search

DLCIncluded / MQ

Monotropism Questionnaire - I am not a doctor. This is not diagnosing anything. Questions taken from: https://osf.io/4wru2
28 stars 6 forks source link

Questions marked N/A should probably be excluded from the scoring entirely #7

Closed ethereal-engineer closed 1 year ago

ethereal-engineer commented 1 year ago

At the moment, a question answered with N/A reduces the overall score without also reducing the total that the score is divided by. If a question is N/A (or Not Applicable), it means that the question cannot be answered usefully by the participant. Although there is an edge case where this kind of logic could lead to a significantly biased score (mark all but a handful as N/A), in most cases, this would make the scoring more accurate.

In short, if a question is marked N/A, subtract the maximum possible score from that question - from the total divisor, and set the score for that question to zero.

DLCIncluded commented 1 year ago

Hey there! While I see your point, I am not sure if that is how the original assessment was designed to be scored. I have reached out to one of the co-author's to confirm this point and see if it is valid, thank you for the suggestion! Once I have an answer I'll reply here 😄

DLCIncluded commented 1 year ago

I have a commit ready to push once this is confirmed, if this is indeed the way it should be scored If the value is 0 aka N/A, it subtracts 1 from the number of questions(starting at 47). So if you have 2 N/A, it will go through and add everything, and divide by 45

Before this change if you had 5 N/A your avg. score would be 3.87 (assuming none of the n/a are reversed scored 182/47) EX: image

After this change if you had 5 N/A your avg. score is 4.33 (assuming none of the n/a are reversed scored 182/42) EX: image

Waiting on confirmation from @Oolong before comiting these changes

oolong commented 1 year ago

Yes, I think go for it. It definitely shouldn't be treated as a 'super-strong disagree'!

On Wed, 26 Jul 2023, 00:28 David Cary, @.***> wrote:

I have a commit ready to push once this is confirmed, if this is indeed the way it should be scored If the value is 0 aka N/A, it subtracts 1 from the number of questions(starting at 47). So if you have 2 N/A, it will go through and add everything, and divide by 45

Before this change if you had 5 N/A your avg. score would be 3.87 (assuming none of the n/a are reversed scored 182/47) EX: [image: image] https://user-images.githubusercontent.com/23517228/256059769-022183d5-1b0a-4d47-8850-76a0d8ad0136.png

After this change if you had 5 N/A your avg. score is 4.33 (assuming none of the n/a are reversed scored 182/42) EX: [image: image] https://user-images.githubusercontent.com/23517228/256059718-b929398b-dff8-43d4-9805-0a6d0bfddc31.png

Waiting on confirmation from @oolong https://github.com/oolong before comiting these changes

— Reply to this email directly, view it on GitHub https://github.com/DLCIncluded/Monotropism-Questionnaire/issues/7#issuecomment-1650696954, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABDKUPJLAZMRZ7K2Y6BHP3XSBIPPANCNFSM6AAAAAA2UGUV2I . You are receiving this because you were mentioned.Message ID: @.***>

ethereal-engineer commented 1 year ago

The revised calculation seems a more realistic projection of score. Great.