# SUDAAN Course Descriptions

## SUDAAN One-Day Descriptive Procedures Course

Description of Training
The one day SUDAAN Descriptive Procedures Course provides an in-depth presentation of SUDAAN's descriptive procedures: PROC CROSSTAB and PROC DESCRIPT. It also provides an introduction to PROC VARGEN. This course will prepare students for running SUDAAN descriptive procedures on any operating system in which SUDAAN is available.

Introduced in SUDAAN 11, PROC VARGEN is a new descriptive statistics procedure. This procedure computes point estimates and their associated design-based variances for user-defined parameters that can be expressed as complex functions of estimated means, totals, ratios, percents, population variances, population standard deviations, and correlations. Point estimates can be computed within subgroups and subgroup contrasts can also be estimated in similar fashion to other descriptive procedures in SUDAAN.

What is Covered?
A brief introduction to cluster-correlated data is presented, as well as detailed instructions for data input and data analysis. The course covers the major SUDAAN programming statements for the descriptive procedures, CROSSTAB, DESCRIPT, and VARGEN. Each lecture includes extensive examples and a discussion of results. Course content covers survey sampling applications, as well as correlated data analysis from observational studies.

What is Included?

• A three-ring binder of class notes, hands-on lab exercises, and complete examples.
• A CD-ROM containing an electronic version of the SUDAAN Manuals, plus data sets and results for the hands-on lab exercises, and
• A printed version of the SUDAAN Manuals will be available to interested attendees.

After attending this course, you will be able to:

• Understand when and why SUDAAN is appropriate for your data
• Specify your data for input to SUDAAN
• Specify your sample design and the appropriate variance estimation procedure in SUDAAN
• Estimate descriptive statistics and their standard errors from complex sample survey data: e.g., means, totals, proportions, ratios
• Estimate one-way and multi-way frequency distributions and Cochran-Mantel-Haenszel statistics from complex sample survey data
• Perform basic contrasts among a set of estimated means, totals, and proportions

## SUDAAN Modeling Procedures Course

Description of Training
The one day SUDAAN Modeling Procedures Course, with hands-on computer labs, provides an in-depth presentation of SUDAAN's modeling procedures:  PROC REGRESS, PROC LOGISTIC, and PROC MULTILOG. Each course is taught by statisticians who use SUDAAN for analysis of clustered data on a regular basis. Small class sizes allow one-on-one interaction. This course will prepare students for running SUDAAN modeling procedures on any operating system in which SUDAAN is available. Prospective students should have taken the one day SUDAAN Descriptive Procedures Course or have knowledge about basic SUDAAN functions.

What is Covered?
A brief introduction to survey data and a sketch of the theory for parameter and variance estimation used in the SUDAAN software package is presented, as well as detailed instructions for data analysis. The course covers the major SUDAAN programming statements for the modeling procedures, REGRESS, LOGISTIC, and MULTILOG. Each lecture includes extensive examples and a discussion of results. Applications of SUDAAN for cluster-correlated and repeated measures data from epidemiology, clinical trials, and toxicology experiments are also demonstrated.

What is Included?

• A three-ring binder of class notes, hands-on lab exercises, complete examples, and notes on the variance estimation theory underlying SUDAAN
• A CD-ROM containing an electronic version of the SUDAAN Manuals, plus data sets and results for the hands-on lab exercises, and
• A printed version of the SUDAAN Manuals will be available to interested attendees.

After attending this course, you will be able to:

• Understand when and why SUDAAN is appropriate for your data
• Specify your sample design and the appropriate variance estimation procedure in SUDAAN
• Estimate model statistics and their standard errors from complex sample survey data: e.g., betas, model based odds ratio, and predicted marginals.
• Perform basic contrasts among a set of estimated betas or marginals.
• Apply the same set of procedures in SUDAAN for the analysis of clustered data in non-survey situations: e.g., repeated measures and longitudinal data problems in clinical trials, epidemiology, toxicology, and other public health applications.

## Computing Weight Adjustments and Deriving Imputations with SUDAAN Course

Description of Training
The 3-day weighting and imputation course, with hands-on computer labs, provides an in-depth presentation of computing weight adjustments using PROC WTADJUST and PROC WTADJX, and deriving imputations using PROC IMPUTE. This course will prepare students for performing these pre-analysis tasks in SUDAAN. Prospective students should have knowledge about basic SUDAAN functions. Each course is taught by statisticians who use SUDAAN for analysis of clustered data on a regular basis. Small class sizes allow one-on-one interaction. This course will prepare students for running SUDAAN imputations and weighting procedures on any operating system in which SUDAAN is available.

What is Covered?
A brief introduction to cluster-correlated data is presented, as well as detailed instructions for data input and data analysis. The course covers the major SUDAAN programming statements for the weighting and imputation procedures, WTADJUST, WTADJX and IMPUTE. Each lecture includes extensive examples and a discussion of results.

What is Included?

• A three-ring binder of class notes, hands-on lab exercises, and complete examples,
• A CD-ROM containing an electronic version of the SUDAAN Manuals, plus data sets and results for the hands-on lab exercises, and
• A series of short handouts on selected topics (e.g. references for further reading).
• A printed version of the SUDAAN Manuals.

After attending this course, you will be able to:

• Understand when and why SUDAAN is appropriate for your data,
• Specify your data for input to SUDAAN,
• Estimate weighting adjustments, including nonresponse adjustments, post-stratification adjustments, and weight truncation using a model-based weighting approach,
• Perform multivariate and multiple imputations using the weighted sequential hot deck imputation method, and
• Perform imputations using other methods such as cell means imputation and model-based imputation.