SDTL Working Group 20 May 2021

 

  1. Change WeightVariable to ExpressionBase

    1. WeightVariable is currently defined as VariableSymbolExpression, which means that it points to a variable. But it is possible for a weight to be an expression including a variable. We can accomodate this by defining WeightVariable as ExpressionBase, which will allow both simple variables and complex expressions.

    2. The range of the WeightVariable will be changed to ExpressionBase.

  2. Case-wise and column-wise deletion options

    1. Commands in Stata and SPSS corresponding to the SDTL Collapse command include options to exclude data by case or by column. Case-wise deletion means omitting all of the data on a row if any variable has a missing value. Column-wise deletion means omitting all of the data in a variable if any value in the column is missing.

    2. We discussed two ways of handling these options in SDTL:

      1. Create additional properties for case-wise and column-wise omission

      2. Model the behavior of these options using other SDTL commands, like DropCases and IfRows

    3. Dan argued for the first option. He pointed out that SDTL provides high-level descriptions of commands. These high-level descriptions are more useful to users than detailed descriptions of lower-level steps required for executing the commands.

    4. We also discussed whether the properties should be true/false or enumerated lists. We decided to use enumerated lists, because statistical languages may implement these operations in different ways. For example, the case-wise deletion option in Stata does not apply to all aggregation functions. An enumerated list can point the user to documentation in the source language for specific features.

    5. Collapse will include properties for CaseWise and ColumnWise deletion properties

      1. The enumeration list for each property will give the name of the source language, so that users can refer to documentation for details about the behavior of the property in context.