It’s been a disappointing week for poll junkies, with the phone pollsters including Newspoll evidently waiting until after the Australia Day long weekend before ending their New Year hibernation. Since this is an off-week in Morgan’s fortnightly cycle, that just leaves Essential Research. All told, there have only been three poll results published so far this year two from Essential and one from Morgan so you’re more than welcome to take BludgerTrack with a bigger-than-usual grain of salt for the time being. For what it’s worth though, the one new data point has driven the Coalition to a new low of 39.3% on the primary vote, and pushed Labor’s two-party lead to a new high of 52.5-47.5.
That might seem counter-intuitive given that the one new poll had the Coalition leading 51-49, but there are three factors which have made it otherwise. First, in adjusting the pollsters for their house biases, a unique approach has been adopted for Essential Research to acknowledge that its bias is in favour of stability, rather than one party or the other. For example, Essential overshot on the Labor vote during the election campaign as momentum swung towards the Coalition, but it’s been doing the opposite since the Coalition started heading south in November. So rather than the usual method of determining bias with reference to past performance in late-campaign polls, I’m plotting a trend of Essential’s deviation from BludgerTrack so its bias adjustments change dynamically over time. With Essential stuck at 51-49 to the Coalition while other pollsters are being fairly unanimous in having Labor leading 52-48, you can pretty much work out for yourself what the Essential bias adjustment currently looks like.
The second point is to do with rounding. While Essential’s two-party result was unchanged this week, the primary vote had the Coalition down two points, Labor down one and the Greens up one. Most of the time that would mean a one-point shift to Labor on two-party preferred, but this is one of those occasions where the shift went missing after the remainders were pared away. However, BludgerTrack doesn’t actually use pollsters’ published two-party results, instead determining primary vote totals and deriving a two-party result from them using 2013 election preferences. So the Essential result looks like a slight shift to Labor compared with last week, so far as BludgerTrack is concerned. The third point is that Essential’s numbers are a two-week rolling average (though last week’s result, being the first from the year, was a sample for that week only), so any change that occurs in a given week is a bigger deal than the published numbers suggest.
So it is that BludgerTrack gives Labor a 0.5% gain on the two-party preferred projection and a boost of three on its seat tally. The state relativities haven’t changed much since last week, so the Labor seat gains are evenly spread, with one each provided by Victoria, South Australia and Tasmania. Full results as always on the sidebar.