I think your intuitions are correct, and the reason is that Likert scales often suffer from poor construct validity.
One example that comes to mind is Alter et al. (2010), who use a mixed design, and note that participants in both conditions gravitate toward the middle of the scale on their first rating. They suggest that participants do not know how to use the scale at first; however, their second rating is informative because it can be compared to the first.
Several other studies (e.g. Dittrich et al., 2007; Olson et al. 2007) suggest that Likert scales are best used for relative rather than absolute judgments. Within-subject designs may abide by this rule because participants will implicitly use their previous ratings as anchors.
Although within-subject designs are often preferred, many times they are not feasible. (For instance, an experimental manipulation may make it impossible for a subject to be in both conditions without jeopardizing the effect of the manipulation). However, a properly designed scale may be usable in a between-subjects design. The key here is wording the question such that subjects will make judgements relative to some other criterion. So you may say something like:
"Compared to the last time you received one free beer, how happy were you upon receiving
ten free beers?"
-5 -4 -3 -2 -1 0 1 2 3 4 5
less happy than last time neither more or less happy much happier than last time
This example is a little contrived--in fact it's not that great since the question is leading--but hopefully you get the idea.
Since you mention happiness, I will note that this issue has arisen in the happiness literature. Happiness is often assessed by a Likert scale, but the wording of the question
can evoke different comparative standards. For instance, Steffel & Oppenehimer (2009) report that people respond differently whether they think about intra- vs inter- personal happiness (i.e. "How happy are you compared to others?" vs "How happy are you compared to how happy you would like to be?") When left to their own accord, they typically adopt an inter-personal comparison. However it is best to make this comparison explicit whenever possible to avoid these issues.
Alter, A. L., Oppenheimer, D. M., & Zemla, J. C. (2010). Missing the
trees for the forest: A construal level account of the illusion of
explanatory depth. Journal of personality and social psychology,
99(3), 436. PDF
Dittrich, R., Francis, B., Hatzinger, R., & Katzenbeisser, W. (2007).
A paired comparison approach for the analysis of sets of Likert-scale
responses. Statistical Modelling, 7(1), 3-28. PDF
Olson, J. M., Goffin, R. D., & Haynes, G. A. (2007). Relative versus
absolute measures of explicit attitudes: Implications for predicting
diverse attitude-relevant criteria. Journal of Personality and Social
Psychology; Journal of Personality and Social Psychology, 93(6), 907.
Steffel, M., & Oppenheimer, D. M. (2009). Happy by what standard? The
role of interpersonal and intrapersonal comparisons in ratings of
happiness. Social indicators research, 92(1), 69-79.