How can the researcher (in management) decide whether to
- to use an existing scale
- to modify the existing scale
- to develop a new scale?
What are the advantages / disadvantages/ statistical tests/ implication on validity and reliability for each option?
Thanks and best regards.
Mohan Karambelkar. I think you ask some very good questions. On the one hand, existing scales can be just taken off the shelf and used as they are. However, they are often not really suitable for particular research purposes, so researchers can end up barking up the wrong tree. In their "original" form, many existing scales have problems with their instructions, items, and response options, also, and under those circumstances I think it's reasonable to consider modifying them. Why persist with something mediocre? An unrelated issue is that some scales have copyright over them and researchers might need to pay a fee (even per participant) for using them. Even modifying those scales could infringe copyright, so care might need to be exercised regarding that.
One advantage of existing scales is that scores on them can probably be compared between different samples.
Modifying scales is very sensible, at times - particularly in order to overcome the problems I mentioned above such as unsatisfactory instructions, items, and response options. Many apparently respectable and widely used scales have problems of that kind, and I don't blame researchers for trying to improve those scales in the hope of collecting more valid data.
The third option you consider, creating new scales, can require a lot of work. I know. In fact, a colleague and I are doing exactly that at the moment, and simply getting enough participants (two groups of 400 in order to be able to run factor analyses appropriately) is a challenge. Having sufficient numbers of participants is only one issue, however. Determining how some forms of validity might be assessed is another issue: convergent and discriminant validity, and known groups validity, for example are not as simple as might initially seem to be the case. It is also necessary to have statistical skills in such domains as factor analysis.
There are some good books and journal articles about scale development. I could suggest some if you like, but for the moment you might be interested in an article that I posted on RG recently concerning scale development. I had written that article about 10 years ago because I was concerned about self-assessment of foot health, but the principles in my article generalize to other areas of research.
I'll download it here for you.
I'm sure that other RG members could provide advice additional (some perhaps contradictory) to what I have written above. May they do so!
Systematic lit review, ideally synthesis narrative review.
Hope this helps
Matt
Thank you Dr. Robert Trevethan For the detailed answer and very useful inputs.
Thank you Dr. Matthew Kerry for the suggestion.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publications