Visitor Travel Survey (VTS)
Detailed information for fourth quarter 2019
The Visitor Travel Survey (VTS) provides statistics on U.S. and overseas visitors to Canada, their characteristics of travel and spending levels.
The Visitor Travel Survey was introduced in January 2018 to replace the U.S. and overseas visitors to Canada component of the International Travel Survey (ITS record 3152 Archived).
Data release - July 8, 2020
The Visitor Travel Survey was developed to fully replace the visitor component of the International Travel Survey. The Visitor Travel Survey collects information about international travel to Canada by U.S. and overseas residents. Starting in 2018, the Canadian residents component of the former ITS became part of the new National Travel Survey (record 5232).
The electronic questionnaires (e-questionnaires) and Air Exit Survey (AES) are the two components of the Visitor Travel Survey (VTS). The objective of the VTS is to provide more detailed information about the characteristics of U.S. and overseas travellers to Canada, such as expenditures, activities, places visited, and length of stay.
Information from the VTS is used to meet the requirements of the Canadian System of National Accounts (Balance of Payments (BOP)). Moreover, the information collected in the questionnaires is used by the Tourism Satellite Account (TSA), Canada Border Services Agency (CBSA), Destination Canada, provincial tourism agencies, the United States Department of Commerce, the OECD, banks, investment companies, other private sector industries and independent researchers. The information is also used for reporting to international organizations such as the World Tourism Organization (WTO), the Organisation for Economic Co-operation and Development (OECD) and the Pacific-Asia Travel Association (PATA).
The AES started in the year 2000 for overseas visitors and in the year 2011 for U.S. visitors. The primary objective of the AES is to improve the quality and reliability of trip and traveller estimates for foreign air travellers to Canada, from major and emerging markets. Starting in 2017, tablets were introduced to replace the AES paper questionnaires. Although the main mode of collection for the AES interviews is the electronic questionnaire on the tablets used by the interviewers, paper questionnaires can still be used as a back-up in case of tablet failure, or for eight commonly used languages other than English and French.
The e-questionnaire component of the survey began in 2013, with the distribution of invitation cards to travellers (Canadian, American, and Overseas) who entered at one of 137 designated Canadian ports of entry. The mail-back questionnaires were last used in 2014. As of 2018, the invitation cards will be distributed to non-Canadian residents only as the VTS is only in-scope for U.S. and other overseas travellers.
Reference period: The Visitor Travel Survey collects information on a monthly basis and is disseminated quarterly. The quarter is determined by the date the traveller crossed the border into Canada.
Collection period: Brief collection periods, known as stints, occur at air and land ports of entry across Canada every month of each quarter.
- International travel
- Travel and tourism
Data sources and methodology
The Visitor Travel Survey targets all American and overseas residents entering Canada, except diplomats and their dependents, refugees, landed immigrants, military and crew. With the Air Exit Survey, it also targets American and overseas travellers returning directly to the USA or to selected overseas countries. The targeted countries are those from which we attract the most travellers.
The population observed by the e-questionnaire survey differs slightly from the target population. The questionnaires are directed at international travellers who use popular modes of transportation to enter Canada and who enter through ports with a significant amount of traveller traffic. Accordingly, invitation cards for the e-questionnaire are distributed at 137 designated ports of entry. These ports of entry constitute approximately 53% of all Canadian ports but account for the vast majority of international travellers who enter Canada. Invitation cards to complete the questionnaires on-line are actively distributed to American and international travellers who enter Canada by one of the following modes of transportation: automobile, commercial plane, commercial bus or commercial boat (Vancouver Seaport only). In 2018, 97% of all international travellers entered Canada using these modes of transportation. For each port involved, a sample proportional to the number of travellers who crossed the border the previous year is created, using data from the Frontier Counts (record number 5005).
In 2000, the AES was added to gather information on non-resident international travellers who leave Canada on direct flights to countries other than the United States. The number of interviews targeted at each airport for a particular month and a particular overseas country is approximately proportional to the square root of the number of travellers from that country who cleared customs through the port during the same month of the previous year. Based on the targeted sample sizes by port, month and country of residence, the commercial flight schedules are consulted in order to choose a representative sample of flights; balancing as best as possible Canadian versus foreign carriers, chartered versus non-chartered flights, and days of departure. The airports surveyed in 2000 were Halifax, Montreal, Toronto, Calgary and Vancouver.
In August 2011, the AES was extended to the United States air commercial travellers in the United States waiting areas of Toronto, Vancouver and Montreal airports. In January 2012, it was extended to Ottawa and Calgary airports (until December 2013 for Calgary). In February 2012, it was extended to Halifax airport.
In October 2017, the AES was expanded to five additional airports: Quebec and Edmonton (overseas and US visitors) and, Winnipeg, Saskatoon and Victoria (US visitors).
In 2018, 23,758 questionnaires from non-resident travellers were used to produce estimates, including ITS e-questionnaires and paper Air Exit Survey questionnaires.
In August 1999, during the development of the Air Exit Survey for overseas visitors, some testing occurred in different airports. The main points observed during the testing were the time required to conduct the interviews, the quality of the data obtained from the distribution of the questionnaire by Statistics Canada interviewers versus personal interviews, the potential response rate as well as logistics required to conduct the interviews. As a result, personal interviews were chosen over the distribution of questionnaires. Data collection for the AES started in January 2000.
Testing of the AES for American visitors occurred in 2011, which consisted of informal, face-to-face interviews with respondents awaiting their flight at Ottawa International Airport. A team of four interviewers approached travelers and administered the questionnaire to non-Canadian residents. Respondents were also asked a few follow-up probing questions once the questionnaire was completed to gauge their understanding. The use of tablets for collection of the AES data was tested in January 2017 in Toronto Pearson and Ottawa.
A redesign of the International Travel Survey commenced in 2011 and was implemented in 2013, with a reduction in paper questionnaires from five to two and the introduction of the e-questionnaire. The mail-back questionnaires were last used in 2014.
The e-questionnaire version of the paper questionnaire was thoroughly tested by members of the project team before implementation. The first phase consisted of specially-prepared test cases to verify specific functionality. In the second phase, data corresponding to actual trips was entered, to test both the performance and user-friendliness of the tool. Cognitive testing of the e-questionnaire was conducted by Statistics Canada's Questionnaire Design Resource Centre (QDRC) in both Ottawa and Montreal. A pilot test was carried out between March 26 and March 31, 2012. The goals of the pilot test were to test operational procedures and the on-line response application, and to estimate the take-up rate.
In 2018, the Visitor Travel Survey replaced the U.S. and overseas visitors to Canada component of the International Travel Survey (ITS). The Canadian traveller component of the ITS e-questionnaire was removed and the application was tested before implementation. No other changes were made.
This is a sample survey with a cross-sectional design.
The Visitor Travel Survey is a sample survey that targets all American and overseas residents entering Canada (with the exceptions mentioned previously). Historical data extracted from the Frontier Count system provides information about the number of travellers that came to Canada by the ports of entry (land ports, airports, seaports) and is a basis for determining the number of invitation cards to complete the e-questionnaire that should be distributed at each port and the number of Air Exit Survey interviews that should take place at each airport.
Survey processing serves to make the data suitable for tabulation and producing estimates through data editing, coding, and imputation. All reported data on electronic and AES questionnaires are captured, coded and verified.
As part of the Visitor Travel Survey modernization initiative, a new editing system has been developed with a greater focus on automation. Manual intervention will only be used when data fail edits and an automated resolution is not available.
Some errors occur in the data capture process such as an incorrectly recorded field, or missing data on the questionnaire. These errors must be located and corrected. Edit rules are designed to ensure the data are consistent with known characteristics, and that questionnaire entries are valid responses given the question.
For example, some ports of entry can only permit certain modes of entry. If an automobile is recorded in a response as a traveller's mode of entry into an airport, then the data entry is corrected. To make this correction the instance is investigated and either the port is changed, or the mode of entry is changed accordingly.
In addition, if all transportation costs are included in a package cost, but are also included in the entries for fare to enter Canada and fare to leave Canada, then the entries are corrected to remove the duplication of reported expenditures.
A number of variables have been derived on the data file by combining items from the questionnaire in order to facilitate data analysis. For example, expenses reported on package tours must be broken down into spending components and reported data for trips that involved multiple entries into Canada must be separated.
The VTS also asks respondents for information on any packages or inclusive trips purchased. This purchase is not reported in the expenditure section on spending while in Canada. Consequently, packages are separated according to the components covered within the collective package cost, and added to the relevant expenditure categories.
From the VTS definition of a trip, any exit from Canada marks the end of a trip. Respondents might report several trips under this definition within their completed questionnaire. These trips must be separated to allow for consistent analysis.
Finally, all variables that are required in further analysis are created and calculated.
A certain amount of data imputation is performed in the VTS whenever fields are missing within a questionnaire, or when an insufficient number of questionnaires are received to meet the VTS objectives.
In the Visitor Travel Survey, the target population (international travellers) are partitioned into imputation groups, based on port of entry and on selected traveller characteristics, such as country of residence, mode of entry and duration of stay. Total imputation (i.e. imputation of entire questionnaires) is performed for any imputation group, also called Port Factor Group (PFG), for which we have received an insufficient number of questionnaires for a given quarter.
Estimates are generated by PFG, and thus a minimum number of questionnaires must be used to provide quality estimates for the PFG. In instances where the minimum number is not met by received questionnaires from the current quarter, a sufficient number of questionnaires, selected amongst questionnaires from the same quarter of the five previous years belonging to the PFG are added to the current year's sample of that PFG for the reference quarter. For 2018, rates of total imputation are as follows: United States residents 69% and overseas residents 7%.
Total imputation is also carried out for several PFGs or strata that are outside the scope of the questionnaire distribution. These PFGs refer to modes of entry that are not targeted by the questionnaire distribution (train, private plane or boat, motorcycle, bicycle, foot, etc.) or to ports of entry that do not distribute questionnaires. There are 52 PFGs for which Statistics Canada does not receive any questionnaires. These represent less than 3% of the target population. For these out of scope PFGs, the characteristics of travellers are estimated using total imputation.
These imputed questionnaires are duplicates of questionnaires that were obtained in 1990 for the same quarter and same PFGs. When total imputation has been used in this manner, the quality of the subsequent estimates may be compromised due to limitations in available questionnaires that can be brought forward to the current quarter's sample. For example, if there are very few questionnaires available in past quarters, the current quarter's estimates may be of poor quality.
For estimation purposes, the responses obtained through the questionnaires are treated as a simple random sample from the total traffic in each stratum although the sample selection methodology is typically stint sampling. A stint is a predetermined time period during which sampling activity is taking place. The calibration groups (or post-strata) are defined by three elements: (1) port or group of ports of entry, (2) type of entry formed by combinations of mode of transportation used (car, air or other modes), length of stay (same-day or overnight), route of entry to Canada (direct from the U.S., direct from overseas of from overseas via the U.S.) and (3) the quarter of the trip. The data may in fact be subject to some degree of "distribution bias" due to the fact that not all categories of travellers are represented in the questionnaire distribution or to a non-response bias due to the fact that the individuals replying may not be representative of the travelling population.
Also, because it is known that travellers who conduct long trips have a tendency to respond less than travellers on shorter trips, a calibration, also known as "bias adjustment", is made to address this for certain U.S. commercial air travellers. The adjustment is done by using information obtained from the E311 Declaration forms, completed by all travellers entering by commercial air. A distribution of the travellers is obtained by trip purpose (personal or business) and by length of trip. These categories vary depending on the airport of entry to Canada and the country of residence of the traveller. This distribution is used to adjust survey estimates such that they align better with E311 Declaration cards' data (population data from the Frontier Counts). The progressive implementation of the Primary Inspection Kiosk (PIK) by the Canadian Border Services Agency, which started in March 2017, will bring changes to the way the bias adjustment is done.
For the United States travellers coming to Canada by car through the largest land ports equipped with the Integrated Primary Inspection Line (IPIL) system, another calibration is performed and is related to the state of residence. It is based on the vehicle state registration information obtained from the IPIL system. This adjustment is done so that VTS survey estimates better reflect the distribution of travellers by state of residence, especially those from the states that provide the largest volume of travellers.
From 2013 onward, a generalized variant of the raking-ratio is used for weighting adjustment of overseas travellers. This method, based on linear programming, ensures exact matching to constraints defined by country of residence, purpose-duration groups, and combinations of region and mode of entry. Region of entry is included since the majority of overseas travellers arrive at a small number of airports.
The following tables (see Additional documentation link, below) provide for the American Visitors and for the Overseas Visitors, the calibration groups (or post-strata) that are used for estimation purposes. There are 154 calibration groups for the American Visitors and 81 calibration groups for the Overseas Visitors.
For each of the stratum (or calibration groups), also referred to as Person Factor Group, it is possible to obtain the number of travellers arriving to Canada from the Frontier Counts system. These control totals are used for weighting the questionnaires returned by international travellers entering/re-entering Canada.
Supplementary distribution of the International Travel Survey took place from January 2010 to August 2013 at high volume land ports such as Fort Erie, Niagara Rainbow, Windsor, Sarnia and Sault Ste-Marie in Ontario, Lacolle in Québec, as well as Douglas and Pacific Highway in British Columbia. As a result, a significantly increased number of questionnaires of the International Travel Survey were filled and returned by travellers during this time period.
This resulted in better estimates of the characteristics and spending of international travellers at these border crossings that affect their comparability with estimates of previous and subsequent years. Caution should be exercised in time series analyses for the implicated provinces and at the national level.
The extension, in August 2011, of the Air Exit Survey to include United States commercial air travellers has improved the quality of current estimates concerning United States commercial air travel to Canada as compared to historical estimates. Consequently, use caution when comparing current and historical estimates for this flow of travellers.
