The SURVEYFREQ Procedure

REPWEIGHTS Statement

  • REPWEIGHTS variables </ options> ;

The REPWEIGHTS statement names variables that provide replicate weights for replication variance estimation methods (which include bootstrap, BRR, and jackknife). Each variable should contain the weights for a single replicate, and the number of variables should correspond to the number of replicates. The variables must be numeric, and their values must be nonnegative unless you specify the NEGATIVE option.

You can provide replicate coefficients by specifying the JKCOEFS= or REPCOEFS= option. These options are available for bootstrap and jackknife variance estimation (VARMETHOD=BOOTSTRAP and VARMETHOD=JACKKNIFE), in addition to general replication variance estimation (when the VARMETHOD= option is not specified). When you specify VARMETHOD=BRR, the replicate coefficient is fixed at 1 slash upper R, where R is the total number of replicates.

For information about replicate weights, replicate coefficients, and replication variance estimation, see the sections Bootstrap Method, Balanced Repeated Replication (BRR) Method, Fay’s BRR Method, and Jackknife Method. When you provide replicate weights but do not specify the VARMETHOD= option, PROC SURVEYFREQ computes the variance estimates as described in the section Jackknife Method.

When you provide replicate weights, you do not need to specify a CLUSTER or STRATA statement.

If you specify a REPWEIGHTS statement but do not include a WEIGHT statement, PROC SURVEYFREQ uses the average of each observation’s replicate weights as the observation’s weight.

You can specify the following options:

DF=df

specifies the degrees of freedom for the analysis. The value of df must be a positive number. By default, the value of df is the number of REPWEIGHTS variables. For more information, see the section Degrees of Freedom.

PROC SURVEYFREQ uses df to obtain the t distribution percentile for confidence limits (for proportions, totals, and other statistics). For more information, see the section Confidence Limits for Proportions. PROC SURVEYFREQ also uses df to compute the denominator degrees of freedom for the F statistics in the Rao-Scott and Wald chi-square tests. For more information, see the sections Rao-Scott Chi-Square Test, Rao-Scott Likelihood Ratio Chi-Square Test, Wald Chi-Square Test, and Wald Log-Linear Chi-Square Test.

JKCOEFS=value | <(>values <)> | SAS-data-set

specifies the jackknife coefficients for jackknife variance estimation (VARMETHOD=JACKKNIFE). You can provide a single jackknife coefficient value to use for all replicates, or you can provide a separate value for each replicate by specifying a list of values or a SAS-data-set. The jackknife coefficient values must be nonnegative numbers. For more information, see the section Jackknife Method.

You can provide jackknife coefficients by specifying one of the following forms:

JKCOEFS=value

specifies a single jackknife coefficient value to use for all replicates, where value must be a nonnegative number.

JKCOEFS=(values)

specifies a list of jackknife coefficient values, where each value corresponds to a single replicate that is identified by a REPWEIGHTS variable. You can separate the values with blanks or commas, and you can enclose the list of values in parentheses. The coefficient values must be nonnegative numbers. The number of coefficient values must equal the number of replicate weight variables that you specify in the REPWEIGHTS statement.

List the jackknife coefficient values in the same order in which you list the corresponding replicate weight variables in the REPWEIGHTS statement.

JKCOEFS=SAS-data-set

names a SAS-data-set that contains the jackknife coefficients. You must provide the jackknife coefficients in a variable named JKCoefficient in the SAS-data-set. Each coefficient value must be a nonnegative number. Each observation in the SAS-data-set is assumed to correspond to a replicate that is identified by a REPWEIGHTS variable. The number of observations in the SAS-data-set must not be less than the number of REPWEIGHTS variables.

The order of the jackknife coefficients in the SAS-data-set should match the order in which you list the corresponding replicate weight variables in the REPWEIGHTS statement.

By default, the jackknife coefficient is left-parenthesis upper R minus 1 right-parenthesis slash upper R, where R is the total number of replicates.

NEGATIVE | NEG

removes the restriction that the REPWEIGHTS variable values must be nonnegative. When you specify this option, PROC SURVEYFREQ accepts negative replicate weight values and includes them in the analysis.

By default, replicate weight values must be nonnegative; when PROC SURVEYFREQ encounters a negative replicate weight value, it displays an error message and does not perform the analysis unless you specify the NEGATIVE option.

REPCOEFS=value | <(>values <)> | SAS-data-set

specifies the replicate coefficients for replicate-based variance estimation. You can provide a single coefficient value to use for all replicates, or you can provide a separate value for each replicate by specifying a list of values or a SAS-data-set. The replicate coefficient values must be nonnegative numbers.

This option is available for bootstrap and jackknife variance estimation (VARMETHOD=BOOTSTRAP and VARMETHOD=JACKKNIFE, respectively) in addition to general replication variance estimation (when the VARMETHOD= option is not specified).

When you specify VARMETHOD=BOOTSTRAP, the default value of the replicate coefficient is 1 slash upper R, where R is the total number of replicates. When you specify VARMETHOD=JACKKNIFE, or when you do not specify the VARMETHOD= option, the default value of the replicate coefficient is left-parenthesis upper R minus 1 right-parenthesis slash upper R. For more information, see the sections Bootstrap Method and Jackknife Method.

You can provide replicate coefficients by specifying one of the following forms:

REPCOEFS=value

specifies a single coefficient value to use for all replicates, where value must be a nonnegative number.

REPCOEFS=(values)

specifies a list of coefficient values, where each value corresponds to a single replicate that is identified by a REPWEIGHTS variable. You can separate the values with blanks or commas, and you can enclose the list of values in parentheses. The coefficient values must be nonnegative numbers. The number of coefficient values must equal the number of replicate weight variables that you specify in the REPWEIGHTS statement.

List the replicate coefficient values in the same order in which you list the corresponding replicate weight variables in the REPWEIGHTS statement.

REPCOEFS=SAS-data-set

names a SAS-data-set that contains the replicate coefficients. You must provide the coefficients in a variable named RepCoefficient in the SAS-data-set. Each coefficient value must be a nonnegative number. Each observation in the SAS-data-set is assumed to correspond to a replicate that is identified by a REPWEIGHTS variable. The number of observations in the SAS-data-set must not be less than the number of REPWEIGHTS variables.

The order of the replicate coefficients in the SAS-data-set should match the order in which you list the corresponding replicate weight variables in the REPWEIGHTS statement.

REPWARN <=NO | YES>

controls the log message that PROC SURVEYFREQ provides when a parameter cannot be estimated from one or more replicates. REPWARN or REPWARN=YES displays this log message as a warning. By default, REPWARN=NO, which displays the log message as a note.

For more information, see the subsection Variance Estimation in the section Jackknife Method.

Last updated: December 09, 2022