Psychometric analytics refers to the statistical analysis of test data to evaluate the quality of measurement, by closely examining the performance of items, subscores, and the test as a whole. For example:
If the correct answer is C but lots of the smart students chose B?
Did the smarter students choose C or B?
Is the item biased with differential item functioning?
Unlike most platforms that provide a perfunctory analysis, we provide a deep, professional level report, with options for item response theory, which is the paradigm for all modern assessment.
What proportion/number of students selected each answer, and what was the point-biserial?
What was the overall difficulty and discrimination? IRT parameters? Poor items are automatically flagged for you.
What was the mean, standard deviation, reliability, standard error, and other key stats for each subscore?
Descriptive stats, multiple reliability indices, test information function, CSEM function, and much more.
Automatically pull classical test theory analytics, which is based on simple mathematics like proportions, averages, and correlations. This allows you to quickly and effectively improve items, especially via discussion with subject matter experts.
IRT is far more complex than CTT, but solves some more advanced psychometric issues far more effectively. This includes adaptive testing, automated test assembly of parallel forms, and equating.
