Spanish Domestic and Outbound Tourism Survey (Familitur)
Author: Eva Aranda Palmero
Person in charge of the “Spanish Domestic and Outbound Tourism Survey”.
1. INTRODUCTION
This presentation aims to give an account of Spain’s experience of research into the tourism behaviour of residents in Spain by using a panel-type statistical survey. Spain is one of the few countries that use this information-collection method for studying national tourism; we therefore believe it to be of interest to share our experience and describe the advantages of this system for obtaining useful, reliable and up-to-date information.
The statistical survey upon which this study is based is called Familitur (Movimientos turísticos de los Españoles: Tourism Movements of the Spanish), and it is drawn up by the Tourism Studies Institute (Instituto de Estudios Turísticos: IET)[1]. Its main objective is to collect information on all trips made by the Spanish (national tourism), both inside Spain (domestic tourism) and abroad (outbound tourism), providing they involve at least one overnight stay outside the usual environment.
A Short History of Familitur
The choice of a panel-type modality over other sample-design alternatives has followed upon a whole process of reflection being carried out by the Instituto de Estudios Turísticos since 1996, the first year of setting up the previous survey, at which time, albeit only experimentally, the choice had already been made to panelise a small fraction of the total of the selected census sections.
The results of the partial panelisation in the first year of the sample prompted IET to opt for a completely panelised sample in the two following years, 1997 and 1998. This modality was then continued in the new statistical survey Familitur begun in 1999, on the basis of a totally renewed household panel.
In 1999 a series of survey improvements was also made, with an increase of the sample (rising from 3,800 households in each data collection to 10,800) plus a series of modifications in the questionnaire. The main advantages of these changes were the obtaining of figures at a regional level and the study of a phenomenon not hitherto dealt with: short-duration trips to second residences.
2. METHODOLOGICAL REFERENCES OF FAMILITUR
Since 1996 the Instituto de Estudios Turísticos has been the body responsible for running the Familitur statistical survey, in fulfilment of the statistics commitments taken on by Spain in section C of Council Directive 95/57/EC on the collection of statistical information in the field of tourism.
A brief explanation will now be given of the methodological process of the Familitur statistical survey; any readers who desire more information on the matter should refer to the working document 23: “Movimientos Turísticos de los Españoles (Familitur). Referencias metodológicas (methodological references)”, available in the Instituto de Estudios Turísticos and on the organisation’s web page:
2.1. Sample Type and Units
To carry out the survey a selection is made of a panel of households by using a two-stage cluster sample with subsample and stratification of the first-stage units. The first-stage units are the census sections of the Instituto Nacional de Estadística (National Statistical Office). The second-stage units are the main family dwellings. Thus the working sample universe excludes collective establishments, the latter concept embracing all hotel-type establishments.
The sample consists of 1,200 sections distributed between the 17 Autonomous Communities making up the whole national territory, in accordance with criteria other than those of the strict proportionality of their respective populations, with a non-proportional allocation. Thus, the distribution chosen for the sample stratification is representative in the smaller communities, population wise, increasing in all of them the number of sections that would otherwise correspond to them in terms of strict proportionality. Nine households are chosen randomly in each of the 1,200 census sections, the final sample resulting in a panel of 10,800 households.
2.2. Reference Period: Four-monthly versus Three-monthly
The survey is effected in three data collections throughout the year, each involving questions on the trips made during the four months running up to the interview month.
In 1997 and 1998 the data was collected on a three-monthly basis, dividing the sample each time into 3 groups. This made it difficult to obtain information during interviews scheduled for holiday months such as July or August, when the families were on holiday.
Another problem posed by this three-monthly method was the splitting up of holiday periods. The inclusion of different data collections in the same holiday period (such as the summer season) meant that a complete analysis of holiday seasons had to wait for the following data collection to be carried out.
These factors led to a rethink of the reference periods in the methodological change of 1999, the conclusion being drawn that the best data-collection period would be four months. The following three reference periods were thus established: the first to record trips made from February to May in a June interview. This meant that this reference period would always include trips made in Easter, whether this fell in March or April, a crucial factor for making year by year comparisons. The second period covers the four months of the summer season (June to September) in an October interview, when most people have returned from their summer holidays. Lastly the October to January period records all Christmas trips, the interview to be conducted in February.
