pyequate

pyequate is an open-source package containing different test score equating methods in one location. It includes methods for random groups, common-item/non-equivalent groups, and IRT equating. The version on GitHub will be the most current, with updates pushed to pyPI.

Areas Currently Under Development

• Standard errors of equating for common-item/non-equivalent groups methods
• post-smoothing for equipercentile equating

Required Packages

pyequate requires NumPy, Pandas, SciPy, and Statsmodels.

Public Datasets

A number of publicly available datasets are bundled with pyequate to allow testing of functions. They are also available as CSV files in the src/pyequate/datasets directory on GitHub.

Once pyequate is installed, the datasets can be loaded directly from within the package — no need to download the files separately:

import pyequate

# List the names of all bundled datasets
pyequate.list_datasets()

# Load a dataset by name (returns a pandas DataFrame)
act = pyequate.load_dataset("ACTmath")

The following datasets are included:

ACTmath, ADMneatX, ADMneatY, CBdatax1y2, CBdatax2y1, KBneatx, KBneaty, Math20EG, Math20SG, PISAbooklets, PISAitems, PISAstudents, itemparx, itempary