Overview

Here are the entities that define the mechanisms by which raw user responses are converted to numeric results, whether in the context of "scoring" a "validated instrument" or "grading" a "test".

Specific Entities

• Scales (as_scales) map 1..n Items into some kind of post-submission processing that generates numerical scores. Illustrations are "grading a test -> % right" or Likert scoring that takes ordinal responses and normalizes them to a 0-100 scale. We abstract out any subsequent interpretation of a numeric score into nominative "bins" (like grades "A, B, ..., F" or designations "Good, Bad, Ugly") to Scale-Interpretations. Scales themselves have:
  • scale_id
  • name
  • definition
  • algorithm_name
  • call_back - proc that takes a list of user Item-Data and returns the Score
  • min_value - bottom of scale's range -- useful for calculating user's "percentage"
  • max_value - top of scale's range -- useful for calculating user's "percentage"

• Scale Item Map (as_scale_item_map) maps Items to a Scale, which is obvious and doesn't require more comment.

• Scale-Scale Map (as_scale_scale_map) however isn't so obvious and does require comment. There are types of scales that are calculated from other scale results, not raw results. For instance, a "total effect" scale might be the arithmetic mean of several other scales scores.