The Role of Web Interviews As Part of a National Travel Survey
L Christensen, C Jensen, DTU Transport, DK
A travel survey by personal interviews is a very expensive way to collect information about travel activities. On the other hand the detailed information about destinations and travel purpose which is necessary for a good travel model regardless of it is being a trip based or an activity based model can only be collected trough information from the traveller. National or local authorities might therefore always look for ways to collect these data cheaper. If the data at the same time could be of a higher quality it would be extra attractive.
The continuous Danish National Travel Survey which was collected by telephone interviews (CATI) was stopped in 2004 partly because of a really bad quality of data in 2000-2001 (Christensen, 2005 and 2006) and partly because of lack of money for collecting enough interviews. In the following year experiments with collecting some of the data on the web was carried out. The web survey showed up to be successful and it was decided to take up the continuous survey again from 2006 in a new organisation and with a web interview as part of the method.
Since May 2006 respondents are sampled from the Central Person Register and contacted by letter. They are asked to log into the internet and answer a questionnaire about their travel activities on a certain day. If they haven?t answered after two days they are called by telephone for a CATI if their telephone number is found. 13 % of the sampled persons are answering the questionnaire on the web which is 23 % of all respondents in the final database. This response rate on the web means that web interviews are considerable contributors to the economy of the survey. In fact the response rate on the web is rather low related to other Danish web interviews but this is between others due to the very short response period which is necessary to get precise travel information for an exact day. The information in the letter that the respondent will be called by telephone if he is not answering is also known to reduce the response rate (Alsnih, 2006).
The relevant question is therefore, is the web interview of the same quality as a CATI? And quite as relevant, do we get contact to some of those we do not get into contact with when we only use a conventional CATI interview? The method can bring us into contact with persons of whom we have no telephone number. And it can also make it possible to interview persons who are out in the afternoon and evening hours when the calls for the CATI are made.
In the paper quality is studied from two angels. The first problem is to what extent we get information of all the trips of the respondent. This is measured by the share of respondents who are not having a trip at the actual day and from the number of trips per person who is travelling. These two figures show up to be very vulnerable to bad quality in the survey process. And especially the no trip rate might be expected to be lower in a web interview than in a CATI if the interviewer is not doing his job quite well. The other measure is the quality of the answers where respondents at the web are answering less correct especially about destinations as the quality of this information can be improved by instruction and routine of the interviewer.
Another relevant question addressed in the paper will therefore be whether it is possible to improve the quality of the geographical and other problematic data in a process of aftercare. Furthermore it is discussed to what extent it will be economically feasible to increase the number of web interviews even when some of the interviews have to be scrapped because of bad quality. And at last it is discussed if the web interviews might introduce the possibility to get some other information which is not possible to collect in a CATI interview, for instance pointing out the exact route.
Alsnih, Rahaf (2006): "Characteristics of Web Based Surveys and Applications in Travel Research". Chapter 32 p. 569-92 in Stopher, P & Stecher, C: "Travel Survey Methods. Quality and Future Directions". Elsevier. The Netherlands
Christensen, Linda (2006): "Possible Explanations for an Increasing Share of No-Trip Respondents". in Travel Research". Chapter 16 p. 303-31 in the above book.
Christensen, Linda (2005): "Data Collection Biases in a Transport Survey" ETC Conference. Strasburg
Association for European Transport