Article Abstract:
Howard, Lankford and Havitz committed an error in interpreting statistical data they obtained for their travel expenditure study. Using a matched samples t test, t values were determined given 39 degrees of freedom at the 0.10 probability level. Under these conditions, t values should exceed 1.685 to be considered statistically significant, a level none of the authors' t values attained. Their published two-tailed probabilities were incorrect. They were therefore wrong in rejecting the null-hypothesis and concluding that there was a significant difference between their subjects' estimates and actual recordings of travel expenditures.
User Contributions:
Comment about this article or add new information about this topic:
Article Abstract:
The role of constraints and attributes in the early evoked and late evoked stages preceding the final destination decision of travelers was studied. The convenience sample of respondents showed that the degree of attributes designated as facilitators are initially indicative of choice. However, this trend changed in latter stages, with the degree of perceived inhibitors significantly influencing decision. This implies that destination decision-making is a three stage process and that choice is a satisfying mechanism which is influenced by inhibitors rather than facilitators.
User Contributions:
Comment about this article or add new information about this topic:
Article Abstract:
Wesley S. Roehl's identification of the interpretation error committed on t test results in the pleasure travel expenditure study was welcomed by the authors. They explained that it stemmed from a transcription error were the wrong two-tailed probability value column was matched with t values. The mistake remained undetected despite internal controls set by the authors and their colleagues. While acknowledging their fault, the authors also defended their use of a two-group design and raised their own comments on Roehl's alternative hypotheses.
User Contributions:
Comment about this article or add new information about this topic: