Travel Survey of Residents of Canada (TSRC)

Status:
Active
Frequency:
Quarterly
Record number:
3810

The Travel Survey of Residents of Canada (TSRC) is a major source of data used to measure the size and status of Canada's tourism industry. It was developed to measure the volume, the characteristics and the economic impact of domestic travel. Since the beginning of 2005 this survey replaces the Canadian Travel Survey (CTS).

Detailed information for first quarter 2014

Data release - November 3, 2014

Description

Since the beginning of 2005, the Travel Survey of Residents of Canada (TSRC) has been conducted to measure domestic travel in Canada. It replaces the Canadian Travel Survey (CTS). Featuring several definitional changes and a new questionnaire, this survey provides estimates of domestic travel that are more in line with the international guidelines recommended by the World Tourism Organization (WTO) and the United Nations Statistical Commission. In 2011, TSRC underwent a redesign. Please refer to the document entitled Differences Between the 2011 Redesigned TSRC and the 2010 TSRC available in the "Documentation" section located at the end of the detailed information for this survey, for an explanation of the differences between TSRC from 2006-2010 and TSRC in 2011.

The Travel Survey of Residents of Canada is sponsored by Statistics Canada, the Canadian Tourism Commission, and the provincial governments. It measures the size of domestic travel in Canada from the demand side. The objectives of the survey are to provide information about the volume of trips and expenditures for Canadian residents by trip origin, destination, duration, type of accommodation used, trip reason, mode of travel, etc.; to provide information on travel incidence and to provide the socio-demographic profile of travellers and non-travellers. Estimates allow quarterly analysis at the national, provincial and tourism region level (with varying degrees of precision) on:

- total volume of same-day and overnight trips taken by the residents of Canada with destinations in Canada,

- same-day and overnight visits in Canada,

- main purpose of the trip/key activities on trip,

- spending on same-day and overnight trips taken in Canada by Canadian residents in total and by category of expenditure,

- modes of transportation (main/other) used on the trip,

- person-visits, household-visits, spending in total and by expense category for each location visited in Canada,

- person- and household-nights spent in each location visited in Canada, in total and by type of accommodation used,

- use of travel packages and associated spending and source of payment (household, government, private employer),

- demographics of adults that took or did not take trips, and

- travel party composition.

The main users of the TSRC data are Statistics Canada, the Canadian Tourism Commission, the provinces, and tourism boards. Other users include the media, businesses, consultants and researchers.

Subjects

  • Domestic travel
  • Travel and tourism

Data sources and methodology

Target population

The target population is the civilian, non-institutionalized population 18 years of age or older in Canada's ten provinces. Specifically excluded from the survey's coverage are: residents of the Yukon, Northwest Territories, Nunavut, persons living on Indian reserves, full-time members of the Canadian Armed Forces, and persons living in institutions. Together, these groups represent an exclusion of less than 3% of the Canadian population aged 18 and older.

Instrument design

The implementation of the Travel Survey of Residents of Canada (TSRC) in 2005 was planned by a task force composed of members from Statistics Canada, the Canadian Tourism Commission, several tourism provincial organizations/ departments and other federal organizations. This pan-Canadian task force was involved in the development of the concepts, the platform and the content of the questionnaire.

In order to test several components of TSRC, a qualitative study evaluation and a quantitative survey evaluation were held across Canada. The qualitative study evaluation, which included focus group discussions and in-depth interviews in both of Canada's official languages, preceded the quantitative survey evaluation. The goal of the qualitative study was to test a new introduction to the survey and different trip definitions.

The major objective of the quantitative survey evaluation was to recommend an introduction to the survey that would be the most successful in encouraging Canadians to volunteer information on their travel activities, while complying with the Task Force's interpretation of the World Tourism Organization's (WTO) definition of tourism.

In addition, the TSRC questionnaire was assessed at Statistics Canada by the Questionnaire Design Resource Centre and tested internally. Starting in 2009-2010, the TSRC team worked in collaboration with the pan-Canadian task force to redesign the TSRC questionnaire. All were consulted in order to identify the questions or variables that could be removed, changed or improved and on the implementation of new efficiencies while building the questionnaire and the collection application itself. It had some impact on the way some questions are now asked.

Some effort was also spent on harmonizing questions such as activities, accommodations and transportation modes with the International Travel Survey (ITS).

The methodologists tested the best approach to sub-sampling of trips that would enable the collection of a maximum number of trips within the 15 minutes interview time limit and get as much detailed information as possible. Two qualitative tests were conducted. The first focus groups held in March 2010 were oriented in obtaining feedback from respondents on their overall impressions of and reactions to the proposed content and questions, assessing the respondents' understanding of the concepts, terminology, questions and response categories, their ability to accurately report the main reason for taking each reported trip (two scenarios for these questions were tested), the availability of the information requested and the respondents' ability and willingness to answer the questions. The outcome of the tests led into changing the flow for the main reason questions and changing some wording to make some concepts clearer to the respondents.

To finalize the application, another round of testing was held in the Sturgeon Falls regional office in September 2010. The purpose of this test was to have experienced interviewers go through varied pre-determined paths in the application to assess the changes and suggest improvements where applicable. This was an opportunity to express any concerns about the revised content and collection.

Sampling

This is a sample survey with a cross-sectional design.

The Travel Survey of Residents of Canada (TSRC) is a voluntary supplement to the Labour Force Survey. This means that the TSRC is run using one or more subsamples of households that are in the Labour Force Survey (See LFS survey documentation - record number 3701 - for more details on sampling). The Labour Force Survey is a monthly household mandatory survey of about 54,000 households that provides official estimates of employment and unemployment in Canada with a 90% response rate. Of these responding households, the TSRC has a slightly lower response rate of between 70 to 75%.

The Labour Force Survey sample of individuals is representative of the civilian, non-institutionalized population, 15 years of age or older in Canada's ten provinces. Some people are specifically excluded from the TSRC coverage. These include residents of the Yukon, the Northwest Territories and Nunavut, people living on Native Reserves, full-time members of the Canadian Armed Forces and inmates of institutions. Each month, this survey collects employment data for the week that includes the 15th of the month, during the week after the 15th of the month. Since the Labour Force Survey samples all households in Canada, it is used as a platform to run other household surveys including TSRC.

After the Labour Force Survey questions have been answered, the collection application randomly picks a respondent of 18 years of age or older from the selected household who is then asked to respond to the TSRC questionnaire.

Data sources

Responding to this survey is voluntary.

Data are collected directly from survey respondents.

Only households that respond to the Labour Force Survey are included in TSRC. After the labour force questions have been asked, within a selected household, one person 18 years of age or older is randomly selected to respond to TSRC. If this person is available, the interviewer will ask him/her the TSRC questions. If this person is not at home or not available at the time of the LFS data collection, an interviewer will call back later during the data collection period.

The interviews are conducted over the telephone by regional office interviewers, using the computer-assisted telephone interview (CATI) method.

Proxy interviews are permitted if the respondent does not speak one of the official languages or is ill or absent for the duration of the collection period.

The systematic interviewer monitoring system implemented in the regional offices in 2002 is also used for TSRC. This is an automated system that enables supervisors to see the interviewer's screen remotely and hear the telephone conversation between the interviewer and the respondent. The monitoring of interviewers serves to improve the collection of information from respondents and thereby enhances the quality of the data published by the survey.

View the Questionnaire(s) and reporting guide(s) .

Error detection

The TSRC computer-assisted interview (CAI) questionnaire incorporates many features that serve to maximize the quality of data collection. There are many edits built into the CAI questionnaire which compare unusual responses and that check for logical inconsistencies. CAI controls the sequence of questions in accordance with responses to previous questions. Other checks are also made during the interview to reduce the number of errors attributable to typos and misunderstandings. For example, if the number of nights spent in various types of accommodation does not correspond to the total number of nights spent away from home, an edit message will appear on the screen. The interviewer can then correct the mistake, and less editing has to be performed at the Statistics Canada head office. As well, for each question the interviewer has the possibility to enter "Don't Know" or "Refused" as a valid response if the respondent does not answer the question.

Data are transmitted to head office from the regional offices where a series of processing steps occur to prepare the data for dissemination. Verification of the raw data is completed in order to identify inconsistencies. Trip records are validated to ensure that values in critical fields are acceptable. For example, the number of nights is within the logical range, the type of trip is valid, etc.

For the majority of trip records the coding of the geographic area has been automated and is done in the field during collection. For a small number of records the coding of the geographic area is manually done at head office.

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, MRDTRIP2, which identifies the main reason for taking the trip, has been derived using several questions from the questionnaire as input into this variable.

Several consistency edits are carried out on the data to verify the relationship between two or more variables. An example of this is the number of adult household members who went on the trip cannot exceed the total number of adult household members.

Expenditure imputation is performed on trip records to ensure that all trips have a valid expenditure. If the system is not able to impute a value, a value of 'not stated' is assigned.

Trip imputation is performed for all non-selected trips. The collection application collected key information for all trips declared by the respondent. Then, through a sub-selection process, only certain trips were selected for collection of additional information. Imputation is done on the non-selected trips using the selected trips as donors.

Outliers identification is done using weighted aggregate data and anomalies are treated.

Imputation

Two types of imputation are done on TSRC data: trip expenditures imputation and trip imputation.

The trip expenditures imputation is done for expenditure data to ensure that every record has a valid expenditure value for all expenditure categories. All other missing or erroneous values are either corrected or converted to a "not stated" code.

Expenditure values are imputed based on the average expenditure for this variable from other selected trips with similar characteristics. Common characteristics are used since expenditures vary considerably depending on trip characteristics. Averages are computed for trips with common characteristics like the destination, the duration, the number of people in the travel party, the main reason for the trip, etc. Averages are computed separately for trips from the first and second recall months.

An additional step in the imputation process is the distribution of tour package expenditures to specific expenditure categories. This is accomplished in the same fashion as expenditure imputation.

Imputed expenditures are then re-edited to ensure that no outlier values have been created by the expenditure imputation process.

After this procedure, an extra adjustment is performed for same-day trips with identical trips and high expenditures.

The trip imputation is used because of the sub sampling strategy. First, respondents' trips are rostered and basic information is gathered. Then, a subset of the respondents' trips is selected for the collection of additional information which is referred to as the detailed trip information.

This strategy leaves all of the detailed trip information variables with missing values for unselected trips. This information is imputed by means of a donor imputation strategy. For each unselected trip (the "recipient") one single trip (the "donor") is selected to impute all the detail variables. The information from this donor is borrowed to fill in the missing values in the recipient. Therefore, it is important to select the donor carefully. Also, note that the imputation is performed separately by collection month.

The process of finding a donor for the unselected trips resembles the way averages are calculated in the expenditures imputation stage. The donor is selected from a set of trips with similar characteristics to the recipient, called "donor pool".

Only trips from the same recall month and trip type (same-day or overnight) as the recipient are considered to be part of the donor pool. These are mandatory matching categories. In addition, trips with high reported total spending are excluded from the donor pools to avoid creating new outliers through imputation.

After the donor pool is finalized, the final donor is selected from the pool using a distance function that indicates how far each of the donors in the pool is from the recipient in terms of some continuous variables. For same-day trips, the donor whose trip distance is closer to the recipient's is chosen. For overnight trips, a more complex distance function, which incorporates the number of nights and the trip distance, is used to select the donor.

Finally, after the donor is selected, all its information is copied over to the recipient. A donor can be used more than once.

Estimation

TSRC estimates are produced based on survey data to which weights are applied, making it possible to inflate these data to agree with the Canadian non-institutionalized population 18 years and older. The weights calculated to produce estimates are person, trip and person-trip weights.

The starting point in creating the person weights is the Labour Force Survey (record number 3701) sub-weight. The person weight is then adjusted to reflect i) the subsampling of rotation groups, ii) subsampling of people (18+) within a household, iii) non-response and iv) calibration to known control totals (age/sex groups, CMA totals). Person records from the second month of collection are also treated for recall bias.

From the person weight, the person-trip weight is derived by adjusting for i) identical trips, ii) the ratio of declared to reported trips, iii) reported trips missing essential data, and iv) trip-level nonresponse. Person-trip weights are also treated for outliers, and records from the second month of collection are adjusted for recall bias. These weights are used to estimate trip volume.

Finally, the trip weight is derived by dividing the person-trip weight by the number of adults (18 and over) in the household who went on the trip. Trip weights are used to estimate expenditures.

The non-response groups are created by estimating response propensity based on demographics and paradata. For each province, person records are clustered in homogeneous response propensity groups. Within each non-response group, the estimated response rate is computed and the inverse of this rate is used to adjust the weights.

The TSRC uses the coordinated bootstrap method, a replicate based method, for calculating variances. The bootstrap method involves taking sub-samples from the sample and re-weighting. This is done many times, 500 times for the TSRC. Similar to the regular bootstrap, the coordinated bootstrap takes into account the subtle correlation in the TSRC. This correlation arises because the TSRC is a supplement to the LFS and in any given month, there is a large overlap between any 2 consecutive reference months (because any given interview contributes to the estimates of two consecutive reference months).

Quality evaluation

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 TSRC. Furthermore, we work in close cooperation with provincial tourism departments, which provide additional viewpoints and information sources, helping us evaluate data quality at a more refined geographic level.

Disclosure control

Statistics Canada is prohibited by law from releasing any information it collects which 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.

Disclosure of results is very closely monitored, especially with regard to the public use of microdata files. For TSRC the smallest geographical area for which information is available to the public, regarding traveller and travel records, is that of the census division and census metropolitan area. Furthermore, socio-economic variables such as the respondent's age and occupation have been deleted or recoded in order to eliminate any possibility of identification. Before the official disclosure of microdata for TSRC, its content will need to be assessed and approved by the Microdata Release Sub-committee.

Revisions and seasonal adjustment

The quarterly preliminary results for TSRC are revised when final annual estimates for that year are made available.

Data accuracy

The TSRC overall response rate for the reference year 2013 was 75.6%. Response rates varied by province and territory.

Sampling variability is the error in the estimates caused by the fact that we survey a sample of respondents rather than the entire population. Standard error and the related concepts of coefficient of variation (CV) and confidence interval provide an indication of the magnitude of sampling variability. The standard error and coefficient of variation do not measure systematic biases in survey data that might affect estimates. Rather, they are based on the assumption that sampling errors follow a normal probability distribution.

Usually, the larger of the two estimates will have a smaller CV, and therefore be more reliable. Also, for two estimates of the same size, the one associated with a characteristic more evenly distributed throughout the population will tend to have a smaller CV.

Simply speaking, the CV is used to identify three major quality levels of data:

- 16.5% and below Acceptable;
- 16.6% to 33.3% Marginal, with cautionary note; and
- Above 33.3% Unacceptable estimate.

Estimates that do not meet an acceptable level of quality are either flagged for caution or suppressed.

Documentation