- Analysis of Variance
- Bayesian Analysis
- Categorical Data Analysis
- Cluster Analysis
- Descriptive Statistics
- Discriminant Analysis
- Distribution Analysis
- Exact Inference
- Finite Mixture Models
- Group Sequential Design and Analysis
- Longitudinal Data Analysis
- Market Research
- Missing Data Analysis
- Mixed Models
- Model Selection
- Multivariate Analysis
- Nonlinear Regression
- Nonparametric Analysis
- Nonparametric Regression
- Post Processing
- Power and Sample Size
- Predictive Modeling
- Psychometric Analysis
- Quantile Regression
- Regression
- Robust Regression
- Spatial Analysis
- Standardization
- Structural Equations Models
- Survey Sampling and Analysis
- Survival Analysis
- SAS/STAT Procedures A-Z

The CAUSALTRT procedure implements causal inference methods that are designed primarily for use with data from nonrandomized trials or observational studies. The CAUSALTRT procedure provides methods for estimating causal treatment effects controlling for confounding between the outcome and characteristics of the subjects. Specially, the procedure estimates the average causal effect of a binary treatment on a continuous or discrete outcome in nonrandomized trials or observational studies in the presence of confounding variables.

You can adjust for confounding by modeling the treatment assignment or the outcome or both. Modeling the treatment assignment leads to inverse probability weighting methods, and modeling the outcome leads to regression adjustment methods. Modeling both leads to doubly robust methods that can provide unbiased estimates for the treatment effect even if one of the models is misspecified.

The following highlights are available in the CAUSALTRT procedure:

Provides the following methods to estimate causal effects

- inverse probability weighting methods
- regression adjustment
- doubly robust methods

Estimates two types of causal effects:

- average treatment effect (ATE), also sometimes called the average causal effect (ACE)
- average treatment effect for the treated (ATT or ATET)

Computes standard errors and confidence intervals for the causal effects by the following methods:

- asymptotic methods
- bootstrap methods (which you request by specifying the BOOTSTRAP statement)

Provides the following types of graphical output:

- diagnostic plots for the propensity score model, including various plots of propensity scores or weights
- histograms for bootstrap estimates

Saves the propensity scores, inverse probability weights, and the predicted potential outcomes in a SAS data set.

Performs BY group processing, which enables you to obtain separate analyses on grouped observations.

Supports the following diagnostics for assessing the balance produced by a propensity score model:

- weighted and unweighted standardized mean differences and variance ratios for the covariates used to fit the propensity score model
- plots of the propensity scores and weights by treatment conditions
- weighted and unweighted kernel density plots for the continuous covariates in the propensity score model

For further details see the SAS/STAT User's Guide:
The CAUSALTRT Procedure

( PDF | HTML )