Geography Reference
In-Depth Information
METHODOLOGICAL ISSUES
Latham and Edwards (2003) review the major types of data collection used for tourism
statistics. They report that among state-sponsored tourism research in the United States,
conversion studies are a popular method to examine and evaluate advertising campaigns
and visitor surveys, to assess a sample of visitors to individual states. The use of other
methods of data collection are also discussed (e.g. diary questionnaires, participant
observation and personal interviews—see Mullins and Heywood 1984; Perdue 1985). Yet
few studies consider the issue of sampling, sample design and the sources of error which
may arise from such surveys (Aaker and Day 1986; Cannon 1987). In fact the lack of
research on the reliability of the estimate from a sample survey (the standard error) is
rarely discussed in most tourism surveys (for a more technical discussion of this point,
see Latham and Edwards 2003:70-1)). In many cases, large tourism surveys focus on the
logistics of drawing the sample and the bias which may be reflected in the results.
Therefore, any tourism survey will need to pay careful attention to the statistical and
mathematical accuracy of the survey, especially the survey design and the effect it may
have on the results, a feature which is discussed in great detail by Ryan (1995).
Ryan (1995) provides an excellent review of survey design, questionnaire design,
sampling and also an insight into the statistical techniques to use for different forms of
tourism data. As a result it serves as an important reference point for issues of
methodology and the technical issues associated with the statistical analysis of tourism
data. Without reiterating the excellent features of Ryan's findings, it is appropriate to
consider some of the main accuracy problems associated with the collection of domestic
and international tourism statistics.
Ryan (1995) argues that errors in data collection can lead to errors in data analysis.
Among the most frequently cited problems associated with domestic and international
tourism statistics are
• the methods by which the data are collected, which are influenced by administrative,
bureaucratic and legislative factors in each country
• sample sizes which are too small and lead to unacceptable sampling errors and in some
instances where the sample design is flawed
• the procedures for collecting tourism statistics are not adhered to by the agency
collecting the data.
In addition, Edwards (1991:68) argues that a 'fourth potential reason—arithmetic
mistakes and data processing errors—only occasionally produce significant errors'. In
fact, Edwards (1991:68) supports the cause of 'tourist statisticians [who] are both
knowledgeable and conscientious, but are having to work with tools which they know
could produce inaccurate or misleading data', concluding that for any set of tourist data,
potential sources of error obviously depend on the method of collection employed. This,
in turn, tends to be largely determined by the legislative and administrative framework
and by the financial and manpower resources available.
In the case of tourist expenditure and receipts data, organisations such as the
International Monetary Fund (IMF) issue guidelines for the compilation of balance of
payments statistics. But errors may occur where leakage results from tourist services paid
for in overseas bank accounts and in extreme cases, where a black market exists in
