Search results

Data-driven market segmentation is heavily used by academic tourism and hospitality researchers to create knowledge and by data analysts in tourism industry to generate market insights. The stability of market segmentation solutions across repeated calculations is a key quality indicator of a segmentation solution. Yet, stability is typically ignored, risking that the segmentation solution arrived at is random. This study aims to offer an overview of market segmentation analysis and propose a new procedure to increase the stability of market segmentation solutions derived from binary data.

The authors propose a new method – based on two independently proposed algorithms – to increase the stability of market segmentation solutions. They demonstrate the superior performance of the new method using empirical data.

The proposed approach uses k-means as base algorithm and combines the variable selection method proposed by Brusco (2004) with the global stability analysis introduced by Dolnicar and Leisch (2010). This new approach increases the stability of segmentation solutions by simultaneously selecting variables and numbers of segments.

The new approach can be adopted immediately by academic researchers and industry data analysts alike to improve the quality of market segmentation solutions derived from empirical tourist data. Higher quality market segmentation solutions translate into competitive advantage and increased business or destination performance.

The proposed approach is newly developed in this study. It helps industry data analysts and academic researchers to reduce the risk of deriving random segmentation solutions by analyzing the data in a systematic way, then selecting the most stable solution using the segmentation variables contributing to this most stable solution only.

The purpose of this paper is to critically review past recommendations to correct for cultural biases in empirical survey data sets, and propose a framework that enables the researcher to assess the robustness of empirical findings from culture‐specific response styles (CSRS).

The paper proposes to analyze a set of derived data sets, including the original data as well as data corrected for response styles using theoretically plausible correction methods for the empirical data at hand. The level of agreement of results across correction methods indicates the robustness of findings to possible contamination of data by cross‐cultural response styles.

The proposed method can be used to inform researchers and data analysts about the extent to which the validity of their conclusions is threatened by data contamination and provides guidance regarding the results that can safely be reported.

Response styles can distort survey findings. CSRS are particularly problematic for researchers using multicultural samples because the resulting data contamination can lead to inaccurate conclusions about the research question under study.

The proposed approach avoids the disadvantages of ignoring the problem and interpreting spurious results or choosing one single correction technique that potentially introduces new kinds of data contamination.

The existence of variable response styles represents a major threat to the correct interpretation of market research findings. In international marketing, this threat is further increased due to samples of respondents from different cultural backgrounds. The purpose of this paper is to extend the investigation of differences in cross‐cultural response styles by studying full response patterns instead of extreme values, quantify the extent of the potential mistake of not accounting for cross‐cultural differences in response behaviour and present a simple way of testing whether or not data sets from various cultural backgrounds can be used without correcting for cross‐cultural response styles.

Two independent data sets are used. Extreme response style (ERS) scores are compared by testing for equality of proportions. Respondents' answer patterns are partitioned using the k‐means algorithm, the resulting differences between cultures tested using a Fisher's exact test for count data. The extent of inter‐cultural difference in responses is assessed using ANOVA.

Asian and Australian respondents differ significantly in ERS and full response patterns. Differences in cross‐cultural response patterns account for up to 6 per cent of the variance in the data, thus representing a significant potential source for misinterpretation in cross‐cultural studies.

International market researchers using samples including respondents from more than one cultural background have to be aware of the potential source of misinterpretation caused by cross‐cultural differences in response patterns. A simple ANOVA‐based procedure allows researchers to determine whether data can be used in its uncorrected form.

The paper investigates cross‐cultural response styles for new groups of respondents (Australian vs Asian), extends the study from the investigation of extreme values to full response patterns and gives market researchers in the international marketing context an indication of how high the level of potential misinterpretation can be and presents a simple means of checking how necessary it is to account for cross‐cultural differences in response behaviour.

This paper reflects on the state of quantitative tourism research.

Prior literature, observation and introspection form the basis of this article.

Key questions raised include: Do we choose methods because they are suitable or because they impress? Are our results just another number? Why temperature is temperature and loyalty not loyalty? Why are we retesting the same things over and over again? Do we have enough suitable reviewers? Why don’t we study what we are trying to understand: tourist behaviour?

This paper is limited, in that it does not conduct a comprehensive review to provide information on how frequently the observed phenomena are occurring in tourism research.

This paper calls for a change in quantitative tourism research. Specifically it calls for refocusing on the study of actual behaviour, tackling novel research problems or – when existing constructs are studied –building on existing definitions, applying the simplest possible appropriate methodological approach rigorously, dedicating manuscript space to a detailed interpretation and discussion of findings and being open-minded and generous as reviewers while not compromising on methodological rigour.

To the best of the author’s knowledge, this special issue of Tourism Review – including the present article – represents the first attempt to critically reflect on the state of quantitative tourism research.