The paper reports the survey practice from a unique study covering all local and regional political representatives in Sweden To our knowledge there are no previous survey conducted in any other country targeting all local political representatives. In this paper, we first report our successful multi-mode design that combines web-survey and paper questionnaires to maximize the response rate. We then compare the response patterns in the results from the web-survey and the paper questionnaires via post to evaluate whether the data collecting method affect the responses. The political representatives were first contacted via e-mail and invited to fill in the questionnaire in our online web-survey. Already during the first day, we had managed to achieve a response rate of over 10 percent. Thereafter, we approached the non-responding representatives several times, via e-mail and then via telephone up to ten times). Finally, each and every one of those who lacked an e-mail address were targeted and offered a paper questionnaire via post. All in all, after four and a half month of fieldwork, we had achieved nearly 10 000 responses and a response rate over 70.2 percent. This multi-mode design employing a combination of web-survey and paper questionnaire gave us the opportunity to test how the data collecting method affect the responses. We are therefore able to present comparisons of the different response patterns observed as regards item non-response and the response scale variance in the web and paper questionnaires respectively. In the web-survey result we find that non-item response is gradually increasing as respondents make progress in the questionnaire, while the non-responses in the paper questionnaire stem from questions on controversial and private issues. As regards the respondents use of scale end-points and variance response patterns we find significant differences between the two modes used.