Foreword METHOD

Method

An online survey was uploaded to a Web Server (surveymonkey.com) and was made available between June 1st and June 30th, 2006.  

Survey respondents were asked to rate their agreement with 32 statements on a seven-point Likert scale as specified below.  Each statement was designed to correspond with one of the trust-related constructs as indicated in section 2 above.  

Citizens were also asked to respond to 10 demographic questions based on those asked on the well established Euro-barometer survey .  

We requested that only EU citizens should respond to the survey which was offered in eight different languages in order to maximise the diversity of respondents.  The languages offered were: English, German, French, Spanish, Hungarian, Greek, Czech and Polish.  

The survey was promoted by all the participating organisations of FIDIS and in all eight languages.  The basic promotion strategy was to have members of the research team use their personal contacts, professional groups and the mailing lists of groups to which they belong.  Links to the survey were also placed on the LSE (www.lse.ac.uk) and FIDIS (www.fidis.net) websites and in various press releases.

Response rate. Overall, there were 2,918 responses to the survey.

Table 1: Breakdown of Respondents by Country 

While respondents to the survey came from 23 out of the 25 EU countries, low response rates from some of the countries prevented a valid comparison across countries.  Despite efforts to promote the survey, and like most web-based surveys, the current one cannot be said to be representative in this sense. Owing to this obvious bias in the data set, we have decided to analyze the responses by region, rather than by country, so as to allow a valid output. Thus, the analysis contains a regional comparison for each survey statement. In creating the regions for analysis, similarities in cultural background and the legal framework were taken into consideration. The approach resulted in the following regional clusters:

1. UK & Ireland 

2. Austria, Germany & Scandinavia 

3. Benelux & France 

4. Central and Eastern Europe 

5. Southern Europe 

Besides systematically comparing response means across these five regions, aspects of demographic variation were analyzed for some of the questions. Those variables were not used for groups that were highly under-represented. The following demographic variables were used for this analysis: 

1. age  

2. gender  

3. settlement size 

4. education  

The questionnaire also contained a question about the extent to which users provide personal data online in the past month. Similar comparisons by region were carried out for this variable. 

As indicated above, survey respondents were asked to rate their agreement with a list of statements on a seven-point Likert scale. The advantage of using this scale is that valid responses can be represented by numbers ranging from 1 to 7. In our case, 1 represents strong agreement with the statement and 7 strong disagreement. In the analysis, we treated the Likert scale as an interval-scale for which mean and standard deviation can be interpreted. The midpoint of the scale is at 4 (as the scale starts at 1). In the section that follows, we maintain the original structure of answers, where numbers less than 4 indicate degrees of agreement with the statement and numbers greater than 4 indicates disagreement. Four is the middle point of the scale, which we interpret here as neither agree nor disagree.