ABSTRACT
Whereas the drivers of tourist satisfaction with a destination have been largely studied, we know little about whether the serial order of the destination in multi-destination trips plays a role in explaining satisfaction. Based on a large longitudinal data set and using ordered probit models, we show that tourists are systematically highly satisfied with those destinations visited later within a multi-destination trip. This finding is conditional on a large set of controls and individual random effects. Interestingly, length of stay at each destination, age and travel party size are found to be significant mediators of the relationship between the serial order of the trip and satisfaction, partially counterbalancing recall bias. The results provide evidence of recency effects in tourist satisfaction recall and have important implications for the interpretation of tourist satisfaction studies.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1 There is a discussion in the econometrics literature about the need for exclusion restriction. While authors like Wilde (Citation2000) indicate the model can be identified due to its non-linearity, others like Puhani (Citation2000) advocate for the need of at least an exclusion restriction.
2 We use the Autonomous Community of origin rather than the province because the Domestic Travel Survey only provides information about the Autonomous Community of origin.
3 Alternatively, Autonomous Community of origin fixed effects could be used as exclusion restrictions. Apart from their potential overlapping with the province of destination fixed effects for those who travel within the borders of the region of residence, their lack of temporal variation makes them less suitable as predictors of the heterogeneity in the probability of engaging into a multi-destination trip over time and across seasons.
4 Heckman two-step regressions assuming Satisfaction is continuous rather than ordinal produce similar results, since the inverse Mills ratio is not found to be significant (Supplementary Material, Table A3). As discussed in Greene (Citation2000, p. 843), selection bias can be understood as a problem of omitted variables. If the coefficient associated with the inverse Mills ratio in the second stage is not statistically different from zero, this implies a regression on the subsample of interest produces unbiased and consistent estimates of the parameters.
5 In linear regression models, the indirect effect of a covariate can be easily obtained as the difference in the coefficients of Destination Order from regressions without (full effect) and with (direct effect) the control
. However, in non-linear regression models like ordered probit this becomes more cumbersome because of the differences in the residual variances across models (Kohler et al., Citation2011). The KHB method distinguishing between mediation and scale effects. See Kohler et al. (Citation2011) and Karlsson et al. (2012) for further details.