I seem to recall a recent paper investigating whether ratings created using Likert or Visual Analogue Scales formats had the cognitive properties of interval or ratio data (since they are generally analysed as such).
I think a central question was whether, for example, participants understood the difference between a 2 and a 3 to be the same as the difference between 5 and 6, but I can't find it! Does this ring bells for anybody?
I did eventually find something which may be what I was thinking of, examining the normality & psychometric properties of Likert scales with differing numbers of points. They concluded that the 11-point scale was closest to interval-type data in scaling and normality.
Reference:
A comparison of psychometric properties and normality in 4-, 5-, 6-, and 11-point Likert scales. Leung, Shing-On. Journal of Social Service Research37.4 (Jul 2011): 412-421. http://dx.doi.org.ezproxy.library.tufts.edu/10.1080/01488376.2011.580697
