I'm trying to develop a diagnosis tool and I expect that it would mix of objective & subjective questionnaires.
There's many ways to develop objective questionnaire in a rigorous way but I am not sure how I can develop the subjective questionnaires.
I wonder if there's any way to make the subjective questionnaires more rigorous.
Thanks!
hi Yoonyee. You could try two approaches. First, design 'pairs' of questions where you structure the second question as the opposite of the first (or the negative). For example: please indicate you level of agreement with the statement "I love apples". Then reverse it later in the survey please indicate you level of agreement with the statement "I hate apples". Then check that the two answers are consistent in that if it's a high score for the first question it should be a low score for the second. The second method is to add a monetary aspect. So if your first question is how much do you hate pollution then ask a second question about how much you would pay to reduce pollution. Hope this helps? Cheers, Marc
I think it is important to point out that what we understand as rigor is rather more volatile than it has been since the middle of the last century when modernism still held sway. I don't seek to discourage the process of instrument refinement, and Marc's suggestions above are sound, but with items intended to be "subjective," sometimes it is more useful to cultivate authentic dialogue than it is to provide validity checks. Otherwise you would be using something more "objective" in nature as a measure. I would assume meaning is more important for these items, and if so, a subject-centered technique, like photo-elicitation interview might be more fruitful. Peace and Love, Michael
I think the real problem is the validity of the indicators. For example, working with a attitudes scales (subjective indicators) implies validating the scale through factor analysis. The relationship between concepts and indicators should be tested before integrating a questionnaire. Ana
An alternative approach is to focus on the specific "source" that you use to generate the contents of the items. Typically, the use of "multiple sources", including "expert ratings" of the items [i.e., scope of relevancy to the target construct] will reduce the "garbage-in and garbage-out" factor analytic approach.
If you are familiar with IRT, you can attend to the LD-statistics to guide in selecting items with overlapping contents...Good luck with your project
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire