Proc surveyselect for two-stage cluster sampling

[slowlydoesit]

Hi there, I would really appreciate some help on this issue. I've tried searching through these forums and the documentation as well as google but can't find enough detail or any examples.

My task is to select 15 clusters (PSU) of 30 units (SSU) each. My population contains 50 clusters of varying sizes, and the 15 clusters need to be selected using PPS (or PPS_SYS). The 30 units within these clusters will then be selected using SRS and it's highly likely that a cluster may get selected more than one due to a couple having high n's.

Is there any way I can do this within one step in SAS?

Also considering that a cluster may be chosen twice during the first stage, is there any way to get around having to specify individual cluster sizes even if it means doing this process in more than one step? This is because the number of clusters may change and I would prefer to have this done automatically.

Finally are there any examples of the two-stage cluster procedure using PPS in use?

The documentation mentions SAMPLINGUNIT | CLUSTER variables < / options > ; but I don't understand how that works to process both the cluster stages.

This is the link to the documentation I need assistance with
http://support.sas.com/documentation/cdl/en/statug/63347/HTML/default/viewer.htm#statug_surveyselect...

Thanks, Audrey Message was edited by: slowlydoesit

[mfisher]

I think you will need to do this in more than one step: 1) draw your first stage sample of n=15, using pps and without replacement; 2) merge your sample back with the original data to get your second-stage data included; 3) draw your second stage sample of n=30 from the merged sample, using the first stage information as a stratum, not a cluster. Surveyselect will draw 30 samples from each of your 15 first-stage samples if you treat them as strata.

Regards,
Mark

[slowlydoesit]

Thanks for your response, yes this is how I've been doing it so at least it confirms that the method I've been using seems logical to another person!

It is interesting that I can't get the samplingunit | cluster command to work, as I assume that has the possibility to do the procedure in one step.

[carolet]

I need to create SAS code for the following SPSS code-- It is 2 stage sampling --- does anyone know how to program this in SAS?

<?xml version="1.0" encoding="utf-8" standalone="no"?>

<SPSSComplexSamples version="1.0">

                <Header copyright="Copyright (c) SPSS Inc., 2006. All Rights Reserved."/>

                <AnalysisDesign SRSestimator="wor" numberOfStages="2">

                                <AnalysisStage estimationMethod="equal_wor" label="Stage 1" stageNumber="1">

                                                <StrataVarList numberOfVariables="3">

                                                                <Variable name="region_num"/>

                                                                <Variable name="sub_region_num"/>

                                                                <Variable name="comm_size_num"/>

                                                </StrataVarList>

                                                <ClusterVarList numberOfVariables="1">

                                                                <Variable name="residence_merge"/>

                                                </ClusterVarList>

                                                <IncProb>

                                                                <Variable name="inc_prob_1"/>

                                                </IncProb>

                                </AnalysisStage>

                                <AnalysisStage estimationMethod="equal_wor" label="Stage 2" stageNumber="2">

                                                <StrataVarList numberOfVariables="1">

                                                                <Variable name="age_gender_group"/>

                                                </StrataVarList>

                                                <IncProb>

                                                                <Variable name="inc_prob_2"/>

                                                </IncProb>

                                </AnalysisStage>

                                <Weight>

                                                <Variable name="weight_final"/>

                                </Weight>

                </AnalysisDesign>

</SPSSComplexSamples>