Dr Kevin Bonham
They say even a week can be a long time in politics. When I last wrote for TT, Paul Lennon was still Premier, Peg Putt was still leading the Greens, and the global financial turmoil of September-October was still months in the future. Some things, however, change more slowly. We still have a state Labor government that presides over an at least superficially much sounder economy than Tasmania is historically used to, but that government is still struggling to avoid a seemingly endless procession of probity and governance concerns and arguments in the popular media.
IN MAY this year I accepted a short-term position that, while completely apolitical in nature, was nonetheless covered by conditions concerning public comment that could affect my comments on a range of issues. Rather than punch with one hand behind my back in certain exchanges on this site, or punch unsure of whether that hand needed to be there or whether I was even allowed to punch at all, I simply quit posting on this site until that contract was finished. Now I’ve moved on to other things and I am free to post here again (although I have been so busy with those other things it has taken me another month to do so.)

They say even a week can be a long time in politics. When I last wrote for TT, Paul Lennon was still Premier, Peg Putt was still leading the Greens, and the global financial turmoil of September-October was still months in the future. Some things, however, change more slowly. We still have a state Labor government that presides over an at least superficially much sounder economy than Tasmania is historically used to, but that government is still struggling to avoid a seemingly endless procession of probity and governance concerns and arguments in the popular media.

And some things, alas, never change. In trying to make sense of the impact of factors both local and global upon little Tasmania’s future political landscape we have little to go on but for EMRS polls that record such a high undecided rate as to be only slightly more reliable than reading the entrails of a pademelon. Worse still, by the time these results are interpreted by the more excitable scribes of the popular media, any similarity between what is served up for public consumption and the likely electoral realities has well and truly exited the building.

So firstly: the poll. The full details are at http://www.emrs.com.au/pdfs/State%20Voting%20Intentions%20November%202008%20Report.pdf. The November EMRS poll shows support figures of 30, 26, 18 and 1% for Labor, Liberal, Green and Independent respectively, with a massive (though not so massive by EMRS standards) 23% supposedly undecided. The Liberals are down four points since August and the Greens up two. The preferred premier poll shows David Bartlett down three points to 37, Will Hodgman down four points to 29, and Nick McKim up three points to 15, with uncertain/someone else up five to 20. The sample size is 1000 voters, presumably evenly distributed between electorates.

There is a slightly curious comment concerning the trends in the leaders’ ratings. According to EMRS, “Support for Nick McKim as premier has continued to increase”, but leaving aside the question of statistical significance, EMRS actually have only two known readings of Nick McKim’s preferred premier rating to comment on (12% in August 2008 and 15% now.) Before August, they had last polled preferred premier ratings in June, at which time Peg Putt (8%) was still the Green leader.

I have mentioned before (see http://oldtt.pixelkey.biz/index.php?/weblog/article/poll-worse-for-lennon-than-for-labor/) that EMRS polls are of limited use for two main reasons: firstly, between elections, they record many voters as undecided who past evidence suggests are mostly actually Labor voters. Secondly, they overstate the Green vote. Things can sometimes usefully be made of them by comparing from poll to poll, but this itself is difficult because the changes from poll to poll are not always statistically significant, and even where they are, they may simply reflect good or bad news cycles at the time of polls, rather than long-term changes. As shown by Peter Tucker in his November 11 post at http://www.tasmanianpolitics.blogspot.com/ there is, however, a long-term trend of roughly linear decline in Labor’s EMRS vote.

The Examiner, the Mercury and the ABC all reported the poll on 11 November. The Examiner’s report simply summarises the results of the poll as provided by EMRS. It is commendably factual although a warning note about the lack of statistical significance of some findings would have improved it. The ABC’s report (http://www.abc.net.au/news/stories/2008/11/10/2415739.htm) makes a little too much of the three points lost by David Bartlett and the two points gained by the Greens. Neither change is statistically significant.

