Raph Frank wrote: > > > It might also be worth discussing possible renormalisation of votes > > > here to cancel out participation bias. Also, it might be worth adding > > > in actual participation bias in the final election. > > It also has the advantage that it encourages voters turnout in the > main election. > > For example, if the last election was 4 years ago, then the turnout > estimate for the 22-25 age group might be determined (at least partly) > by the turnout for the 18-21 group in the last election (as they are 4 > years older now).
The adjusted results can then be offered as alternative views of the election. Just like we offer results that include/exclude votes based on the criterion of authentication method (open cast, private coin, etc.), we will also offer results that are/are not weighted for participation bias. The viewer can choose which view is most appropriate for her purposes. A wider range of alternative views and combinations could be offered by third parties. The design allows for provision of extended electoral services/analyses by external sites, independent of the core server. -- Michael Allan Toronto, 647-436-4521 http://zelea.com/ ---- Election-Methods mailing list - see http://electorama.com/em for list info
