search

Qub3k / subjective-exp-consistency-check

Software implementation of the subjective experiment consistency checking framework (based on a p-value P–P plot).
MIT License
2 stars 0 forks source link

Process data with stimulus identifier not unique across different experiments #4

Closed Qub3k closed 3 years ago

Qub3k commented 3 years ago

This PR introduces the functionality of processing data that must be grouped by both stimulus ID and experiment ID. (The old code assumed that grouping by stimulus ID is sufficient.)

One has to group by both stimulus ID and experiment ID if input data contains stimuli with non-unique IDs across many experiments. Naturally, this makes sense only if your input data contains results from more than one experiment.

Please keep in mind that code in this PR still plots only one P–P plot for all input data. If your input data includes responses from multiple experiments then the resulting P–P plot is much harder to interpret.