I'm using a particular scale in my research (I did not develop this scale myself), and one of the subscales consists of just 3 items. One of those items is significantly affecting the internal reliability of the subscale in my sample - deleting this item would lift the subscale's alpha coefficient from .67 to .83.
Is it better to:
a) delete the item, to make the results from that subscale more interpretable, or
b) keep the item in, given that a subscale consisting of 2 items is hardly a subscale at all?
Thanks in advance!
You should delete it since that will improve the subscale internal reliability
Keeping subscale will be a better situation, as it is more realistic at the same time I guess not making the computation more complex. Sometimes in order to make the computations easier we neglect subscale especially when we feel it is not playing major role. Tracy Boateng is correct in saying deletion may improve the subscale internal reliability but in the process we may miss its contribution in reliability. Please be aware sometime we find a gap between guess and output. Decide judiciously.
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation