Below are some tips for running or supervising a peace poll:

• Cover all major aspects of social and political life effected by public institutions and government departments as 'the people' and their 'political representatives' often have very different views (and interests) about the nature of the conflict and its resolution.
• Encourage key decision makers to become involved in drafting the research questions and designing the methodology so that they will take the results more seriously.
• As politicians may be skeptical about the benefits of public opinion polls first undertake a program of pure research to demonstrate the independence and validity of the work.
• If politicians disagree with the results of the pure research poll ≠ this is welcome - invite them to help design the next survey to their satisfaction.
• Do not exclude any serious parties from the applied research - it is most helpful to test support for mainstream opinion, centre party compromises and radical reforms together.
• If the large established parties do not show willing try the small centre parties first after which the larger parties may decide they do not wish to be left out.
• Start with some simple confidence building questions about the peace process in general and other confidence building measures (CBMs) that could easily be implemented.
• Deal with all of the principal procedural or 'shape of the table' issues before getting into too much detail over substantive or 'negotiated settlement' issues.
• In public opinion polls the elimination of extreme positions, those with little cross community support, is just as important and just as easy as finding compromises and common ground.
• Sometimes questions that have been drafted can not be run in a poll for lack of space. This is not entirely a bad thing as it provides a working foundation for later polls.
• Systematically deal with all preconditions and objections to a peace process ≠ people generally want 'jaw jaw' in preference to 'war war'.
• Do not avoid sensitive issues because others might take on those same questions in a less helpful way that is potentially more damaging to the peace process.
• Give 'the people' every opportunity to answer questions about the exercise of their democratic franchise ≠ they like it ≠ and the results should send a message to their elected politicians.
• Devise questions that can produce a ranking of the major problems in a conflict and their potential solutions.
• Develop questions that include all of the potential elements of a final agreement by way of informing both the negotiators and the general public.
• Do not be put off by complexity. The people living with a conflict often have a very sophisticated understanding of that conflict.
• Use a method of analysis that reflects the voting procedures used in the negotiations proper in terms of both constituencies and levels of support required.
• Test comprehensive agreements as a 'package' as many of its problematic elements will be acceptable as part of a balanced settlement.
• 'Underline' the politically unacceptable alternatives to a comprehensive settlement when it is opportune to do so. For example when radical groups are actively opposing a 'deal'.
• Timing is of the essence. For example testing a 'Comprehensive Settlement' would be almost useless if run months before the parties are ready to 'cut a deal' or the day after talks collapsed!
• Try to retain control over funding so that the parties involved with the polls will not be able to exercise a veto if they think the work is not going to go their way.
• Don't use public opinion polls to renegotiate agreements. Regrettably much of the partisan media will do this anyway.
• Don't assume the work is over once the deal is signed particularly if many of the issues raised in the research are not dealt with in the agreement.
• Even when a very difficult decision has to be made try and include all the critical parties to that decision - however difficult that makes the work.
• When key players refuse to negotiate use neutral parties to feed in constructive suggestions.
• When key players introduce questions designed to produce an unhelpful result get neutral parties to critique the value of such questions.
• Design and run 'cold shower' questions that explore the consequences of failure when the point of 'do it or lose it' is reached. Public opinion polls are an excellent medium for dealing with 'contextual' issues.
• Try not to end the research arbitrarily. Let the parties have a say in when to run the last poll as they are ultimately responsible for the success of the peace process.
• When support for running a public opinion poll is 'mixed' consult widely and do not be afraid to temporarily poll against the wishes of some parties.
• Have an experienced board or advisory group at hand to back up difficult polling/ethical decisions.
• As an independent facilitator or mediator it is generally inappropriate to express personal opinions but reviewing the work done and progress made can sometimes be very helpful.