Difference between revisions of "Condition in Student-Step Rollup"
|Line 20:||Line 20:|
See [[DataShop ]]
Revision as of 14:08, 29 October 2010
Status: Done (DataShop v4.0 December 2009)
As a researcher, I'd like to see the condition included in the student-step rollup table, so that I can compare across conditions using the data in this format.
- "Is the condition supposed to be included in the step rollup table? Datasets, 'Example Wilkinsburg Spring 2007' and 'Example CWCTC Spring 2007' don't seem to have them in the step rollup data." -- Vincent 4/25/2008
- "In my experience, when you analyze data from an experiment, you (pretty much) invariably want to compare what was different between your conditions. The reason for having the tutor data is typically to examine whether any differences, between the conditions, in the students' learning outcomes (i.e., pre/post test gains) can be explained in terms of differences during their learning process (i.e., the differences in the log data - more errors, more hints, steeper learning curves, etc.). Without the condition data you would not be able to do this. (And so with the step rollup table, it means we need to extract the condition information from transaction table, and then insert it in the step rollup table. Doable, but work that the DataShop could have done for us, especially given that I would think that this is an oft-recurring need.)
- So I would say anytime you export something from DataShop, having the condition in there is a good thing.
- (BTW, is there an easy way to get a student-condition table out of the DataShop? If that were so, then having the condition information in any and all exports is less critical.)"
- Alida, 4/30/2008: You can create a sample per condition, then select all those samples and then export the Step Rollup Table. Sample is one of the columns in the Step Rollup Table and export. This would be better than trying to get the same information from a transaction export.
- Ken, 8/14/2009: To use the step roll-up table in 'R', the condition is important in this table. Also take out the '#' sign from the headers as that is a comment and causes problems in 'R'. [Ken Meeting 8/14/2009]
- How to show the conditions, multiple columns? one column? use labels or 1's and 0's? -- Koedinger, DataShop Team Meeting, Aug 28, 2009
- Variable number of condition columns in step rollup is OK, but so is concatenating condition names with a delimiter in between them.
- Vincent votes for a single column describing condition.
- By definition, each student is in one condition, so a single column should suffice. (I believe what you are describing is a "factor" in a factorial design. That could be useful but could always be derived from the condition in a very straightforward manner. And vice versa. So to keep things simple I'd go with a single column. -- Vincent email 9/3/2009