Some variables, such as length of stay and trip purpose are revised in consideration of the data given by the Frontier Counts Survey. In this context, "revised" refers to the bias adjustment that is made in the production of weights. The weights are adjusted to reflect benchmarks obtained from the administrative data which are used to produce frontier counts. Since additional information is obtained for some portion of the population, (such as; purpose and duration of trips for a sub-population of air travellers or state of origin of US Visitors coming to Canada by automobile), an adjustment of the estimates can be made to reflect these categories. This can help in addressing potential bias present in the samples of questionnaires.
Under the present operational conditions, the size of these samples is adequate to permit quarterly estimation of expenditures at the national level, and annual estimates for certain provinces, provided that the assumption of negligible bias is not violated. However, some data for lower levels of aggregation and cross-tabulations are not sufficiently reliable to be published on a regular basis.
Data quality is systematically evaluated every quarter. Statistical tables required for analysis are produced and compared with related data sources. A set of indicators is also produced. They are used to determine whether general tourism trends reflect those of the VTS. Furthermore, Statistics Canada works in close cooperation with Destination Canada and the provincial tourism departments who provide additional viewpoints and information sources to help evaluate the data quality at a more refined geographic level.
Statistics Canada is prohibited by law from releasing any information it collects that could identify any person, business, or organization, unless consent has been given by the respondent or as permitted by the Statistics Act. Various confidentiality rules are applied to all data that are released or published to prevent the publication or disclosure of any information deemed confidential. If necessary, data are suppressed to prevent direct or residual disclosure of identifiable data.
For this reason, identifiers are not included on the public user micro-data file (PUMF), as well as some socio-demographic and geographic variables which could have been used to identify respondents. Other variables that are on the PUMF have been grouped into broad categories. Before the official disclosure of microdata for VTS, its content will need to be assessed and approved by Statistics Canada's Microdata Release Sub-committee.
Revisions and seasonal adjustment
The quarterly preliminary results of VTS are produced using questionnaires for trips with a duration of up to 5 months. The quarterly final results are produced with the annual release of VTS. These final results, as well as the Public User Micro-data File are produced using additional questionnaires, with trip duration of up to 8 months. All the processing as well as weighting are redone to include the longer trip questionnaires.
Because Statistics Canada does not control the distribution of the invitation cards by CBSA border service officers, response rates cannot be calculated for the VTS. It is known that distribution, particularly at high volume ports, can be poor. In 2000, the survey had 54,600 returned and completed questionnaires as a result of CBSA distribution; this number dropped to 14,000 in 2012. As a result, in the corresponding period the number of imputed records has increased from 39,900 in 2000, to 57,200 in 2012. This is less of a problem on the overseas residents' file where the Overseas Air Exit Survey supplies many questionnaires. It is also less of a problem for US travellers because of the AES (US). Initial trend analysis on the United States file has shown that much of the imputation is of short duration trips and the trend, in terms of spending, is not changed much by the imputation. Statistics Canada is working with CBSA to improve invitation card distribution and is looking into developing alternative methods of imputation.
It is not possible to determine the scope of the non-response bias. However, the low return rate of mail-back questionnaires has led the VTS to question the reliability of some of the data. In the mail-back questionnaire component, almost 1,000,000 questionnaires were sent in 2008 to Canada Border Services Agency for distribution, while the sample included approximately 36,600 returned questionnaires, including the AES. The AES is an addition to the mail-back questionnaires component and obtains a cooperation rate superior to 95%. In 2012, 400,000 questionnaires were sent to CSBA for distribution, 161,000 questionnaires were distributed by Statistics Canada's interviewers as a supplementary distribution and 10,000 questionnaires were sent for the AES. 40,000 were completed and returned to Statistics Canada.
In 2018, 200,000 invitation cards to complete the survey on-line were provided to CBSA to distribute to every visitors to Canada crossing the border during a one-week stint each month. The number of completed questionnaires received back were less than 1,000, requiring heavy reliance on the approximately 9,000 AES questionnaires returned as well as the use of questionnaires from previous years in the weighting strategy.
Because of the complex nature of the weighting procedures of the VTS, simple calculations of the variance using standard methods tend to underestimate it. Consequently, Statistics Canada uses the "bootstrap" method for estimating the VTS variance. Under this method, the sample units are sub-sampled and re-weighted many times over (at the moment, the ITS uses 200 bootstrap subsamples). From these sub-samples, an estimated variance is obtained to calculate the estimated coefficient of variation (C.V.). The coefficients of variation are less than 5% for the main variables at the Canada level, however, the presence of imputed records may lead to systematic underestimation of variance.