Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 1 down vote favorite

I recently re-ran a survey for year to year comparisons. In the recent survey I changed the Likert scale slightly by anchoring each box rather than just # 1 and # 6.

In other words, my scale went from this:

1 - agree completely, 2 , 3 , 4 , 5 , 6 - disagree completely

To this:

1 - agree completely 2 - agree 3 - somewhat agree 4 - somewhat disagree 5 - disagree 6 - disagree completely

Will the year to year comparisons still be accurate? I have looked online but have not found any standard best practices in this area. Thanks.

share|improve this question
1  
Your use of 'anchor item' does not correspond to what it ought to mean in psychometrics. I would suggest to update your title to reflect that this question is about wording of response categories or item format that changed over time. – chl♦ Jan 25 '12 at 22:33

1 Answer

up vote 3 down vote

I think there are some studies claiming that it makes a difference regarding the distribution of responses. The main concern is to justify the interval scale. Likert scales anchor every point and try to encourage equal interval sizes. Strictly speaking, a Likert scale always requires that all points are anchored.

However, I doubt it makes a difference. Whatever anchoring you use, data will always be subject to scale usage heterogeneity (http://www.jstor.org/stable/2670337). This effect will be much stronger than any effect caused by anchoring.

share|improve this answer
The other issue will be compositional effects due to sampling changes from one year to the other. I imagine scale usage heterogeneity and sample compositional effects will swamp any other factors. – Michelle Jan 25 '12 at 22:25
What is your definition of anchoring? – chl♦ Jan 25 '12 at 22:34
ok, my answer is a bit sloppy mixing anchoring and labeling. In the context of likert scales anchoring refers to labeling in a way, such that there is some kind of absolute reference point which calibrates the scale. Again, I don't think that we can ever be certain that it works. – joint_p Jan 26 '12 at 16:39

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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