Chapter 3. Exploratory study on travel information acquisition and use
3.1 Data availability
Many existing datasets are available online containing surveys or diaries for a wide array of topics: travel behaviour, travel information and attitudes to travel. Table 3.1 presents these datasets and their main characteristics including their accessibility. The main issue is that these datasets are usually isolated by topics, surveying different populations of individuals. To investigate the use of travel information, a unique dataset grouping travel information, travel attitude and behaviour is necessary. Several methods exist for data pooling or data fusion, i.e. to “merge” the datasets. D'Orazio et al. (2006) explains that, by using statistical matching, it is possible to statistically impute the missing data using standard inferential procedures.1 Although the advantage of statistical matching is to obtain one huge dataset
1 For example, the Office for National Statistics Internet Access Module contains demographics and
questions about Internet use, including some about travel information use. The London Travel Demand Survey contains demographics regarding demographics and travel patterns. Demographics are the set
Chapter 3. Exploratory study on travel information acquisition and use grouping variables from various categories across the same population of individuals, many questions are raised about the validity of its application.2 In addition to the uncertainties lying in the statistical matching literature (D'Orazio et al. 2006), it is not probable that such a synthetic dataset could be a better than a new survey questionnaire used for data collection.
Most relevant existing datasets are listed in Table 3.1. The Scottish Household Survey (SHS) is the only one containing an extensive questionnaire on travel information sources that is readily available in the public domain. In addition to demographics and travel patterns, it includes many variables relating to awareness, acquisition and use of travel information about three different sources. It has the further advantages of being readily available, of having surveyed a large number of individuals and of avoiding statistical matching methods. It was therefore deemed sufficient for use as a preliminary study.
The Scottish Household Survey (Scottish Government 2010) is a cross-sectional survey undertaken by the Government of Scotland that collects information on individual and household demographics, individual travel behaviour (via a 24-hour travel diary drawn from eight waves conducted between 1999 and 2008). In addition, and critically from the perspective of the current study, it collects information on the nature of the travel information sources consulted by travellers. The entire dataset consists of information on 27,238 households and 24,615 individuals from the 2007/2008 SHS dataset.
In the following preliminary studies, the effect of demographics and travel patterns are examined on the acquisition and use of travel information. These variables, which are also specific to Traffic Scotland (TS) are used for the analysis in subsection 3.2.3 and are listed and described in Appendix A.
of common variables. Internet use and travel patterns are the set of variables that are not jointly observed and could be imputed.
2 It assumes that both sets of variables that are not jointly observed are independent given the set of
common variables. This is the conditional independence assumption. In this case, the assumption could be easily challenged since Internet and information use may be related.
Chapter 3. Exploratory study on travel information acquisition and use Table 3.1 Potential datasets
Country Acronym Name Source Datasets of relevance (in addition to
demographics) Availability Year Respondents UK SHS Scottish Household Survey3 Economic and Social Data
Service, ESDS Travel information, travel behaviour, ICT use Online 2007-2008 27,238 UK NTS National Travel Survey Statistics Department for Transport / National and official
statistics Travel behaviour Online 2002-2010 20,000+ UK LTDS London Travel Demand Survey4 Transport for London Travel behaviour Upon permission 2005-2013 38,800
UK ICT ONS internet Access Module ESDS ICT use Online 2010 3,307
UK NPS National Passenger Survey National Passenger Survey, NTS Travel behaviour Online - - UK Att. Transp. Statistics on public attitudes to transport from NTS Travel behaviour Online - -
UK UKTUS UK Time Use Survey ESDS ICT use Online 2000 20,981
UK CC&T Climate Change and Transport Choices UK government, GOV.UK Travel behaviour, ICT use Online 2010 3,923 UK BSA British Social Attitudes NatCen Social Research ICT use Online 2010 3,297 UK Census Census 2011 Office for National Statistics Travel behaviour (regional level) Upon permission 2011 - UK IPSOS NHT survey IPSOS Mori (private company) Travel behaviour, travel information Private - - US NHTS National Household Travel Surveys United States Department of Transportation, USDOT Travel behaviour Online 2009 308,901 US WOC Work-Optimism-Cars Pew Research Center Travel behaviour Online 2008 2,260 US Mob. Mobility Pew Research Center Travel behaviour, ICT use Online 2006 2,003
3 Chosen for the exploratory study developed in this chapter.
Chapter 3. Exploratory study on travel information acquisition and use Demographic variables such as age, income and travel pattern (frequency of car driving) are continuous variables and are scaled back in amplitude to respect the relative homogeneity of scale across all variables. The degree of familiarity (and unfamiliarity) of respondents with the Traffic Scotland (TS) information source was determined by calculating the proportional number of correct (and wrong) information features mentioned by the respondents out of all the information features provided by Traffic Scotland.