Greetings everyone. I was wondering if any of you can help me solve this problem. If we have data from a survey and one of the items (question) of the questionnary is missing (i.e we don't have any response because it was not included by error), is there any method we can use to solve this problem? or we have to analyse the rest of the items. What if the item is part of a dimention? do we have to eliminate the rest of the items of that dimension?
Many thanks in advance
José
I feel your pain.
I worked on a pretty incompetently-designed study where, as well as burdening the participants with futile and bothersome measurements, they left a couple of items out of the depression scale that would have provided a valid way of assessing the intervention. I suggested that we could use the items we had and look for change in the participants, but the PI refused – he seemed to think that admitting a mistake made him look incompetent.
My advice: start by admitting the mistake. The have a look and see what the psychometrics on the remaining items are like. While you have to be up front about the fact that the scale differs to an unknown extent from the published version, you can make a case that it tells you something, and that differences between groups or trends over time are interpretable.
So be honest. Go with what you've got, and make sure you point out the limitations of the results.
Dear Dr. Conroy. Thank you very much for your answer. That's what I suggested the PI to do. To analyse the information provided by the rest of the items and not discussing the results of the dimension to which this item contributes.
Greetings,
- Badges
- Science topic