Notes on Methodology

European Survey
The German Marshall Fund commissioned MORI (Market & Opinion Research International) to collect the data for this survey in six European countries: Germany, France, Great Britain, Italy, the Netherlands, and Poland. The survey was conducted by telephone interviews in all countries via Computer Assisted Telephone Interviewing (CATI) except Poland, where the telephone penetration is lower, and a face-to-face Computer Assisted Personal Interviewing (CAPI) approach was used. In each of the six European countries, a representative sample of 1,000 adults living in private households was selected. Households were selected by a random digit dialing approach. Pre-tests were conducted in all countries between May 13-22, 2002. The fieldwork for the survey was conducted between June 5 and July 6, 2002, by the institutes in the table below.

Each partner agency used a random digit dialing approach that best suited the country’s market with the aim of achieving interviews with a representative national sample. The random last/next birthday method was used, where interviewers asked to speak with the member of the household 18 years or older who had the last/next birthday (except in Great Britain and Poland). This systematic respondent selection technique has been shown empirically to produce samples that closely mirror the population in terms of age and gender. In Great Britain and Poland, respondents were chosen randomly, but quotas were set to ensure that a representative cross section of the population was interviewed.

The completed interview lasted approximately 20 to 25 minutes. The overall average response rate was 37%. Nonresponse in telephone interview surveys produces some known biases in survey-derived estimates because participation tends to vary for different subgroups of the population and these subgroups are likely to vary also on questions of substantive interest. In order to compensate for these known biases, data for each country was weighted according to known demographic characteristics of the population (by age, sex, and education). For each individual country, there were no significant differences between the weighted and unweighted figures. The results in this report for the individual countries are therefore unweighted. The figures for Europe as a whole are weighted on the basis of adult population in each of the six countries surveyed.

All the figures given for Europe as a whole in this report include the “don’t know” response alternative. In the documents released on September 4, 2002, the “don’t know” response alternative for Europe as a whole were not included for 6 of the 32 questions. The inclusion of the “don’t know” response alternative will slightly modify the results for the other response alternatives (often by 1 or 2%).

For results based on the total sample in each of the six countries, one can say with 95% confidence that the error attributable to sampling and other random effects is plus or minus 3 percentage points. For results based on the total European sample, the margin or error is less than plus or minus 1.4 percentage points.

U.S. Survey
Harris Interactive conducted 2,862 telephone interviews in the United States among men and women 18 years of age and older, using a random digit dialing technique with a national probability sample. In order to ensure comparability with the in-person Chicago Council studies of 1998 and previous years, personal in-home interviews with a national probability sample of 400 men and women 18 years of age and older were also conducted, using an abridged version of the telephone questionnaire that concentrated on questions repeated from 1998. All interviewing of the general public was conducted between June 1 and June 30, 2002. Data for the telephone and in-person interviews were weighted separately according to known demographic characteristics of the population and merged to form a combined sample (n=3,262).

In order to explore a very extensive set of topics, many questions were asked only of randomly selected subsamples of approximately 700 telephone respondents. “Core” questions, including most of those repeated from 1998, were also asked of the 400 in-person interviewees. Certain key questions were asked of all 2,862 telephone interviewees or all 3,262 respondents.The margin of sampling error in response frequencies varies negatively with the number of respondents asked a question and positively with the closeness of opinion division. For a fifty-fifty division of opinion (where margins of error are highest), at the p = .05 level the margins of error in this study range from 1.7 percentage points (for questions asked of all respondents) up to 4 percentage points (for questions asked of 700 respondents).

Separate analysis of the telephone and in-person data reveals that, as the literature would predict, there tend to be certain systematic “mode” differences in responses. Telephone interviewees, for example, tend to give fewer “don’t know” responses and to give more “positive” or first-option responses (e.g., more perceptions of vital interests and more ratings of goals as 'very important'). This does not mean that either method is incorrect; both meet professional standards and accurately reflect responses by the populations from which they sample. But mode differences do complicate the assessment of opinion changes from the in-person surveys of 1998 and previous years. This report is based on the combined 2002 telephone and in-person data set, which mitigates mode differences. In addition, only those contrasts with previous Council surveys that appear in both the 2002 combined data set and the 2002 in-person interviews taken separately are interpreted in this report as demonstrating opinion changes. An exception is made for the “active part in world affairs” question, for which the in-person responses do not show a significant change from 1998, but the much higher level of activism displayed in the 2,862 telephone responses is confirmed as indicating a real opinion change by others’ surveys conducted in 2002.

Chicago Council surveys have been carried out every four years since 1974. Prior to 2002, all but the first were conducted by the Gallup Organization; Harris conducted the 1974 survey.