WesVar Features

General Features

Usability

• Easy-to-use Windows® interface
• No coding of procedure statements needed
• Simple workbook structure for organizing analysis requests
• Workbook output can be shared without having the source data sets
• Multiple files with replicate weights can be imported in one run
• Preferences control default settings for output, tables, regression models, and file locations

Data Import/Export

• SAS V7/V8/V9
• SAS transport file
• S-Plus
• Microsoft Access®
• Paradox
• Quattro Pro
• ASCII
• SAS for PC/DOS
• SPSS
• STATA
• Microsoft Excel®
• Lotus
• Minitab
• ODBC
• Many more

Working with WesVar Data Sets

• Add variable labels
• Perform recodes
• Subset populations
• Log file for tracking the history of changes made to a data file

Weighting

Create Weights

Import existing replicate weights or create them in WesVar.

Adjust input base weights using:

• Nonresponse adjustments
• Post-stratification
• Raking

Make adjustments for the full sample and separately for each replicate.

Replication Methods of Variance Estimation Supported

Stratified two primary sampling units (PSUs)/stratum

• Balanced repeated replication (BRR)
• Fay’s BRR method (Fay)
• Jackknife 2 (JK2)

Stratified >2 PSUs/stratum

• Jackknife n (JKn)

Unstratified

• Jackknife 1 (JK1)

Maximum Replicates

• BRR, Fay—up to 512 with bundled orthogonal matrices, up to 9,984 with user supplied matrices
• Jackknife (JK1, JK2, JKn)—up to 9,999

Tables

General Features

• Specify up to 8-way tables
• Easily generate all two-way and three-way tables
• Analyze data sets with multiple imputations

Generated Statistics

• Estimates
• Standard errors
• Coefficients of variation
• Confidence intervals
• Wilson method for confidence intervals on extreme proportions
• Unweighted sample size in each table cell
• Design effect for estimate
• Effective sample size
• t statistic for test that estimate is zero, associated p values
• Standardized rates

Computed Statistics

• A statistic that is a function of one or more survey variables or of statistics derived from a survey (e.g., RATIO=INCOME/ASSET). For example, the ratio of two means, the log-odds in a two-way table, and the difference in two ratios can be calculated using computed statistics.
• Other functions include standard operators (+, -, x, /), Log, Exp, Sqrt, Mean, Median, Quantile, and Geometric Mean
• Plausible values can be computed for education assessment tests or for multiple imputation analysis

Tests of Independence

• Pearson’s chi-square
• RS2—Rao-Scott statistic using a design effect adjustment to the Pearson chi-square
• RS3—Rao-Scott statistic based on Satterthwaite approximation

Options for Tables

• Set alpha for confidence intervals
• Set finite population correction factor for variance estimates
• Set degrees of freedom for estimated variances
• Choose method for handling missing values
  • Listwise deletion (complete case analysis)
  • Separate treatment of each analysis variable and computed statistic (available case analysis)
• Control method for computing medians or quantiles
  • Group method
  • No group method (default)
  • Compute direct replicate estimates of standard errors
• Print diagnostic to measure whether replicate contributes heavily to variance estimate

Function Statistics

• A statistic computed as a function of the cell entries in a table, such as the difference between males and females (up to 4-way tables)

Output for Tables

• Requested statistics for all cells and marginals of table
• Statistic type
• Sum of weights
• Percentage of total
• Row percentages
• Column percentages
• Labels for variables and values
• Fixed decimal or scientific notation, with ability to control number of decimal places
• Replicate estimates saved in auxiliary output file

Table Viewer

• Allows users to view the output of a Table Request in a tabular format and copy the data into Microsoft Excel or other programs

Regression

General Features

• Point-and-click selection of dependent and independent variables
• Independent variables—continuous and categorical variables are allowed (class variables can have up to 256 categories)
• Automatic creation of dummy variables
• By variable for subgroup modeling
• Models with or without intercept
• Interactions can be specified among predictors
• Significance of linear combinations of parameters can be tested
• Degrees of freedom can be set for estimated variances

Types of Models

• Linear regression
• ANOVA—up to three-way interactions
• Dichotomous logistic regression
• Multinomial logistic regression (unordered categories)

Linear Regression

• Specify a transformation of a source variable as the dependent variable
• Find parameter estimates using weighted least squares algorithm
• Analyze data sets with multiple imputations

Logistic Regression

• Use modified Newton-Raphson iteration procedure to find parameter estimates
• Specify maximum number of iterations
• Set convergence criterion—relative difference of log-likelihood
• Control starting values for parameter estimates
• Define success value for the dependent variable in dichotomous logistic regression
• Define reference category for the dependent variable in multinomial logistic regression
• Compute odds ratios—estimates and confidence intervals

Output

• Model, error, and total sums of squares
• Parameter estimates
• Standardized linear regression coefficients
• Standard error of parameter estimates
• Confidence intervals for parameters
• t tests on individual parameters
• F tests on main effects and interactions
• User-defined tests on linear combinations of parameters
• Score tests for logistic parameters
• p values for hypothesis tests
• R2 values
  • Linear—squared multiple correlation coefficient
  • Logistic—Cox-Snell, entropy, Estrella
• Correlations of parameter estimates
• Covariances of parameter estimates
• Estimates and confidence intervals for odds ratios in dichotomous and multinomial logistic regression
• Fixed decimal or scientific notation, ability to control number of decimal places
• Replicate estimates saved in auxiliary output file
• Iteration history for logistic saved in auxiliary output file

Descriptive Statistics

Generated Statistics

• Extreme observations—five smallest and five largest unweighted observations
• Influential observations—five most influential weighted observation with respect to the mean and five most influential observations with respect to the total
• Univariate statistics:
  • Weighted and unweighted count of observations, sum, mean, variance, skewness, kurtosis, percentiles, and geometric mean
  • Standard errors for weighted statistics
  • Both replication and Woodruff standard errors for percentiles
  • Minimum, maximum, and unweighted count of missing observations
• Weighted and unweighted correlation coefficients
• Replication standard errors for weighted correlation coefficients