I am doing my undergraduate thesis, and I wonder if I need to do the Cronbach's alpha from my data set, or if I should use the Cronbach's alpha from the commercial questionnaire I will use for this paper. What is the standard way to approach this?

  • $\begingroup$ What do you mean data set? How is the commercial questionnaire not a data set? Explain, please. It is unclear what you are asking. $\endgroup$
    – Carl
    Commented May 13, 2017 at 1:33

1 Answer 1


The standard practice—at least in psychology—is to calculate and report a Cronbach's alpha for any scale that you create.

Say that you have a commercial questionnaire that is four items long: Item 1, Item 2, Item 3, and Item 4. The standard approach is to simply average the participant's responses across those items. When we do that, we often report a Cronbach's alpha as a quick-and-dirty measure of reliability.

Here is an example from a paper I co-authored:

We employed the same free speech items as in Studies 4 through 6 ($\alpha = $ .97), and used the symbolic racism scale ($\alpha = $.89; Henry & Sears, 2002) to assess prejudice. Our measure of expressive threat was slightly adapted from the autonomy subscale of the basic psychological needs scale (Johnston & Finney, 2010): “I am free to decide for myself how to live my life,” “I feel pressured to think what others want me to think,” “I feel free to express my ideas and opinions,” “I feel like I can pretty much be myself,” “I feel free to be who I am,” “I can voice my opinion,” and “I feel controlled and pressured to have certain beliefs” ($\alpha = $ .92). Items were scored such that higher scores indicated more threat to expression.

They are easy to calculate in common statistical software you are likely using—like R, SPSS, or SAS. Make sure that if any of your items need to be reverse-scored you do so before calculating Cronbach's alpha.

  • $\begingroup$ Thanks Mark! So even if I use a scale from a comrercial test, it is recommended to calxulate the alpha based on your own dataset. Is that what you are saying? $\endgroup$ Commented May 13, 2017 at 14:01
  • $\begingroup$ Yes, you should run it on your own dataset. $\endgroup$
    – Mark White
    Commented May 13, 2017 at 14:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.