Technique and theory in tourism analysis

It is argued that, in much tourism research, there has been an overconcentration on technique and not enough concentration on tourism itself. The current practice of demand modelling in tourism does not sufficiently take into account some of the particular characteristics of tourism. Important variables and significant dynamics are routinely ignored. The practices of tourism analysis in other areas, such as strategy and time series modelling, face a similar criticism of applying general techniques to tourism data without sufficiently adapting them to important characteristics of tourism. From the argument, the following propositions are put forward: (a) the results from demand models incorporating seasonality and seasonal data will prove to be better guides for policy and strategy than models of annual data; (b) the results from demand models, estimated using weighted least squares or percentage error loss functions, will yield more robust results than the usual ordinary least squares (OLS) models; (c) discrete choice models will come to be preferred as more reliable than aggregated demand models; and (d) structural equations modelling will decline in use as its weaknesses become apparent.