Closed jpschewe closed 4 weeks ago
If the new report looks good, then remove the normalization code from the software and merge.
Thinking about the scaling calculation, find the range of possible values for all categories and pick the largest range. Then scale all scores to this.
The range if a category is maximum_value - minimum_value
.
Scaled score is ( ( score - category.minimum_value ) * max_range ) / category.range
Now that judging is done to the rubric, there is less need for normalization. However we still need scaling to be able to sum scores across categories. To facilitate this we need to have the maximum performance score in the XML description and then we can scale everything to a common range and generate a new final scores report.
subjective_computed_scores
final_scores