tsostarics / anrlab Goto Github PK

Data collection instruments are itemized, and the summary score instruments aggregate by summing over the number of correct items within tasks and conditions. Sometimes these scores are utilized not as raw counts, but as proportions. The tricky part of this task is accurately finding the correct denominator size, as it varies in size from 2 to 30+, and items are incorporated into multiple aggregate scores.

There are a few ways this might be done:

1. Hard coding each calculation (dispreferred)
2. Grouping by subtest and type within each instrument (I anticipate there will be some scores that aren't able to be done this way)
3. Extracting the number of field names specified in each calculated field, which should be available in the data dictionary (probably the most robust way, but will need to handle aggregates that are sums of other aggregates differently, recursion might help)