0
$\begingroup$

I'm new here :)

I'm conducting a survey questionnaire which gathers feedback on English-as Second Language (ESL).

I have two measurement scales in my survey. Questions on how much time they spend learning English in different ways, and statements to indicate whether they "Strongly Agree to Strongly Disagree". Both are 5-point Likert scale.

I'm wondering if it's appropriate to conduct an overall reliability analysis for all the items combined, or should I conduct them separately?

Improve this question
$\endgroup$

1 Answer 1

Reset to default
1
$\begingroup$

If by 'reliability analysis' you mean that you want to compute internal consistency (Cronbach's alpha) of the scale, combining items with different response formats could be problematic. The reason for this is that mixed response formats can lead to multidimensionality in the scale. In practice this can lead to biased internal consistency estimates (see e.g., Shu & Schwarz (2014) J Educ Measurement, 51:163-177). Thus, it may be a better to compute internal consistency for both sets of items separately (if the number of items does permit this).

Improve this answer
$\endgroup$

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.