Filed under: Canadian Politics, - Liberal Party, - - 2006 Leadership
METHODOLOGY
The projections (see SUMMARY) use four sets of data, each weighted equally:
1. POLLING DATA (25%)
2. PROVINCIAL ANALYSIS OF CONTRIBUTIONS (25%)
3. PROVINCIAL ANALYSIS OF MEMBERSHIP SIGN-UPS (25%)
4. NUMBER OF DELEGATES RUNNING BY REGION (25%)
1. POLLING DATA (25%)
Three polls provided information about Liberal members’ voting intentions: Sep 12 Gandalf poll (national), Sep 19 Strategic Counsel (national and regional), Sep 20 EKOS (Ontario and Quebec). Within the 25% weighted for polling data, half was weighted for national polls (i.e. 12.5% overall weight) and half was weighted for regional polls (i.e. 12.5% overall). National polls were weighted proportional to population, to give an overall % for each candidate per region. National poll data averages the Sep 12 Gandalf and Sep 19 Strategic Counsel polls (both weighted equally because they both had sample sizes of 1000). Regional poll data for Ontario and Quebec averages the Sep 19 Strategic Counsel and Sep 20 EKOS polls (EKOS was weighted 2/3 and Strategic Counsel 1/3 because EKOS sample size was twice that of Strategic Counsel). Regional poll data for B.C. and the Prairies equals the Sep 19 Strategic Counsel poll numbers (since the sample size is smaller, this data is less reliable). No regional polling data was available for Atlantic Canada/North, so these numbers were estimated. NOTE: only decided voters were included.
(remainer will be uploaded soon)
If your comment doesn't appear, it is because our automatic anti-spam software is blocking it. If so, just send us an email and we will post it for you.
