Travel Reference
In-Depth Information
14
Advanced analytical methods
in tourism marketing research
Usage patterns and recommendations
Josef Mazanec, Amata Ring, Brigitte Stangl and Karin Teichmann
Usage and application patterns
When the American Marketing Association launched its
Journal of Marketing Research
in 1963, it
triggered off what is nowadays called the 'multivariate revolution' in marketing. It took about a
decade for marketing researchers to discover the intriguing application area of tourism that poses
plenty of challenges for advanced analytical methods.
Table 14.1
highlights the most prominent
analytical tasks together with conventional and nonstandard methods encountered in these fi elds
of marketing decision making. Nontrivial methods are instrumental for decision-support since
the analyst must come to terms with problems such as data reduction for purifying redundant
observations, qualitative variables, latent constructs and multiple indicators, ambiguity in the
direction of causality, unobserved heterogeneity, nonlinearity or multicollinearity. Given these
intricacies the need for advanced analytical methods is apparent and the investigation may
proceed with highlighting the actual method use in tourism marketing research.
The usage frequency of advanced analytical methods in tourism research was analyzed in a
survey article by Mazanec
et al
. (2010). The following text is based on this article (by courtesy of
Cognizant Communication Corp.). The survey covers six journals (
Annals of Tourism Research
,
Journal of Information Technology and Tourism
,
Journal of Travel Research
,
Journal of Travel and Tourism
Marketing
,
Tourism Analysis
and
Tourism Management
) during the time period 1988-2008. If
tourist behaviour research is considered a key element of tourism marketing one may safely
assume that the six-journal fi ndings are representative of empirical studies with a marketing
background. Under this assumption linear and nonlinear regression methods along with
Exploratory Factor Analysis (EFA) were heading the frequency list of popular methods (23 per
cent each; see also Lee and Law 2012). Confi rmatory Factor Analysis (CFA) was far less frequent
as a stand-alone technique (3 per cent), but necessarily features in a fast growing number of
Structural Equation Models (SEMs) with latent variables (7 per cent). Advanced Time Series
Models (5 per cent) were relatively rare compared to the regression-based econometric models.
Hierarchical and Partitioning Clustering methods put together attained an application frequency
of almost 10 per cent.
Multivariate Analysis of (Co-)Variance (MAN(C)OVA; 5 per cent) and Multiple Discriminant
Analysis (MDA; 3 per cent) were among the more popular techniques from the traditional
