[Andrew Ross]

Under “Edit Attribute Properties”, one of the options for “type” is qualitative . My first thought was: how is this different from categorical? I’m asking because I’m writing a plugin that will (hopefully) compute different types of correlation/association (Pearson r, eta-squared or Cramer’s V) for column pairs of various data types.

I played around a little: in the Mammals dataset, I set “diet” to qualitative. This resulted in the usual words “meat”, “plants”, “both” not being displayed in the case table or card view; instead, an orange bar spanning the column appears. But a graph with Diet as the x variable seems unaffected.

My guess is that “qualitative” was originally meant for things like text or sentences, where it’s unlikely that you’d have two or more items that are exactly the same. But then it just gets handled like categorical, mostly.

So I guess my plugin could either treat it as categorical, or just not try to compute anything about it. I’m curious to hear your thoughts.

Thanks,

Andrew

[ I also tried searching the codap code in github for it but didn’t learn much: https://github.com/search?q=repo%3Aconcord-consortium%2Fcodap+qualitative&type=code&p=1 ]

[Zac Opps]

Good question, Andrew. I’ll let others weigh in with more detail, but the info on the Qualitative data type found on the Edit Attribute Properties how to guide might be helpful background:

qualitative – numeric data displayed using relative rather than absolute measures. In the example below (attached), the percentage of people in each state with graduate degrees has been converted from a numeric to a qualitative attribute, so the percentages are represented as propotionally sized red bars (the % is highest in DC and lowest in FL) rather than as precise numbers.

Attachments:

1. 

   numeric-data.png
   

[Andrew Ross]

Thanks! Sounds like I should actually treat it like Numeric then.

[Author]

Posts