+ Reply to Thread
Results 1 to 1 of 1

Thread: Redesigning a questionnaire (a bit long...)

  1. #1
    Points: 2,942, Level: 33
    Level completed: 28%, Points required for next Level: 108

    Posts
    8
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Redesigning a questionnaire (a bit long...)




    A quick background:

    I am trying to evaluate a questionnaire used to asses patients. It is used around the world to establish whether a patients suffers from a certain condition, or not. It is filled in by an assessor (typically a nurse) based on the patient's performance.

    It consists of 22 questions/parts:
    - Some are simple: did the patient passed/answered correctly or not. The score given for these is 0=fail, 1=pass.
    - In others the patients needs, for example, to repeat a list or 5 items read to him/her by the assessor. The score here is based on the number of items the patients manages to repeat correctly (0-5).

    The overall score is the sum of all the scores for each question/part. If a patients scores below a certain threshold it is deemed that the patient suffers from the condition.

    Looking at a small number (~100) of completed questionnaires, it appears that some of the questions are either not well-phrased, or, perhaps, completely unnecessary, as all (or nearly all) of the patients score the same for them.

    What I, and my clinical advisor, would ideally like to achieve is to significantly reduce the number of questions used without, of course, significantly affect the overall outcome of the questionnaire. She believes, based on her experience, that something is the region of 5-10 questions should be sufficient.

    My plan of attack is to use brute force approach: systematically choose any combination of questions (ranging from, say, 4 to 15), and use multi-variable regression for each combinations..

    Once the results of the above are ready I can then try and identify the “sweet spot” with regards to the number of questions required, and which ones they are.

    PS: I will have access to 1000’s of completed questionnaires for the purpose of this exercise.

    My questions:

    1. From a statistical point of view, will the YES/NO and 0-5 scoring system mention above work with his approach?
    2. Any suggestion as to the tool to use? I was thinking about using Excel and write a macro for the iterations.

    Many thanks in advance.

    J.
    Last edited by j_cocker; 10-19-2015 at 08:31 AM.

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats