See Regulation 692/2011, Delegated Regulation 2019/1681 and Methodological Manual for Tourism Statistics

Information on non-mandatory variables are available under request (NSI).

The statistical unit of the survey is the resident of Luxembourg aged from 15 to 85 years and not living in collective household.

Participation in tourism: the individual.
Tourism trips: the tourism trip with at least one overnight stay made by the individual.
Same-day visits: the same-day visit made by the individual.

This is an individual survey, and not a household survey.

Participation in tourism: Residents aged from 15 to 85 years (collective households are excluded).

Trips: All leisure and business trips with at least one overnight stay made outside the usual environment which ended during the reference period.

Same-day visits: visits without overnight stays, made outside the usual environment, originated from the place of usual residence. It includes leisure and business same-day visits.

The first data were delivered in 2012 to Eurostat, but the information has been collected since 1997.

The first data were delivered in 1997 to Eurostat.

The first data were delivered in 2014 to Eurostat.

The first data delivered to Eurostat were collected in 2018, but the data exist since 2017.

Expenses in overnight trips: break in series in 2019. In 2019, data collection on expenses for leisure trips with overnights has been modified from a collective cost to an individual cost. Until 2018 included, for example: cost of a one week leisure trip made by a couple was collected for the 2 persons together (cost for transport, for acommodation, for "food, drinks and other"). The NSI used to divide the costs by coefficients taking into consideration the size of the group, the age of the members and the means of transport. From 2019 onwards, expenses are collected for the respondent only. When costs are shared (transport by car for exemple), the individual cost is calculated by the respondent. Between the two years, the average cost of a leisure trip decreased about 20% (from 991€/person to 787€/person).

Regulation 692/2011

Regulation 1051/2011

Delegated Regulation 2019/1681

The Law of July 10, 2011 on the organisation of the National Institute for Statistics and Economic Studies sets out the missions of STATEC and its internal organisation (See this website).

Formal contract of STATEC with subcontractor TIP Markforschung, Trier.

The survey on Tourism is conducted on a mixed mode data collection process, telephone and web.

The sample is divided into 2 groups according to the presence or absence of individuals in the telephone directory. Persons with registered telephone number in the freely available telephone register of the national postal office (Editus.lu) belong to the CATI sample. Persons with no registered phone number belong to the CAWI sample.

In 2023, the quarterly sample size was reduced from 5,000 to 4,000.

To compensate, the CAWI subsample receives reminders twice, once after 3 weeks and again after 5 field weeks. In contrast, the CATI subsample receives only one reminder.

Our NSI sends an invitation letter (in two languages) to every individuals of the sample. The language version depends on the nationality and the age of the individuals. The letters are sent in Luxembourgish/French, Luxembourgish/German, German/English, French/English and French/Portuguese. A hotline is available for questions or to fill in the questionnaire by phone. A flyer, available in the same languages as the letters, informs about how to access the web application and shows some short statistics on tourism. A notice about the protection of data is sent together with the invitation letter.

The online questionnaire is available in Luxembourgish, French, German and English, which offers the opportunity for the respondents to answer the questions in the language in which they are most familiar.

The online questionnaire provides a possibility to make comments. We use these comments to regularly revise and improve formulations of the questions where necessary.

Fieldwork starts on the first opening day following the reference period.

Response rate reached 46% for the year 2023 (+5 points compared with the previous year). It varies from 44% to 46% according to the quarter.

The participation rate has increased again, since the introduction of a second reminder in the CAWI sub-sample.

The decision to add a second reminder for the CAWI subsample is an improvement. 

This decision give us the opportunity to reduce the burden of the surveys of Luxembourg's residents. Our country being of small size (650 000 inhabitants), the burden of socioeconomic surveys is more important here than in bigger countries.

See: Regulation 692/2011

Requests from international institutions such as OECD and UN Tourism.

Users such as STATEC (Balance of Payments, Tourism Satellite Accounts-TSA), tourism entreprises and organisations, Ministry of Economics, Ministry of Tourism, students and journalists have asked for more detailed data, absolute data or percentages.

Compliant with the requirements of Regulation on tourism statistics 692/2011, Delegated Regulation 2019/1681, as well as recommendations laid down in the Methodological Manual for tourism statistics.

450 651

1.1%

54 701

4.2%

3 281 481

1.7%

198 025

5.8%

3 083 455

1.2%

2 831 225

1.7%

450 256

5.4%

81 483

7.8%

116 543

8.3%

4 059 845 241

2.7%

1 343 072 420

2.8%

The sample for the tourism survey is drawn from the national register for residents every quarter of the year.

The over-coverage consists of individuals who are not eligible for survey and are included in the sample due to lack of information:

• people who left the country without reporting to the authorities,
• people who died in the meantime between preparing the sample and conducting the survey,
• persons living in collective households (prisons, retirement home).

Some of this over-coverage is corrected: individuals living in the main collective households of the country are withdrawn from the sample according to their addresses. For some other over-covered people, we got the information from a relative (by mail or phone).

The known over-coverage reaches 1.7%.

Under-covered immigrants recently arrived in the country, who are not in the national register for residents yet.

268 (death, live in collective households, age, live abroad)

15 732

8 046 (letters returned by Post, non-assigned phone number, target person unknown at the phone number, after 10 attemps no interview was made, non-contact in the CAWI sub-sample)

72

278 (partially filled questionnaires)

145 (language problems, physical or mental illness, temporarly not reachable person)

8 541

54.3%

54.3%

54.3%

The participation is possible through two modes: CATI and CAWI. Both modes are available for all participants.

Interviewers of the CATI center are trained.

Letters and leaflets are sent to the individual of the sample.

The leaflets give information about how to connect to the online questionnaire, how to change mode and results selected from previous waves.

The documents are bilingual, among 5 languages (Luxembourgish, French, German, English and Portuguese), chosen acoording to the nationality and the age of the individual.

A reminder is sent to the non-participants, as well as to those that partially filled in the questionnaire.

A second reminder is sent to the CAWI sub-sample.

Nearly all item non-responses consist of data on expenditure.

This is the case of 4 to 7% of the trips according to the topic: accomodation, transport, food and other, and valuable goods.

Few additional items non-responses concern booking through package trips.

To minimise item non-response, most of the questions are compulsory.

In addition, pop-ups and filters are integrated into the questionnaire.

Pop-ups inform the respondents in case of inconsistency in their answers.

Filters help reduce errors.

The questions on expenses and income are the only ones not compulsory. In both cases, besides the possibility of indicating a real amount, intervals are given to facilitate the answer.

When all costs for a trip are collected, a sum of all expenses is automatically calculated and presented to the respondent. He/she may correct the reported values if necessary.

To reduce the share of non-responses in the questions on expenses, an introductory text has been added in the questionnaire, to explain the difficulty of the questions and the flexibility allowed in the answers.

180

160

160

-27 days before the deadline.

-5 days before the deadline.

-31 days before the deadline.

Despite changes in the data collection method used over years (CATI, mixed-mode), in the samples size, more or less important participation rates to surveys, data on participation in tourism remains comparable over time.

Like for tourism participation, despite of the adaptations made year after year in the questionnaire, the method of data collection, ... , we consider the data on tourism trips stay comparable over time.

Break in series in 2022 due to the modification of the labels of the questions.

Individuals survey

Ëmfro iwwer den Tourismus (Luxembourgish)

Enquête sur le tourisme (French)

Umfrage über Tourismus (German)

Survey on tourism (English)

Survey on tourism

We are working with a sample drawn from the National Population Register where the residents of Luxemburg are registered. 

The composition of the sample is representative of the resident population in terms of age, sex and nationality.

16 000 residents aged 15 to 85 years old

4 000 residents aged 15 to 85 years old

16 000 residents aged 15 to 85 years old

7 191 residents

1428 residents

7 191 residents

Simple quarterly random sample from the population register

Quarterly for trips and same-day visits

Annually for participation data

90% used CAWI; 10% used CATI

See annexes

See annex

Not available

Proxies are allowed as far as the person responding is very well informed about the journeys of the sampled person. The given answers must reflect the behaviour of the sampled person.

CATI: 14 minutes

CAWI: 17 minutes

19 minutes

The validation process is built by steps:

• In the questionnaire: bounds, filters, revisions of answers by the respondents (trips expenses).
• Syntax checks: process that systematically compares the formats and values of the answers given in the questionnaire with theoretical formats and values compiled in the VDB (Variables DataBase).
• Coherence checks: process that controls the coherence between the answers given in the questionnaire (f.i.: destination and mean of transport, type of accommodation and expenses on accommodation ...).
• Outliers: outliers in the expenses are systematically controlled.

Missing values of the following variables have been imputed:

• Expenditures on Transport, on Accommodation, on Food/Drinks/Other, on Valuable Goods and on Package
• Use of package trip

Imputation method for expenses: predictive mean matching. Predictive mean matching was used because of its beneficial properties for skewed and heteroscedastic data as well as for the plausibility of imputed values. Furthermore, predictive mean matching enables an appropriate imputation of semicontinuous variables such as expenditure on valuable goods (Vink et al., 2014).

We conduct the imputation with the R programming language and the add-on package MICE. As provided by the package, we perform a multivariate imputation by chained equations, which enables the user to impute all variables simultaneously.

We estimated separate models for the imputation of leisure and business trips. To each of the imputation models, we included approximately 20 predictors:

•         time variable: the quarter (QUART)
•         social characteristics: sex, age, nationality, size of the household, economic status, education, income
•         expenditure variables: different expenditure variables were used as imputation predictors for each other
•         trip variables: number of trips made during the quarter, destination of the described trip, duration, transport, main accommodation, size of the group and package

For leisure trips, the following variables are added:

•         social characteristics: employment
•         trip variables: reasons, activities, location, package

4 to 7% of expenses are imputed.

The other missing values are predicted by a logistic regression using the proc MI in SAS.

The weights are built on the basis of the population estimation at 1st January made by STATEC. The following characteristics are taken into consideration:

• gender
• age group: 15-24, 25-44, 45-64, 65+
• nationality: Luxembourg-Foreigner.

No second survey or sources are used. All data come from the same survey on the resident population.