By: Sandford BorinsI never participate in focus groups or any other market research. But the Citizens First survey that arrived in the mail was different. For almost a decade The Institute for Citizen-Centred Service (ICCS)has been surveying Canadians about their experience with government services, and I was delighted to have been chosen for their latest questionnaire.This is important work, as it helps governments know if their service improvement initiatives have increased satisfaction and helps them design future initiatives. Canada's extensive use of survey research has been one of the factors responsible for its top spot in Accenture's international rankings of e-government performance.Here are some reactions to the questionnaire, suggestions for improvement, and ideas about how to analyze the data.Many important questions are asked in the survey: assessments of the full range of services by all three levels of government, a detailed assessment of one service experience, a question about whether you are willing to have a common username and password for all government services, a penultimate question about what you consider to be good service, and the standard demographic questions at the end.In survey research, the order in which questions are asked can affect the answers. The questionnaire started by asking you to choose one level of government and then to give your overall assessment of service at that level. Because each level of government delivers a wide range of services, I would have asked this summative question at the end, to be considered after evaluating all the services delivered at each level.I would have put the question about what you consider to be good service at the beginning, to help you be clear about the standards against which you are evaluating specific services.While asking about whether you were willing to have a common username and password was a good way to establish privacy concerns, the questionnaire should also have asked whether you were willing to have government programs (and levels of government) share data about you without your explicit consent, perhaps by laying out a scenario for general authorization of data-sharing.Here are two ideas for analysis. First, the detailed question about one chosen service experience (for example, passports) could be extraordinarily helpful if it generates a sufficient number of responses about particular services.For statistical testing, 30 is generally considered to be the smallest acceptable number, but smaller samples might still be useful for program managers. Second, the data about service standards could be correlated with the demographic questions to see if standards vary by, for example, age, income, education, and Internet use.The covering letter tells us that results of the survey will be summarized on the Institute for Citizen-Centred Service Web site next year. I'd go beyond that to suggest an open-source approach. After the survey's consultants have had their crack at the data, ICCS should make all the data available to anyone who wants to analyze it. This is the practice some major election polling companies (e.g. Decima Research) follow.Citizens First is producing a lot of valuable information, and analysts in academe could come up with some imaginative ways to interpret it. Let's bring as many minds as possible to the table.I won't be posting next week, as I will be on the Digital State tour. I'll post in two weeks to talk about what I've learned on the road.