The BCHOICE Procedure

Displayed Output

This section describes the displayed output from PROC BCHOICE. For a quick reference of all ODS table names, see the section ODS Table Names. ODS tables are arranged in four groups, listed in the following sections: Model- and Data-Related ODS Tables, Sampling-Related ODS Tables, ODS Tables Related to Posterior Statistics, and ODS Tables Related to Convergence Diagnostics.

Model- and Data-Related ODS Tables

Model Information

The "Model Information" table (ODS table name ModelInfo) displays basic information about the model, such as the name of the input data set, response variable, model type, sampling algorithm, burn-in size, simulation size, thinning, and random number seed. The "Model Information" table is produced irrespective of the sampling technique.

The "Model Information" table also displays the random number seed that is used to initialize the random number generators. If you repeat the analysis and specify this seed value in the SEED= option in the PROC BCHOICE statement, you get an identical stream of random numbers. This table is displayed by default.

Class Level Information

The "Class Level Information" table (ODS table name ClassLevels) lists the levels of every variable that is specified in both the CLASS statement and the MODEL statement. You should check this information to make sure that the data are correct. You can adjust the order of the CLASS variable levels by specifying the ORDER= option in the CLASS statement. You can suppress the "Class Level Information" table completely or partially by specifying the NOCLPRINT option in the PROC BCHOICE statement.

Choice Sets Summary

The "Choice Sets Summary" table (ODS table name ChoiceSummary) is displayed by default and should be used to check the data entry. It shows the numbers of choice sets, the number of alternatives, and the number of chosen and unchosen alternatives. You should always check this summary table to ensure that the data are arrayed correctly. This table is displayed by default.

Number of Observations

The "NObs" table (ODS table name NOBS) shows the number of observations in the data set and the number of observations that are used in the analysis. By default, observations that have missing values are not used. This table is displayed by default.

Sampling-Related ODS Tables

Burn-In History

The "Burn-In History" table (ODS table name BurnInHistory) shows the scales and acceptance rates for each parameter block in the burn-in phase for logit models. The table is not displayed by default, but you can request it by specifying MCHISTORY=BRIEF or MCHISTORY=DETAILED in the PROC BCHOICE statement.

Parameters Initial Value

The "Parameters Initial" table (ODS table name ParametersInit) shows the initial value of each fixed regression coefficient. This table is not displayed by default, but you can request it by specifying the option INIT=PINIT in the MODEL statement.

Prior for Fixed Effects

The "Prior for Fixed Effects" table (ODS table name CoeffPrior) shows the prior distribution for the regression coefficients. The table is displayed by default for a model with fixed effects.

Sampling History

The "Sampling History" table (ODS table name SamplingHistory) shows the scales and acceptance rates for each parameter block in the main sampling phase for logit models. The table is not displayed by default, but you can request it by specifying MCHISTORY=BRIEF or MCHISTORY=DETAILED in the PROC BCHOICE statement.

Tuning History

The "Tuning History" table (ODS table name TuningHistory) shows the number of tuning phases used in establishing the proposal distribution. The table also displays the scales and acceptance rates for each parameter block at each of the tuning phases for logit models using random walk Metropolis algorithm. For more information about the self-adapting proposal tuning algorithm, see the section Tuning the Random Walk Metropolis in Logit Models. The table is not displayed by default, but you can request it by specifying MCHISTORY=BRIEF or MCHISTORY=DETAILED in the PROC BCHOICE statement.

ODS Tables Related to Posterior Statistics

PROC BCHOICE calculates some essential posterior statistics and outputs them to a number of ODS tables. Some of the ODS tables are produced by default, and you can request others by specifying an option in the PROC BCHOICE statement. For more information about the calculations, see the section Summary Statistics.

Summary and Interval Statistics

The "Posterior Summaries and Intervals" table (ODS table name PostSumInt) contains basic summary and interval statistics for each parameter. The table lists the number of posterior samples, the posterior mean and standard deviation estimates, and the highest posterior density (HPD) interval estimates. This table is displayed by default.

Summary Statistics

The "Posterior Summaries" table (ODS table name PostSummaries) contains basic statistics for each parameter. The table lists the number of posterior samples, the posterior mean and standard deviation estimates, and the percentile estimates. This table is not displayed by default, but you can request it by specifying the option STATISTICS=SUM.

Correlation Matrix

The "Posterior Correlation Matrix" table (ODS table name Corr) contains the posterior correlation of model parameters. This table is not displayed by default, but you can request it by specifying the option STATISTICS=CORR.

Covariance Matrix

The "Posterior Covariance Matrix" table (ODS table name Cov) contains the posterior covariance of model parameters. This table is not displayed by default, but you can request it by specifying the option STATISTICS=COV .

Deviance Information Criterion

The "Deviance Information Criterion" table (ODS table name DIC) contains the deviance information criterion (DIC) of the model. This table is not displayed by default, but you can request it by specifying the option DIC . For more information about the calculations, see the section Deviance Information Criterion (DIC).

Interval Statistics

The "Posterior Intervals" table (ODS table name PostIntervals) contains the equal-tail and highest posterior density (HPD) interval estimates for each parameter. The default $\alpha $ value is 0.05, and you can change it to other levels by using the STATISTICS= option. This table is not displayed by default, but you can request it by specifying the option STATISTICS=INT.

ODS Tables Related to Convergence Diagnostics

PROC BCHOICE provides convergence diagnostic tests that check for Markov chain convergence. The procedure produces a number of ODS tables that you can request and save individually. For information about calculation, see the section Statistical Diagnostic Tests.

Autocorrelation

The "Autocorrelations" table (ODS table name AUTOCORR) contains the first-order autocorrelations of the posterior samples for each parameter. The Parameter column states the name of the parameter. By default, PROC BCHOICE displays lag 1, 5, 10, and 50 estimates of the autocorrelations. You can request different autocorrelations by using the DIAGNOSTICS=AUTOCORR(LAGS=) option. This table is displayed by default.

Effective Sample Size

The "Effective Sample Sizes" table (ODS table name ESS) calculates the effective sample size of each parameter. For more information, see the section Effective Sample Size. This table is displayed by default.

Monte Carlo Standard Errors

The "Monte Carlo Standard Errors" table (ODS table name MCSE) calculates the standard errors of the posterior mean estimate. For more information, see the section Standard Error of the Mean Estimate. This table is not displayed by default, but you can request it by specifying the option DIAGNOSTICS=MCSE.

Geweke Diagnostics

The "Geweke Diagnostics" table (ODS table name Geweke) lists the results of the Geweke diagnostic test. For more information, see the section Geweke Diagnostics. This table is not displayed by default, but you can request it by specifying DIAGNOSTICS=GEWEKE .

Heidelberger-Welch Diagnostics

The "Heidelberger-Welch Diagnostics" table (ODS table name Heidelberger) lists the results of the Heidelberger-Welch diagnostic test. The test consists of two parts: a stationary test and a half-width test. For more information, see the section Heidelberger and Welch Diagnostics. This table is not displayed by default, but you can request it by specifying DIAGNOSTICS=HEIDEL .

Raftery-Lewis Diagnostics

The "Raftery-Lewis Diagnostics" table (ODS table name Raftery) lists the results of the Raftery-Lewis diagnostic test. For more information, see the section Raftery and Lewis Diagnostics. This table is not displayed by default, but you can request it by specifying DIAGNOSTICS=RAFTERY .