Contributing data to the IRW

We welcome contribution of new datasets to the IRW!

Contributing data

If you would like to contribute data, please get in touch at itemresponsewarehouse@stanford.edu.

We will take information about any data that you think may be worth including in the IRW. Inclusion will be easier if the data is already in a format consistent with the IRW data standard. Below we provide information on how to construct such a dataset.

Constructing an IRW-compliant dataset

Below are critical instructions for formatting data for the IRW. More details are available here.

  1. Numeric values of a response should be meaningful. For example, missing values cannot be coded as numbers (e.g., -9).

  2. Please check that the id and item identifiers have an appropriate number of unique values.

  3. Additional data elements (e.g., rt, date, wave, rater, qmatrix) should all be formatted as specified above.

  4. If there are multiple scales available, the responses need to be split into multiple files (one per scale). If multiple groups are assessed via the same scale, these data can be put into a single file (if desired, a column indicating group membership can be added).

  5. Before sending a file, check it with our validator, the same checks we run before a table is added to the IRW. The quickest way is to drop the file on the browser validator: it needs nothing installed, and your file stays on your computer. From the command line it needs Python but no account or credentials:

    pip install irw-validate
irw-validate mydata.csv

    Both report two things: whether the file conforms to the IRW Data Standard, and whether it passes the IRW’s own intake checks, such as the minimum of 100 respondents. Each problem is an error (must be fixed) or a warning (worth a look; some are legitimate). R users can use the browser validator, or run a smaller set of core checks with source("https://raw.githubusercontent.com/ben-domingue/irw/main/misc/validate_irw.R") followed by validate_irw(df, "mydata.csv").

While we have tried to offer generic guidance on formatting data to the IRW standard, there are innumerable idiosyncrasies that may merit additional conversation. To discuss specific issues associated with formatting your data to the IRW standard, please feel free to reach out to us at itemresponsewarehouse@stanford.edu. We would be happy to talk more! You can also use the IRW Dataset Builder that will help port data to the IRW data standard if you like.