This data-collection method is not incompatible with the three-monthly information requirements of the Council Directive, since, although the survey’s reference period is four monthly, three-monthly information can still be furnished.
IET would like to take this chance to suggest the need to harmonise somewhat the consideration of the various holiday seasons in the various tourism statistical surveys produced, both by the different bodies and by the distinct countries, thus ensuring that the figures would always be comparable.
2.3. Collection of the Information
As for the information-collection system, it is a mixed system combining personal interviews in the homes of the panelists and telephone interviews. During the first panel-capturing operation the information is collected in a personal interview in the home of the members of each household under study, selecting one household member as the main informant. In subsequent data collections, once the household panel has been set up, the method is characterised by the flexible use of two systems combined in the most efficient manner: telephone collection of the information (C.A.T.I.) and personal home interviews when the interviewee prefers or when the household cannot be contacted by telephone.
Telephone calls to households for carrying out the telephone interview are made on different days of the week and at different times of the day with the paramount objective of locating the main informant.
2.4. Validation and Imputation
When all the information is available from the 10,800 questionnaires of each campaign, the first step is then to subject all this information to a series of logical controls designed to ensure its quality before processing it. All data is thus subjected to consistency programmes to correct all wrong information, following validation rules that may be consulted in the various methodological documents published by IET.
Once the information has been debugged, the next step is to carry out lack-of-response imputation tasks, insofar as this has not been corrected in the validation phase.
2.5. Elevation
In broad terms the Familitur survey could be considered as three surveys in one, for there are three analysis units with a totally differentiated treatment. First comes the analysis of the household, involving a study of its characteristics and access to or possession of certain durable consumer goods, such as a second residence or tourism equipment. Second comes the investigation of the household members, including information on their socio-economic characteristics. Lastly comes the study of the trips made by these household members, in which an analysis is made of their basic characteristics such as the type of accommodation used, the means of transport, how the trip was organised, etc. Each of these analysis units is directly related to an area of the questionnaire. It should be stressed that, although they could be considered as three different operations, the interrelationship between each one and the cross checking of data between them is of paramount importance.
As already mentioned, they could be considered as three different operations, insofar as the elevation framework of each one is different. Thus, household data is elevated to the total of Spanish households; the elevation framework of the sample data of household members is the total number of people living in Spain, while the trips made are elevated in terms of the calculated weights of each one, based on the weight of each person who made the trip. As might be expected the balancing matrices used for households and household members are also different. Furthermore the processing of the data obtained for each of the analysis units is totally different.
2.6. Obtaining the Results
In the tabulation plan the information is structured into three basic blocks, to tie in with the aforementioned analysis units: households, household members and trips. Thus, the first block includes a study of the households, with details, for example, on whether or not the household is a traveller one, depending on whether or not it has access to a second residence. The second block concentrates on the analysis of the household members, offering information of the type: travellers or non travellers, by age group, level of education, etc… Lastly, the third block, studying the trips made, concentrates on the study of the characteristics thereof, such as type of accommodation, means of transport, how the trip was organised, etc..
The concept “traveller” to be explained herein, is linked to the household members and is obtained in this information block.
Although the tabulation plan has not included it this year, information can also be offered on trips correlated with the information on the household members, since the socio-economic characteristics of each traveller are known. A study can thus be made of the academic qualifications of persons using the hotel, or the occupation of persons making business trips, to give only a few examples.
2.7. Analysis Tool: Pulsar
It has already been mentioned that certain variables can be correlated with others even though this is not included in the tabulation plan. On certain occasions it is also necessary to effect special operations for comparing certain results with those offered by other sources. It should be pointed out that this type of cross check and operation can be effected simply and immediately, since all survey information is loaded into a user-friendly programme (Pulsar), which carries out all these investigations in a few minutes and without any need for handling complex computing systems.
2.8. Dissemination of Results
IET publishes annually all tables of the tabulation plan, both in hard copy and in computer format. It also draws up reports for each season: spring, winter and summer. All this information is available on the Institute’s web page.
There is also a series of agreements with regional statistics institutes for furnishing the primary files, so that the latter may use them for ad hoc purposes.
3. SPANISH TOURISM BEHAVIOUR: THE IDEAL NATURE OF A PANEL
We are first of all going to sketch in some of the characteristics of the tourism behaviour of Spanish residents. It was precisely these features that prompted I.E.T. to opt for a panel-type statistical survey to study this matter.
