Participants in this study comprise Danish- and Norwegianspeaking
leisure tourists who were passengers on selected international
direct scheduled (non-charter) flights of the airlines Air Berlin,
Cimber Air and Norwegian Air Shuttle returning from Mallorca to airports
in Denmark and Norway. The survey was accomplished on six
selected days, mixing different days of the week, from 10 July to 7 August
2010. Averagely, each survey day included three flights. The
study thus covers the peak summer holiday season for Scandinavians,
from the end of June until early August. A screening question was first
asked in order to identify passengers in the target category. Prospective
respondents were then requested to complete a self-instructing
questionnaire, which was returned to the administrative staff.
The questionnaire was available in Danish and Norwegian. It was
translated from an English original by the research team and its assistants,
which encompassed speakers of a range of languages. Due to fi-
nancial limitations, tourists from other Nordic countries were not
included in the survey. The questionnaire items were partly selected
on the basis of personal interviews with people of various ages,
most of them with experience from summer holidaymaking in the
European part of the Mediterranean. Also, three drafts of questionnaire
formulations were tested on a small number of people from
Denmark and Norway and discussed within the multilingual and
multi-national research team and with other travel survey experts.
The exact response rate among those who were asked to fill in the
questionnaire is not known, for several reasons. Firstly, numbers of
passengers on the selected flights and their distribution in terms of
travel parties were not known. Additionally, children were excluded
from the sample but numbers of children on the flights were not
known. The reply percentage was also influenced by the fact that couples
mostly filled in one questionnaire. Some 23 questionnaires were
rejected because they were incompletely filled in. Based on notes
from the survey staff, the response rate is roughly stipulated to be
just over 45%. This is considered satisfactory since airport surveys
typically have response rates between 40% and 70% (Rideng &
Christensen, 2004). The effective sample size is 405.
Non-charter tourists were chosen partly because they represent a
majority of arrivals in Mallorca and partly because they are assumingly
more independent and organise their own holiday journeys
to a greater degree than fellow holidaymakers going for charter package
tours. The choice of only direct flights was related to resources, as
it would have been prohibitively expensive to cover all departures
from Mallorca to relevant major airline hubs such as Barcelona, Amsterdam,
Frankfurt, Munich and Zurich. Also, it was known that
most Nordic holidaymakers in Mallorca during the peak summer season
have relatively short stays and thus favour direct flights. As the
study relies on self-reporting of a destination choice that for many respondents
was made well ahead of the interview occasion, there is a
potential recall bias or memory effect, that is, under-reporting and
J.K.S. Jacobsen, A.M. Munar / Tourism Management Perspectives 1 (2012) 39–47 41
over-reporting. However, it is likely that possible under-reporting
and over-reporting would be evenly distributed between various information
sources for destination choice. Then again, possible
under-reporting for reasons of social desirability might have occurred
(May, 2001). For instance, tourists with individualistic mind-sets may
possibly underestimate the importance of commercial sources of information
(e.g., airlines, tour operators/travel agencies and hotels),
as they might dislike looking upon their travel decisions as result of
commercial persuasion.
The chosen procedure, including a limited number of scheduled
flights on selected dates, represents sampling error (Hurst, 1994).
Sampling error is encountered in such en route surveys because the
sample is not a perfect representation of the test population. However,
by distributing data collection over a period of four weeks and
varying the days, sampling error is reduced.
Table 1 provides key characteristics of the sample. Some 41% of the
respondents were first time visitors to Mallorca, and 49% had stayed
more than one week in the island. Some 55% were women, reflecting
an actual female majority on such tours (e.g., Jacobsen, 2002). Half of
the tourists have a higher education, while quite a few were too
young to have started at university/college. Half of the sample had
bought only air ticket while the other half had purchased some kind
of tour package including air ticket and accommodation. Some 75%
stayed in hotels or similar, 17% resided in a holiday home, 4% stayed
with relatives or friends, while 4% made use of other types of overnight
lodgings. Moreover, t