The platform calculates overall Scores of specific questionnaires wrongly.

Question

I entered several different questionnaires for a project. The scores of the questionnaires need to be calculated differently for each of them.
So far, when I download the CVS-files of collected answers, these scores are wrongly calculated. For instance, a questionnaire has 4 possible answers for each questions (0-3 points). However, the platform always gives at least 1 point (for the answer that should give 0 points). How can I change this?

Answer 1

The internal coding for selection and scale questions always starts with 1 (for several reasons, one is that Excel does not cleanly distinguish between 0 and nothing).

You may now either use a "recode variable" element to actually recode the values into new variables.

Or you may simply substract from the mean ($score = valueMean(...) - 1) or summary indices ($score = valueSum(...) - $n_items) to calculate the correct scores.

Comments

Can I change this somehow so that in the CSV it is directly calculated correctly? Will this be done by using a "recode variable" element? If so, can you give me a concrete instruction on how to do this?
Or is a PHP needed for that? If yes, can you give me one?

---

You can use both, either a "recode variable" element or PHP. My recommendation would be the first option, because it handles the additional variable and the recording in one step.

1. Create an element "recode responses" in the list of questions
2. Tell it to recode 1 -> 0, 2 -> 1, etc.
3. Under "import items directly" add the current variable names and new names (note, that they must use a different scheme then the A000_* used by SoSci data variables)

GC03_01 -> V1_01
GC03_02 -> V1_02
GC03_03 -> V1_03

Save the element and drag it onto any page of your questionnaire.

However, let me repeat, that there is no "right" or "wrong" coding. The information contained in the data is the same. It depends on the context whether linear transformations make sense and where.