Researchers have utilized DataShop to explore learning issues in a variety of educational domains. These include, but are not limited to, collaborative problem solving in Algebra (Rummel, Spada, Diziol, 2007), self-explanation in Physics (Hausmann & VanLehn , 2007), the effectiveness of worked examples and polite language in a Stoichiometry tutor (McLaren, Lim, Yaron, & Koedinger, 2007) and the optimization of knowledge component learning in Chinese (Pavlik, Presson, & Koedinger , 2007).

Datashop integrates the results of the AFM (Additive Factor Model) algorithm, a logistic regression performed over the “error rate” learning curve data. The AFM logistic regression, a standard regression bounded between 0 and 1, attempts to find the best-fit curve for error-rate data, which also ranges between 0 and 1. The results of this model are shown as the "predicted learning curve" on each line graph of student error rate. The predicted learning curves are the average predicted error of a skill over each of the learning opportunities.

The Model Values page in DataShop presents a quantitative analysis of how well, given the selected knowledge component model, the AFM statistical model fits the data (via AIC, BIC, log likelihood) and how well it might generalize to an independent dataset from the same tutor (via cross validation RMSE).

For more on the Additive Factor Model, see Is Over Practice Necessary? Improving Learning Efficiency with the Cognitive Tutor through Educational Data Mining (Cen, Koedinger, and Junker 2007).

Data is collected from the seven PSLC courses (Algebra, Chemistry, Chinese, English, French, Geometry and Physics) and various studies. There are also sources external to the PSLC that contribute to DataShop, such as middle school math data from the Assistment project at WPI.

If you are developing a course or study and have not yet collected data, then you probably want to log student-tutor transactions as they occur. The page Logging New Data describes a number of scenarios where you would log data from your course or study to DataShop.

If you are interested in storing and viewing data from a course or study that has occurred in the past, then you probably want to import the existing data. The page Importing New Data describes the two main types of data that can be imported into DataShop: XML files and tab-delimited text files.

To gain access to private data sets, first create an account (see “How do I access DataShop?” above), then visit the Other Datasets tab and click the “Request Access” button next to the name of the project you would like to access. In the dialog that appears, enter a brief reason for why you would like access. The request for access will be sent to the project's principal investigator and data provider (if one exists). The status of your request will be shown on the Access Requests page. Any projects for which you have been given access will appear on the My Datasets page.

If you're not sure what data you need, please contact us and we'll do our best to help.

Tutors which produce log data but do not log directly to the PSLC server, such as Andes (Physics LearnLab) or the Carnegie Learning Cognitive Tutors (Algebra and Geometry LearnLabs) must go through a collection and conversion process. The length of this process depends on the availability of the personnel to collect and anonymize the data, as well as the state of the program needed to run the conversion. Also note that conversion of extremely large datasets can add time.

If you need a dataset urgently, please contact us and put "urgent" in the subject of the email.