...
3.108. Such surveys are conducted through face‑to‑face interviews or by telephone. Countries that had achieved high penetration of landline telephones in households found that telephone surveys considerably reduced cost by eliminating the need for interviewers to be sent around the country. However, in recent years, the rapid growth of mobile (cell) phones and the associated lower penetration of landlines in households are generating concerns regarding the representativeness of such a methodology. If interviewing is conducted only via landlines and does not include mobile phones, there could be a significant bias in the sample, e.g., an underrepresentation of younger people who have a higher propensity to use only a mobile phone.
| Tip | ||
|---|---|---|
| ||
Box III.27 Representativeness of CATI: example of Austria For the purpose of achieving representativeness, the gross sample for the Austrian demand‑side survey is drawn from the Central Registration Register of the Ministry of the Interior, which enables a stratified random selection. Telephone numbers for the persons in the gross sample, are obtained from the official telephone directory, by using last name and address. The official telephone book covers landline as well as cellphone numbers. However, owing to growth in the number of confidential numbers (since registration in a telephone directory is no longer required in Austria) and the replacement of landlines by cellphone numbers not available to the public, telephone numbers could not be found for approximately 50 percent of the persons in the gross sample in 2011. Bias is likely due to the fact that the persons with no telephone numbers listed in the official telephone book cannot be covered. __________________________ Source: Statistics Austria. |
| Tip | ||
|---|---|---|
| ||
Box III.28 Mobile‑phone sampling: example of Australia In 2013, Tourism Research Australia (TRA) pilot‑tested mobile‑phone sampling for its National Visitor Survey (NVS). Full‑scale sampling commenced at the beginning of 2014, with the mobile‑phone component constituting 50 percent of the total sample in the first year. In pilot‑testing (entailing 800 interviews), response rates for the mobile‑phone sample were similar to those for the existing landline sample. Further surprisingly, respondents voiced few concerns about being contacted on their mobile phone. As the NVS will be an “overlapping dual‑frame” survey from 2014 onwards, the sampling and weighting will become more complicated, and that a potential source of problems. However, TRA has invested considerable resources to ensure the sample design is appropriate, and that the new weighting process is well tested and well understood. __________________________ Source: Australian Bureau of Statistics. |
3.109. As mentioned in section B.2 above, the frequency of household‑type surveys is basically dependent on cost; however, other factors, relating to the characteristics of resident visitors and their tourism trips, should also be considered when the frequency of observation is being determined.
3.110. Countries may try to establish the characteristics of trips and of persons taking trips independently of,or with frequencies that differ from that used in, the measurement of associated expenditures. Domestic tourism flows should be measured with high frequency (monthly or even more frequently in specific periods such as high season), since their characteristics tend to fluctuate significantly throughout the year (exhibiting seasonality) and, although to a lesser extent, from year to year (see paras. 3.18–3.19). The average expenditure per person or per person per day associated with each characteristic of a tourism trip, on the other hand, tends to be less volatile. Hence, it can be observed less frequently (e.g., every three to five years) and measured through extrapolation using relevant price indices once the volume and characteristics of the trips are clearly established. The advantage of this approach is that it reduces the need for collection of good‑quality expenditure data, which is difficult and can add considerable cost to the survey when included.
| Tip | ||
|---|---|---|
| ||
Box III.29 Selection of trips in a travel survey: example of Canada In Canada’s domestic travel survey (the Travel Survey of Residents of Canada (TSRC)), one adult is selected at random from each household surveyed. At the beginning of the interview, a roster is created consisting of all trips by the respondent that ended in either the first or the second recall month. The roster contains the following information gathered during the interview: main destination, main reason for trip, duration, when trip ended, number of household members on the trip, and mode of transportation. The interviewer collects information on domestic trips and on the Canadian portion of international trips for the roster. From the roster of trips, between one and three in‑scope trips are randomly selected and information is requested on the detail of the trip(s), namely, secondary mode of transportation, travel party, expenditures, and activities engaged in during the trip, as well as locations and accommodations for overnight trips. The subselection of the trip is based on an algorithm that takes into account the nature of the trip, i.e., reference month, inter versus intra‑provincial, overnight versus same day, and number of identical trips. __________________________ Source: Statistics Canada. |
3.111. Apart from other particularities (see sect. D.2), the following issues arise in the use of household surveys for purposes of tourism:
...
- In cases where a household (or the individual being interviewed) has taken more than one in‑scope trip during the reference period, some countries use the following methodology, they collect details of only one trip (usually the last one taken) but count the total number of trips taken. These data would then be used in the sample expansion process to estimate the total number of trips taken.
3.112. Besides the personal characteristics of the person taking the trip and, eventually, of the travel party, which are all characteristics attached to the trip, the following additional characteristics should be observed:
...
3.113. Same‑day trips and trips to vacation homes should be identified separately and given special treatment within a specific sub‑module.
[1] The individuals are also collection units, as they might take more than one tourism trip during the reference period.