Mingyu 's Data Set
I would like the columns documented.
A data set that have one row per kid per question
A data set that has one row per kid per action
I would a table for problems (add something to indicate parent for ML) Its one row per problem with links back to the main assisment it comes from if tis scaffolding It should have number of hint. Number of answers Question type.
Then we need to dump for skills. We should give out two different skill tables for easy of processing.
One should be a matric where columns indicate a skill, and row are for each problem. The value in the madtris that is non-zero indicated that question is tagged with that skill.
A second way of dumping that data is to have a table that one row for each non-zero value in the matrix. Since the matris will be so parse this is easier sometimes.
We should for each probelm sets also dump the prety print of the the problem set so go along.
In fact if you do hit show section labels then you can see the full structure of the problem set. http://www.assistment.org/build/print/sequence/6250?mode=custom&op_scaf=true&op_hint=true&op_answer_op=true&op_answer=true&op_name=true&op_buggies=true&op_sections=true&short_answers=false
We should make a table of problem sets. A row indcates a problem set. Columns include stuff like - this probem set has or has not heircahyr. You should be able to quickly see which problem sets hare complicated. and which are kust simple mastery learning or simple linear or simple random order. --A nother is
Can you add the secoitn idea Matt?
We want to know the different between a mastery learning problem set and non mastery.
Each data set need to list the years its for.
So add all the data set I can think of from 2004-05 , 2005-06 etc. Neil is trying to have one public repositity.
We need a set of files in xml that capcautre the basics of a table. That is the hierarchical structure of the problet set and the types for each section (random, linear, choose condtion) . It should also include
O=If you use one of these data sets please give an acknowledgment to the paper that that data set is from. If you use the general data set where there are no publication please set the book chapter "Advanced in Intelligent Tutoring Systems that will be edited by Roger Nkambou."
- Here are two data sets we used for ITS'2010 submitted paper. (http://nth.wpi.edu/pubs_and_grants/ITS%202010/Feng%20Can%20We%20Get%20Better%20Assessment.pdf)
- one row per student
- Year of ASSISTments usage: 2004-2005 (DA0405.txt) and 2005-2006 (DA0506.txt)
stu_id: student ID
original_count: number of main problems done in static condition
%correct: %correct on main problems in static condition
MCAS: MCAS score
DA_original_count: number of main problems done in dynamic condition
DA_original_%correct: %correct on main problems in dynamic condition
DA_scaf_%correct: %correct on scaffolding questions in dynamic condition
DA_avg_time: average number of seconds spent on each question (main or scaffold) in dynamic condition
DA_avg_attempt: average number of attempt made on each question (main or scaffold) in dynamic condition
DA_avg_hint: average number of hints requested for each question (main or scaffold) in dynamic condition