriate combining rules. Examples of implementation of correct combining rules for synthesized data sets are included in later sections of this guide.
- COMBINE is useful for combining information from multiple sources through multiple imputation. Suppose that Data 1 provides variables X and Y, Data 2 provides variables X and Z and Data 3 provides variables Y and Z. COMBINE can be used to concatenate the three data sets and multiply impute the missing values of X, Y and Z to create large data sets with complete data on all three variables. All item missing values in the individual data sets will also be imputed. The multiply imputed combined data sets can be analyzed using DESCRIBE, REGRESS and SASMOD modules (see Schenker, Raghunathan, and Bondarenko (2010) for an application and Dong, Elliott and Raghunathan (2014a, 2014b) for more details).
- SASMOD (requires SAS) allows users to take into account complex sample design features when analyzing data with selected SAS procedures. Currently the following SAS PROCS can be called: CALIS, CATMOD, GENMOD, LIFEREG, MIXED, NLIN, PHREG, and PROBIT. A multiple imputation analysis can be performed when there are missing values. Unlike the other IVEware modules, SASMOD requires SAS.
IVEware can be used with SAS, STATA, SPSS and R packages or as a standalone in Windows, Linux or Mac OS (except SAS) operating systems.
© The Regents of the University of Michigan, 2016. All rights reserved. Permission is granted to use, copy and redistribute this software for any purpose, so long as no fee is charged and so long as the copyright notice above, this grant of permission, and the disclaimer below appear in all copies made; and so long as the name of the university of michigan is not used in any advertising or publicity pertaining to the use or distribution of this software without specific, written prior authorization. Permission to modify or otherwise create derivative works of this software is not granted. This software is provided as is, without representation as to its fitness for any purpose, and without warranty of any kind, either express or implied, including without limitation the implied warranties of merchantability and fitness for a particular purpose. The regents of the University of Michigan shall not be liable for any damages, including special, indirect, incidental, or consequential damages, with respect to any claim arising out of or in connection with the use of the software, even if it has been or is hereafter advised of the possibility of such damages.