WEIGHT Subcommand (CTABLES command) 

 /WEIGHT=varname

If you have a variable that represents adjustment weights rather than frequency weights, you can use that variable as an effective base weight variable. The concept of effective base or effective sample size weighting is based in methods for the analysis of data from complex samples. An effective base weight allows for approximate handling of statistical inference in analyses involving ad hoc adjustments to data from simple random sampling designs via the use of adjustment weights.

  • The effective base weight affects weighted summary statistics values and column means and column proportions significance tests.
  • If weighting is turned on for the dataset (WEIGHT command), the dataset weight variable is ignored and results are weighted by the effective base weight variable.
  • The effective base weight variable must be numeric.
  • Cases with negative weight values, a weight value of 0, or missing weight values are excluded from all results.
  • The WEIGHT subcommand is a global subcommand. The subcommand applies to all tables specified on the CTABLES command.
  • The WEIGHT subcommand must appear before the first TABLE subcommand.

Weights and Rounding

Data can be weighted with the WEIGHT command or with the WEIGHT subcommand in the CTABLES command.

  • If weighting with the WEIGHT command is used, non-integer weights are rounded at the cell or category level for significance tests, confidence intervals, and standard errors.
  • If weighting with the WEIGHT subcommand of the CTABLES command is used, non-integer weights are not rounded.
  • If both are specified, CTABLES uses the variable specified on the WEIGHT subcommand, and non-integer weights are not rounded.