Default scoring mechanism
The default scoring mechanism of the Revise Text item is not "Score per item" like for most other item types, but "Score per error". This is the most logical default, as "Score per item" would imply that you need to identify, correct and categorize all errors perfectly in order to get the maximum score. As soon as you would make one mistake, you would get the minimum score (normally 0).
How to get the maximum score?
Suppose you use the following score per error:
- 1 point when correct
- 0 points when wrong
How can a user get 1/1 for an error, as he needs to do 3 things:
- identify the error by selecting it
- enter the correction, and
- select the correct error category?
When you use categories
When you use categories, the user gets 1/3 of the maximum score per error for each of these 3 actions (identify, correct, and categorize). You cannot change that in assessmentQ.
For example:
- if the user properly selects the error, enters a good correction, but selects the wrong error category, he will get 2/3 or 0.66 for that error.
- if the user properly selects the error, does not enter a good correction, and selects the wrong error category, he will get 1/3 or 0.33 for that error.
When you don't use categories
When you don't use categories, the user gets 1/2 of the maximum score per error for each of the 2 possible actions (identify and correct the error). You cannot change that in assessmentQ.
For example: if the user properly selects the error, but does not enter a good correction, he will get 1/2 or 0.5 for that error.
Grading
The users' answers and automatic scores for a Revise Text item are visible in the Grading module. Graders can still override the automatically generated score there.
More information