NOTE ADDED 12/7/13: This is an old article as should be clear from the absence of the word "Rudd". Conclusions in it no longer apply because of the change of Prime Ministership, and while it is likely Labor is still behind in the seat, fresh polling is needed.
Just some quick notes about the ReachTEL poll of Bass (federal) that appeared in the Examiner today.
The findings are rather dramatic, with the Liberals' Andrew Nikolic supposedly leading Labor's Geoff Lyons 60.3-39.7 two-party preferred (based on preferences distributed as per the 2010 election), for a neat 17-point swing from the 2010 results. The Liberals are on 54.7%, otherwise known as an outright majority, Labor is on 26.7, the Greens are on 8.7, and about ten respondents believe Bob Katter's hat has been washed more recently than mine.
I think the 17-point swing is overcooking it just a little bit, which is not to say the swing won't be that big come election day. There is a very high vote for "Other" (7.9%), and in the absence of any known, specific force that would receive those "Other" votes, I believe those "Other" votes would include a number of soft Labor voters and possibly also some ultra-Green splitters. Thus I believe the preferences of these "Others" would actually favour Labor more than would be determined by using the 2010 election results, and that a more realistic reading of the swing might be, say, 15 points.
But it makes no difference. If the poll is anywhere near accurate, Labor is headed for a thumping in this seat, which could well be given away within the first hour on election night. It shouldn't be surprising either. The very limited state aggregate data available have been pointing to a low-double-digit swing in Tasmania at federal level, vague reports of internal polling have been similar, and it all makes sense given the deep unpopularity of the state Labor-Green government (which is mirrored at federal level) and given that the Coalition is bouncing back from an abysmal result in the state in 2010.
Note that it is not just " a correction to that very bad Liberal result on top of a swing against the Labor Party elsewhere in the country" (as per a part of Antony Green's interview here though the rest provides the fuller context for the swing). Even in 1996 when Labor's national federal result was about a point worse than its current national federal polling, the Liberals only got a 2PP of 54.6% in Bass. The last times we saw anything like 60:40 in the electorate federally were in 1975-7 (when the seat swung hugely to the Liberals at a famous by-election near the end of the Whitlam era) and 1983 (when it went 58.3:41.7 amid anger over Hawke's intended federal intervention in Tasmanian affairs over the Franklin Dam.)
We should not, on this basis, assume a 17% swing, or even a 15% swing, statewide (on which basis the Liberals would win all four Labor seats). There is a history of the north turning viciously against Labor, and at state level Bass is a seat where the Liberals are doing very well at Labor's expense (see Uneven Swing to Liberals in Tasmanian State Election Polling). Taking into account my reservations about the "Other" vote, this poll does provide some weak evidence that the state swing could be a bit higher than the 10.7% currently modelled in the Bludgertrack aggregator, but probably not by much if so. It doesn't change the pattern deduced from existing polling: that Bass and Braddon are very likely to fall, that Lyons and Franklin are effectively marginal and will come down to seat variations on the state swing (if it stays so high), and that Denison is complicated but probably a bridge too far for the Liberals at this stage.
A notable feature of the poll is a poor Green vote of 8.7%, a swing of 6.9% from the 2010 election (although some Green votes may be hiding in the "Other" pile as noted above.) Does this have state implications? I don't think it tells us anything we don't already know at state level. The Green vote tends to be much higher at state level compared to federal level so this poll provides no reason for believing that the Greens will poll single figures in the seat at state level. As the Uneven Swing article noted, there is a live question of whether a several-point swing will knock Kim Booth substantially below a quota, which might cause him to lose his seat to a fourth Liberal. In that article, on the basis of the last EMRS poll, I assumed a state swing against the Greens in Bass of 7.8 points, and the question of whether Booth lost his seat on that basis was unclear. Various assumptions about the state swing might be made based on this finding but there is so much room for error in the assumptions that we won't learn anything from making them.
If anything, however, I would take this poll as being more consistent with the idea that the November EMRS state poll was fairly accurate than with the idea that it was a "rogue".
Labor's best hope in Bass has long been candidate malfunction by their opponent, Brigadier Andrew Nikolic. Nikolic has been involved in two notable scrapes so far - police were involved after he heckled a protest by the activist group Code Green, and he threatened to contact the employers of people who had "liked" a Facebook post that satirised his military service. This poll shows practically nobody cares about either of these incidents (bar those who were locked in against the candidate anyway), with a measly 14.4% having an unfavourable impression of the candidate - and many of those probably doing so just because he is a Liberal. With a net satisfaction level of +24.8, Nikolic clearly has nothing to worry about from his previous skirmishes, although it might be a good idea still for him to not get in any more.
Meanwhile the incumbent's netsat of -1.9 might seem not too bad against the backdrop of a 40:60 2PP, but it would be a bit concerning for Geoff Lyons that less than a quarter of the electorate viewed him favourably and the commonest response was merely neutral.
This is just a poll, of course, and it just tells us how voters are thinking now, not necessarily what the result will be. But it does indicate that those modelling the next federal election can, for modelling assumption purposes, write off both Bass and Braddon, which should be assumed to be Coalition gains unless evidence emerges showing otherwise.
One could hardly blame federal Labor if it wanted state Labor to get its apparently inevitable election defeat out of the way now in order to stop dragging the federal vote down.