Overview (CSPLAN command)

CSPLAN creates a complex sample design or analysis specification that is used by companion procedures in Sampling and Testing. CSSELECT uses specifications from a plan file when selecting cases from the active file. Analysis procedures in Sampling and Testing, such as CSDESCRIPTIVES, require a plan file in order to produce summary statistics for a complex sample. You can also use CSPLAN to view sample or analysis specifications within an existing plan file.

The CSPLAN design specification is used only by procedures in Sampling and Testing.

Options

Design Specification. CSPLAN writes a sample or analysis design to a file. A sample design can be used to extract sampling units from the active file. An analysis design is used to analyze a complex sample. When a sample design is created, the procedure automatically saves an appropriate analysis design to the plan file. Thus, a plan file created for designing a sample can be used for both sample selection and analysis.

Both sample and analysis designs can specify stratification, or independent sampling within nonoverlapping groups, as well as cluster sampling, in which groups of sampling units are selected. A single or multistage design can be specified with a maximum of three stages.

CSPLAN does not actually execute the plan (that is, it does not extract the sample or analyze data). To sample cases, use a sample design created by CSPLAN as input to CSSELECT. To analyze sample data, use an analysis design created by CSPLAN as input to Complex Samples procedures, such as CSDESCRIPTIVES.

Sample Design. A variety of equal- and unequal-probability methods are available for sample selection, including simple and systematic random sampling. CSPLAN offers several methods for sampling with probability proportionate to size (PPS), including Brewer's method, Murthy's method, and Sampford's method. Units can be drawn with replacement (WR) or without replacement (WOR) from the population. At each stage of the design, you can control the number or percentage of units to be drawn. You can also choose output variables, such as stagewise sampling weights, that are created when the sample design is executed.

Analysis Design. The following estimation methods are available: with replacement, equal probability without replacement, and unequal probability without replacement. Unequal probability estimation without replacement can be requested in the first stage only. You can specify variables to be used as input to the estimation process, such as overall sample weights and inclusion probabilities.

Operations

  • If a sample design is created, the procedure automatically writes a suitable analysis design to the plan file. The default analysis design specifies stratification variables and cluster variables for each stage, as well as an estimation method appropriate for the chosen extraction method.
  • CSPLAN writes design specifications in XML format.
  • By default, CSPLAN displays output that summarizes the sample or analysis design.

Subcommand Order

  • The first DESIGN subcommand must precede all other subcommands except PLAN, PLANVARS, and PRINT.
  • PLAN, PLANVARS, and PRINT subcommands can be used in any order.

Limitations

  • A maximum of three design blocks can be specified.
  • CSPLAN ignores SPLIT FILE and WEIGHT commands with a warning.