Been writing our proposal for a qualitative study and needs a knowledge tool for measurement of parental knowledge on Intellectual Disability. But im having a problem with our tool for this analytic cross-sectional design. Planning to use Knowledge, Attitude, Practices -Intellectual Disability survey tool by Lakhan and Sharma. (https://pdfs.semanticscholar.org/0f4b/8adaf01a4a70030ed68148a317992f0a1439.pdf)
Please help
Just take the questions you want to use and do a new Cronbach's alpha. It should preferable be over .80. Then you write in the methods section that you used parts of the scale KAP and you tell what questions you used.
R.A Tan Chops This is important research work. Thank you for this focus. I have read Lakhan and Sharma and would also hope your questions will not only study the parents lack of knowledge, but also study the parents special first hand knowledge of the situation. You might already have seen this, but my experience, as a researcher and parent of a young person with intellectual disability, is that parents are often underestimated in surveys like this. This pool of knowledge can provide important input when developing constructive strategies for improving the lives of people with ID.
Thankyou so much s. Béatrice Marianne Ewalds-Kvist and Kjartan Skogly Kversøy ! Your points are well taken! thankyou for your help, I appreciate it! :)
In real life, knowledge is not something that exists apart from attitudes and practices. Our "knowledge" is very dependent on what we do, and on what we are disposed to do. For example, the well known saying that you cannot explain something to a person whose income depends on not understanding it. This is behind things like climate change, where the people who cannot apparently know or understand the research are those whose income would suffer.
In a qualitative study, you need to view knowledge not as a collection of facts but as part of a web of beliefs and practices in which the person is immersed. For this reason, I would recommend using an entire KAP scale.
On a pragmatic note, Cohen's alpha does not ensure that the items measure knowledge, just that they correlate with each other. In the case of KAP, this correlation will be strong even if you deliberately sample, say, 5 items from each domain. If you really want to confirm that the items measure a single underlying dimension, I suggest you look at Mokken scaling and use Jane Loevinger's H to measure scale fit. (It's actually called Loevinger's H, but I like to flag up that Loevinger was a woman statistician and psychologist!). You can do Mokken scaling in R or Stata by downloading the community-contributed package mokken (R) or msp (Stata)
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Writing