From the point of view of inbound tourism Spain ranks second worldwide in the number of tourist arrivals. From the point of view of national tourism, however, an analysis of the tourism activity of Spanish residents shows that half of the total Spanish population make trips.
One of the characteristics of the trips made by the Spanish residents is that they are principally (82%) domestic trips. Another of the particular features of tourism in Spain, in comparison with tourism elsewhere, is the phenomenon of the second residence. The Household Panel of the European Union allows for a study of access to such properties in the various countries of the EU, showing the relative importance of such residences in Spain. In 2000 trips to a second residence accounted for over 63% of the total of trips made by the Spanish residents.
The explanation for this phenomenon perhaps lies in the fact that internal migration within Spain has led to a situation where 50% of the Spanish population live in a different place from where they were born. Prima facie this could well lead to a great number of trips to visit family or friends. The figures for 2000 seem to bare out this hypothesis, since 21% of the total trips made by the Spanish were due to this purpose.
As can be seen Spanish tourism behaviour differs from the typical picture in other countries, standing out for the low percentage of the traveller population, the use of the second residence and the low rate of outbound trips.
These factors motivated a thoroughgoing research into both the traveller and non-traveller population, as well as an investigation of the socio-economic characteristics lying behind the tourism behaviour of the Spanish residents.
4. MAIN CONTRIBUTION OF FAMILITUR: THE STUDY OF THE TRAVELLER
4.1. Definition and Concept of Traveller
A Commission Decision[2] defined traveller as “any person on a trip between two or more countries or between two or more places within his/her country of residence”. The interpretation of this definition by the Instituto de Estudios Turísticos understands the concept “traveller” to be associated with a person, so that an individual is to be considered a traveller when he/she has made at least one trip, regardless of whether more than one was made. According to this interpretation the analysis unit “traveller” requires a special treatment, very different from the “trip” unit. Indeed, a trip is not associated with a person but rather with a displacement. Delving a little deeper into both definitions, we can say that displacements can straightforwardly be aggregated, whereas “travellers” need to aggregated in a special way to ensure the same individual is not counted as a “traveller” twice. The calculation of the number of “travellers” therefore requires a longitudinal analysis of each individual.
To help clear up both concepts we are going to give a brief example: If an individual makes two trips during January and three in February, he/she will be counted as one “traveller” in January, one “traveller” in February and as one “traveller” in the cumulative period January-February. For the purposes of counting trips, however, this person will have made two trips in January, three in February and a total of five in the aggregate period January-February.
Surveys conducted at border crossings or destination points use the “trip” as analysis unit. The main advantage of these surveys is that the informants are always the people actually making the trip, so they can offer much more information on the associated characteristics of the trips. As is logical, however, these surveys are not addressed to non travellers, so no information can be obtained on this group nor can the results be brought into relation with them.
It can safely be claimed that the best way of obtaining the “traveller” variable is by a panel-type survey. In surveys with independent samples it is difficult to monitor the same individual over time, so it is impossible to find out whether or not the same person is being counted as a “traveller” twice. If, for example, the same German citizen should visit Spain on repeated occasions during the same reference period, all the trips made to Spain with an overnight stay will be counted as “trips”. Furthermore, in surveys carried out, for example, in hotels, the same person staying in different hotels will be counted under different trips, since it is impossible to follow each individual over time.
Calculation of the “traveller” variable, therefore, allows us to offer results relative to the whole population; for example, that 50.2% of the Spanish population made some sort of trip in 2000. There might well be other means of obtaining this information besides the panel method, such as taking an independent sample at a given moment of time and asking individuals whether they have made a trip during this year. But this method would suffer greatly from the memory effect, and it would also be impossible to establish a trend of the proportion of travellers in different periods of time. For all these reasons, it is thought that the best method of obtaining this variable is the use of household panels.
It should be stressed here that the Familitur survey complies with the information requirements of the Council Directive since, as well as investigating the “traveller” variable it also offers the results of “trips”, which is the analysis unit required by the Directive. The social reality of Spain and the particular tourism behaviour of its inhabitants prompted IET to opt for a panel type study that could offer not only trip information but also a sociological study over time of the trip behaviour of the inhabitants of Spain.