This question was posted by an ACE user through the Ask the ACE Experts function.

The Question
This question was posted by an ACE user through the Ask the ACE Experts function.
I am looking for some bright ideas, and I am hoping you can help me.  I am writing from Toronto, Canada.  We are about to have a mayoral election, and a group I am involved with are organizing the first series of candidates' debates.

Our dilemma at this point, is that there are 18 declared and registered candidates, too many to include them all in a coherent debate. Traditionally it is left up to the editors of major media to nominate who gets to participate, but we are hoping to find a more accountable method of
selection. We are also considering using the results of polling data, and inviting those who appear to be the leaders.
I am wondering if you are able to suggest some alternate method of selecting debate participants that will be reasonable, fair and transparent.

Summary of responses
Both replies suggest hosting two or more debates, to ensure participation from all candidates in at least one debate. This would, however, raise further questions such as time and money constraints, and who should participate in which debate.

It is pointed out that using polling data to select only the most popular candidates can be a fair and transparent method, especially if an average of several polls is taken. This might, however, discriminate against less wealthy candidates with little access to the media.

Other suggestions include a random selection of candidates, or only inviting candidates representing parties with seats in the Council.
One of the experts warns that there is no completely trouble-free way of filtering participates for a debate.

Examples of related ACE Articles and Resources
Encyclopaedia:
• Media and Elections 
• Media in Election Campaigns  

Comparative Data:
• Media and Elections, Are televised debates between candidates or party representatives normally conducted?

Names of contributors
1. Kevin Evans, John Dunn and T.S. Krishna Murthy (joint reply on behalf of the Malaysian Centre for Commonwealth Studies)
2. Hernan Bonomo

On behalf of the Malaysian Centre for Commonwealth Studies:

Beyond selecting participants on the basis of current standing in the polls (which seems quite impartial and transparent - especially if the approach used is to average the various polling figures available, rather than relying on any one individual poll), the following suggestions could be appropriate: 

• restrict candidates asked to participate to those who are the endorsed candidates from parties with seats in the Council. The obvious downside of this approach would be the existence of a candidate without party support but who has a significant following (according to the polls).
• Perform a random choice of participants, or a set of debates, with the participants randomly distributed between them.
• Arrange two separate sessions of nine each. Cut short the duration if time and/or money are constraints.

Replies received from

Kevin Evans (Tiri)
John Dunn (Trinity College, Cambridge University)
T.S.Krishna Murthy (previously Election Commission of India)

I think that any filter for participation you may apply will put you in trouble. It depends on the purpose of the debates to decide whether and which filters are appropriate. If you decide to invite those that are polling on top, you may be discriminating against the poorer candidates, since they are usually less visible on media. If you'd like to provide a space for less visible candidates to discuss with those from wealthier campaigns, then you should try to make two or three debates, combining candidates, arranged so that candidates with little chance can also debate front-runners.

Thanks to Kevin Evans, John Dunn, T.S Krishna Murthy and Hernan Bonomo for your replies. It seems as the following alternatives are among the most popular suggestions: 

1. Include all candidates by hosting 2 (or more) debates with 9 (or less) participants each. This would include all participants but potentially cut the debate(s) short due to time or money constraints. It also raises the question of which candidates that should participate in each debate?

2. Include the most popular candidates according to poll results. A recommendation here was to use several poll results rather than a single poll. This might however discriminate against less wealthy candidates with little access to the media.

Other suggestions included a random selection of candidates or only inviting candidates representing parties with seats in the Council. Any opinions on these alternatives? Are there other ways of selecting candidates?