Restaurant patrons were asked to rate their service using either a semantic differential scale or an unbounded write-in scale. The service indices derived from semantic differential and unbounded write-in scales had different distributions, but did not differ in their relationship to tip percentages. The non-significant service by scale-type interaction suggests that the weak service-tipping relationship in the existing literature is not due to response biases associated with semantic differential scales.
Lynn, M. (2003). Restaurant tips and service quality: A weak relationship or just weak measurement? [Electronic version]. Retrieved [insert date], from Cornell University, School of Hospitality Administration site: http://scholarship.sha.cornell.edu/articles/132