I am conducting a pilot study with a sample size of $n=30$. I am getting a Cronbach's alpha score of $0.933$ to $0.991$. I tried deleting a few items and re-running the analysis. But still, Cronbach's alpha has not changed much. My specific questions are:
Please share any relevant articles pertaining to this.
what could be the reason for this high value?
Generally speaking, values of $\alpha$ ranging from 0.93 to 0.99 indicate your scale has a large number of highly related and/or redundant items. More specifically, an $\hat\alpha$ of 0.93 is considered excellent by nearly all standards, though an $\hat\alpha$ of 0.99 may be seen as problematic, as such a high value is an indication of possible item redundancy or even a coding error. Finally, while $\alpha$ is known to perform well in small samples, a sample size of 30 is still small, so it is very possible that sampling variability contributed to your high $\alpha$ estimates.
Can I justify this high value and proceed to the main study?
You should be good, though the $\hat\alpha$ of 0.99 is worrisome. I would just go back and check for items that have identical responses on certain items.
Pls share any relevant articles pertaining to this.
Regarding next steps and resources, I would look to conduct an item analysis using classical test theory (CTT). See Bichi (2016) for an accessible introduction, as well as this CV post.
References
Bichi, A. A. (2016). Classical Test Theory: an introduction to linear modeling approach to test and item analysis. International Journal for Social Studies, 2(9), 27-33.
