Bodvarsson and Gibson (1999) point out that customer ratings of service quality are
overwhelmingly positive. They argue that this restriction of variability in service ratings biases tests of
the tipping–service relationship and that the relationship would be stronger if the bias were eliminated.
There is some truth to this argument. An examination of data from four published studies reveals that
consumers’ ratings of service quality do fall disproportionately in the upper ranges of the scales used
(Table 1). Clearly, the correlation between tip sizes and service evaluations would be stronger if there
were more variability in service ratings. However, there are two problems with this criticism of existing
research.