If a questionnaire item with corrected item total correlation below 0.2 and it should be deleted from final version of the questionnaire, then how we can do concurrent validation?
1. Is it more accurate to do concurrent and convergent validation while removing results of deleted item? or
2. doing another pilot study after removing the item with low total correlation ?
Such a low value for an item indicates that it is dragging down the reliability of your scale, which will automatically decrease its ability to correlate with anything. So, you should remove the item before doing your validity assessment.
I agree with David L Morgan
By the way, if you think that the item is particularly important, you may retain the questionnair and then retest the reliability using a main study data. I had such an experience and defended well. Article Korean Patient-Perceived Satisfaction Scale of Community-Bas...
- Badges
- Science topic
- Similar topics
- Correction