The SparsityCheck module prevents the release of tables that contain a high proportion of cells with very low values (0,1, or 2). It applies to interior cells only (totals are not included). If SparsityCheck is enabled, each cross-tabulation result is checked to verify that the table is not too sparse for release.
To configure the module, you need to define the sparsity check thresholds, which must be named ThresholdA and ThresholdB.
The names are case sensitive, and the default values are:
- ThresholdA - 0.25
- ThresholdB - 0.50
The module works as follows:
- c is the number of interior cells in the table.
- c0 is the number of zero interior cells.
- c1 is the number of interior cells of value 1.
- c2 is the number of interior cells of value 2.
Then the table will not be released if:
- c-c0=0 /* table is empty, check first to avoid divide by zero error.
- c1/(c-c0) > ThresholdA (the ratio of cells with value 1, to the total number of cells with non-zero value).
- (c1+c2)/(c-c0) > ThresholdB (the ratio of cells with value 1 or 2, to the total number of cells with non-zero value).
Apply the Plugin to a Dataset
Login to SuperADMIN and create a new method:
FREQcommon property to
true(recommended; this will configure SuperSERVER to base the calculation on the contribution count rather than the cross tabulation results).
Add the Data Control plugin to the method (the name of the plugin,
SparsityCheck, is case sensitive):
Set the plugin properties:
Assign the method to a dataset (in this example we are assigning the method to a dataset with the ID
You can review the method details using the command
cat <dataset_id> methods details <method_id>: