I work in Social Science Research (Management)

In my school, Likert Scale is extensively used to measure attitudes.(Strongly disagree to Strongly agree). So, I kind of assumed that it's an absolute standard.

But recently, while I was reading material on designing surveys, I came across the argument that highly criticized this format of response categories, owing to a tendency called "acquiescence response set".

I'm working on my questionnaire and have drafted several questions that measure the attitude of respondents related to an information system. Ex: The system provides accurate information. (strongly agree, Agree, Neither, Disagree, Strongly disagree)

My question is - what are the alternative scales I can use to measure the attitudes of the respondents? what is the more reliable way to measure attitudes?

  • $\begingroup$ A link to info about the acquiescence response set would be helpful. $\endgroup$
    – Krysta
    Oct 29 '14 at 14:38

In attitude psychology Likert response scales are still the norm. They do have their problems (and there are a good deal of problems), but many of them can be addressed satisfactorily.

The acquiescence bias you mention is a small issue. It can be balanced out by reverse coding roughly half of your test items. The acquiescence bias then becomes random noise rather than a systematic bias, which is annoying but not very damaging as long as your analysis has good power.

For example, if you have a 4-item scale that measures roughly the same thing, you could mix items like "The system provides accurate information" with items like "I would not recommend the system to others".

Does that make sense?

  • $\begingroup$ Welcome to cogsci.SE! This has the makings of a good answer, but would be improved by citing sources for more information on the problems with Likert-type scales, or on techniques like reverse coding. $\endgroup$
    – Krysta
    Oct 29 '14 at 19:03
  • $\begingroup$ @adb You mean to say that I can frame roughly half of my questions negatively, and half positively. That makes sense. Thank you! $\endgroup$
    – siri
    Oct 30 '14 at 1:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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