Dear all,

I used a questionnaire for my research. For example, below are four questions I used in my questionnaire. Format: 7-points Likert items.

1. To what extent do you think the explanation is clear for children with ASD? (1 = extremely unclear, 2, 3, 4 = neither clear nor unclear, 5, 6, 7 = extremely clear).

2. To what extent do you think that the example is clear for children with ASD? (1 = extremely unclear, 2, 3, 4 = neither clear nor unclear, 5, 6, 7 = extremely clear).

3. What do you think of the length of the explanation, viewed from a child with ASD? (1 = extremely poor, 2, 3, 4 = Neutral, 5, 6, 7 = excellent).

Number four is more formulated as a statement than a question.

4. After this explanation, a child with ASD understand what to do. (1 = strongly disagree, 2, 3, 4 = neither agree or disagree, 5, 6, 7 = strongly agree).

Participants read the explanation for a task. This explanation is created for children with autism. After reading this explanation, they receive the questions stated above. Thus, the questions measure how good the explanation is.

My question is, is it possible to add those Likert items together (score them). So they become one Likert scale or is this not possible because they have different response anchors? I could not find any information regarding different response anchors. If number 4 cannot be added, can I still add the first three questions together despite the different response anchors?

The participants are divided into two groups. If I can combine the items into one scale and calculate the mean then I can use the mean to compare the two groups via a Mann-Whitney test.

If this is not possible then I will use the Likert items seperately.