EMRS Poll – bad for Labor, cautiously good for the Libs

The latest EMRS poll shows a massive drop in support for the Lennon Labor government. Psephologist Peter Tucker won’t quite jump to the conclusion that Labor is irrevocably stuffed, but cautiously believes the Liberal can take heart.

THE latest EMRS poll reported 12 December 2006 in the Mercury http://www.news.com.au/mercury/story/0,22884,20912204-3462,00.html and the Examiner (no link) is not good news for the Labor government.

Opinion polls in Tasmania, since Jim Bacon’s spectacular defeat of Bob Cheek’s Liberals in 2002, have generally shown Labor support at around 50 per cent. There was a period leading up to the March 2006 election where Labor slumped to the low forties, but recovered again to post another fifty percenter at the polling booth.

Since then, surveys by both EMRS and Morgan has generally shown the government maintaining close to this level of support.

So what can we draw from this poll?

First, be very wary of making predictions from one poll. Too many political commentators and journalists like to link every move in an opinion poll with whatever stories are making the news that week (Dennis Shanahan at the Australian, in his weekly divining of the national Newspoll, is a serial offender). This view of public opinion assumes voters continually adjust their intentions, one day leaning this way, the next the other.

Earth to journalists: the electorate is just not that volatile or fickle. Public opinion tends to be formed cumulatively and incrementally.

For example, analysis in the Mercury on this EMRS poll suggests voters have ditched the government because of the TCC “scandal”. Maybe, but if voters are so ticked-off over TCC why wait until now to show their displeasure? The story has been around for months. And why not blame health or Hawthorn or whatever else is going on? The fact is that there have been bad news stories surrounding this government for years. Allegations of sleaze, incompetence and scandal could not have been thicker on the ground leading up to the last election, yet Labor sailed through scarcely grazed. Perhaps we have now reached the tipping point, but that conclusion should not be drawn from this poll alone.

Second, keeping the above qualification firmly in mind, if the EMRS poll is an accurate reflection of voting intentions, I think the Liberals can start to take serious heart from the outcome. This has been missed in the newspaper commentary because of the 29 per cent Liberal primary vote, but a closer analysis I think reveals a truer picture.

EMRS polls always have a high level of undecideds.  This is a function of the way Hocking conducts his polls.  So the undecideds need to be “rolled back” into the survey, assuming they will distribute the same way as the decided votes.  Doing this reveals the following, compared with the March 2006 election primary vote:

The government has taken a significant hit.  The Libs and the Greens have posted modest gains.  We know that the Greens and “others” seem to do better in opinion polls than at elections, so we could take a couple of percent off each and add that back to the major parties.

Of course, making the assumption that undecided votes will fall the way of the decided votes is a leap of faith. The fact that there are so many undecided voters means that all parties have an opportunity to press their claims and make serious headway.

But, overall from the information we have, it is not unreasonable to generalize that Labor is down about 10 percentage points, while the Liberals are heading, on a three party preferred basis, to the mid 30s.

As I have said before, there are two key requirements for a change of government at election time.  Both have to be present.  (1) voters have to be in the mood to change government, and (2) the opposition have to be seen as a safe and capable alternative.  (1) appears to be in place (although there is time for the government to reclaim voter confidence) while the electorate is still not sure about (2).  The challenge for Will Hodgman and his team is to keep the pressure on the government while cementing in the electorate’s mind that they can govern if given the opportunity. Those undecided voters are there to be convinced.

The next Morgan (http://www.roymorgan.com.au/) poll is due out in early January. Morgan survey a similar number to EMRS (about 1,000) but stretch the information gathering over a period of months. This methodology means less volatile results between surveys, but does not allow for EMRS’s “snapshot” outcome.

In an item for Tasmanian Times in September http://oldtt.pixelkey.biz/index.php/weblog/comments/can-libs-turn-tables/, commenting on polling then, I said this in response to a reader query:

I would not be worried about the opinion polls at the moment.  The trend in polls (here and in other states) is for governments to retain their popularity after an election, slump mid term, then pick up again as the next election approaches.  Voters need time to assess a new leader.  Tasmanian Labor WILL dip in the polls sometime over the next two years.  The voters WILL poke their heads up and take a long, hard look at the alternative.  If that alternative looks OK and doesn’t frighten the children, they might just say “you’ll do” come the 2010 election.

Opinion polls are much more useful and valid when looked at in aggregate and in trend terms over time. Before jumping to any conclusion that Labor is stuffed, I would want to see what a series of polls have to say. The Morgan result in a few weeks should be interesting holiday reading.

Peter Tucker has worked in Tasmania as an advisor for the Liberals in opposition and in ministerial offices for both Labor and Liberal governments.  He is currently a free-lance political commentator and PhD candidate at the University of Tasmania’s School of Government.  http://www.utas.edu.au/government/postgrad/PhD/PT.html