The Mercury’s report by “Sue Neales Chief Reporter” is quite extraordinary in what it attempts to read in to one simple opinion poll. This report (http://www.themercury.com.au/article/2008/11/11/37635_tasmania-news.html) contains one elementary blunder and three speculations worth discussing in details.

Firstly, the blunder … but Neales is not at all alone as this; this particular irritation is increasingly common and hardly any mainstream media source appears immune to it. Neales repeatedly refers to the rates recorded for particular leaders as their “approval rating”. But they are not – the figures discussed are in fact the Preferred Premier score. An approval rating is the result a polling firm obtains by asking a question like “Do you approve of the way Bill Bloggs is doing his job as dictator of Tumbolia?” The difference is one of the most crucial differences in psephology, for two reasons. The first is that it is possible (and not at all uncommon) for a voter to prefer one leader over another while approving of the performance of both, or disapproving of the performance of both. It is even possible for a voter to think, for instance, that A is doing a bad job of being Premier, B is doing a good job of leading the Opposition, but A would still be a better Premier than B.

The second is that the preferred premier score is little but a political “beauty contest” that has been shown to be extremely unreliable as a predictor of election results and changes in party popularity. This is quite different from the true approval rating, which is far more informative and which in some cases can be a useful advance predictor of party support (as John Howard’s waning personal approval apparently anticipated part of the Liberals’ decline, most likely causatively, in the leadup to the 2007 federal election.) It would be more useful if EMRS polled a real approval rating rather than a preferred premier score.

On to the speculations. The first is that David Bartlett is still in a honeymoon period. If this is true then for David Bartlett to hold an eight-point lead as preferred premier over Will Hodgman, and for his party to lead the Liberals by (supposedly) just four points, is a position that is artificially inflated by his newness to the job. Neales provides no evidence that this would be the case, and indeed there is none to find; it could be that this relative state of affairs will be business as normal for the party under Bartlett and hence the “honeymoon” (such as it was) has already ended.

Of course, David Bartlett is clearly better placed than Paul Lennon was in May, but that is no evidence of a continuing “honeymoon” with voters given the kinds of results Paul Lennon was returning near the end of his premiership. So if this is really still a honeymoon as Neales contends, then it must be that Neales sees business as normal to be something like level-pegging between the major parties on both the beauty contest score and the party score, although such an outcome would actually be among the worst opinion poll results for Labor over the last decade.

The “honeymoon” idea thus disposed of, I shall now move on to Neales’ contention that this poll shows that the Liberal Opposition led by Hodgman junior is “in trouble”. There may be rather more to this one, but you’ll have to look outside the Mercury to find it, since Neales’ argument is based on Will Hodgman losing ten points in six months, and his party losing seven, since the exit of Paul Lennon. At least here we are dealing with declines that a statistician would actually get out of bed for, but the decline is in comparison to a time when Labor was still led by an unpopular premier approaching the end of his leadership. The Liberals could hardly be expected to have it as good now in the polls as they did then, so it is no real surprise that they do not.

There is also the question of what should really count as the Liberals being in “trouble”. For the Liberals to actually win the next election outright would be remarkable, since only once in the last 100 years (1982) have Tasmanians replaced a majority government of one party with a majority government of the other. Thus, falling short of that aim is hardly “trouble” for the Liberals. Even failing to win the most seats would be no disaster if the government lost its majority, since the last two Tasmanian governments to depend on Green support have been defeated outright at the following election. “Trouble” for the Libs is a position in which there is a risk of a Labor majority government being returned.

What could spell big trouble for them is if the “undecided” voters in the EMRS poll are overwhelmingly soft Labor voters who will return to the fold come election day. In my Feb 2007 article that I linked to above, I used Morgan polls to suggest that as many as 83% of voters reported as “undecided” by EMRS even when asked which party they were leaning to, would in fact vote Labor. Indeed, in 2006 Labor polled 49.3% at an election despite polling 32% only two EMRS polls before it. If the same corrections I applied there are applied to the November 2008 data, then Labor’s four-point lead becomes a 15-point lead (47-32-18) that with a uniform statewide distribution of swing would probably result in an outright victory for the ALP, losing just one seat in Franklin.

