I am using the Jaccard Index to compare the overlap in items between two sets of instruments. However, the index is strongly influenced by the number of items in each instrument (for example instrument A items=5; instrument B items=12) - see below for an example.
Is there any way to account for the different lengths between the two sets/instruments?
Thanks in advance, Sophia
Example: Instrument A: Items: 1 0 0 1 0
Instrument B: Items: 1 0 1 1 1 1 1 1 1 0
JI = 2/(0+6+2) JI = 0.25
Since instrument B has much more items the overlap is maybe underestimated since instrument A is just a "shorter" instrument and focused on the key items to measure the construct of interest.