I am reluctant to do this, though, because I would like more Morgan data to test whether the undecideds of late 2008 are the same sort of creatures as the undecideds of 2006 and early 2007. I cannot simply assume that they are, and as Peter Tucker shows at http://www.tasmanianpolitics.blogspot.com/ , even if the undecideds split 50-25-10-15 (ALP-Lib-Green-other) that puts Labor in a very difficult position in terms of retaining its majority. And sadly, Morgan have not polled Tasmania at state level for a very long time.

What we can say is that the current state of play between the major parties is almost exactly the same as it was in May 2007 (when the Government also led by four points) except that the Government no longer faces any known looming leadership issue, and the next election is much closer. That in itself sounds not too rosy for the Liberals, and probably isn’t, but Labor have their own problems, particularly in Franklin where with one big vote-getter gone and another incumbent in political limbo, finding the profile to hold three seats in any circumstances will be extremely difficult.

The third Neales speculation is that Nick McKim is a “big winner” from the poll on account of the Greens being up two points and his own preferred premier score being up three. As mentioned above, for the Greens to be up from 16 to 18 is not statistically significant. Furthermore, it is not unusual for the Greens to poll 18 in EMRS polls; indeed, it is the third time in four polls that they have done so this year, albeit with a slightly higher undecided result than before. The difference between Nick’s beauty contest score of 15 points and Peg Putt’s final reading of 8 certainly does mean something, but the latter reading was recorded when David Bartlett really was in a honeymoon period, while Putt’s confrontational style was being seen as irrelevant following the departure of her old enemy Lennon. Nick McKim’s preferred premier position now is similar to Peg Putt’s before the 2006 election. This is not a bad result after only a short time in the job (do Greens Leaders get electoral honeymoons too?) but it would be much more useful to see his actual approval rating, which I’m sure is a great deal higher than 15%, and see from that whether voters have yet formed a view on the progress of his leadership. All we can say from the current poll is that he does not appear to be struggling in the job.

It is still too early to speculate too much about the 2010 state election. Those closely following the US Presidential Election will note how world economic events (with plenty of assistance from a pretty poor Republican campaign) created a national swing of around three points in the two months leading up to the election, turning what looked like a close contest into a decisive victory for Obama. Tasmania currently has the second-lowest unemployment in the nation, and it is unclear that voters will end the majority rule of a government presiding over such conditions no matter how many governance and personnel scandals affect it. But the position of a long-term incumbent government in times of economic instability is always a difficult one, and if the downturn bites in a way that severely affects Tasmanians they are likely to take it out on the party, to at least the extent of returning a hung parliament. On the other hand, if Tasmania gets through the mess more or less unscathed, we may see the government successfully claiming credit. I strongly suspect that the condition of the Tasmanian economy, as it impacts upon Tasmanians generally, will be the main determinant of the 2010 election result.

Another reason it is a little too early to speculate seriously is that we don’t yet know what the playing field will look like. While there appears no serious prospect of a change in the size of the House anymore, there is still the question of the AEC’s current redistribution process. This is expected to conclude in February 2009 and at that time, based on the new boundaries that the AEC has determined, I intend to provide a lengthy analysis of the kinds of swings required to deliver various results, and where these would sit in Tasmanian electoral history. That is, of course, if my employment conditions at the time allow me to do so!


Kevin Bonham is an invertebrate research consultant for a variety of employers, a chess tournament director and an electoral analyst/commentator and scrutineer. His doubtless essential endorsement of the winning Obama/Biden ticket occurred in mid-September in disgust at the illiberal policies of McCain’s running mate.