Labor WON’T get a 20% Swing in NT
The Northern Territory’s only daily paper has got its facts very, very wrong, again – Labor WON’T get a 20% swing at the impending election.
This story www.ntnews.com.au/news/northern-territory/labor-on-track-for-landslide-win-in-nt-election/news-story/2f05cc0cfb5a5b000b21dca80f64b3dc says Labor is currently on track to receive a 20% swing on 27 August, Election Day. That’s hopelessly wrong, and significantly overstates the swing to Labor.
Recent Electoral Act changes are the biggest reason the reported poll results are wrong. The biggest change is from compulsory preferential voting, to optional. The Country-Liberal Party (CLP) and some independents will run a “just vote one” campaign, and to the extent this campaign is successful it will mean Labor has to rely far more on its primary vote for the anti-Government swing.
In this poll, Labor’s primary vote is LESS than at the last election.
As is sometimes the case in the Northern Territory, there may be stooge candidates run, to reduce the prospects of a primary protest (anti-CLP) vote going to Labor.
Second, other Electoral Act amendments ban campaigning within 100 metres of a polling place. This change has the impact of reducing the capacity of voters to vote for the Party/candidate of their true choice, because it increases the prospects of their not knowing who candidates are – will they mark their ballot paper at all, or donkey vote instead? There are other changes, too, which can be found here http://blogs.abc.net.au/antonygreen/northern_territory_elections/ and here www.ntec.nt.gov.au/MediaAndPublications/Documents/Election%20Newsletter%20No.%2003-%2009%20March%202016.pdf. No-one can know how this will impact voting overall.
Third, it is a 887 sample, but it is a mostly (entirely?) urban/rural residential sample, done by phone. The margin of error is said to be plus or minus 3%. The results are weighted demographically, but it’s almost impossible to poll many aboriginal town camps or remote indigenous communities, so they must be under-represented, or not included. It really says nothing about the vote in such locations, and in the NT town camps and remote communities are important determinants of results in individual seats.
Fourth, the report draws conclusions about different parts of the Northern Territory, but the sample size for those geographic and demographic segments is much smaller than 887, so the margin of error will be significantly larger.
Fifth, this reporting of the poll impacts voter intention, and will have multiple outcomes: reducing the enthusiasm for a protest vote amongst those who are nervous about a Labor Government, reducing the enthusiasm for those who want a change in Government to bother voting, increasing the commitment of pro-CLP voters to make the effort, and generating complacency amongst Labor candidates and campaigners. There’s also alleged to be a bandwagon effect, of people backing a perceived winner, but I’ve seen no research suggesting it is significant in state- or national-level elections. Some voters do change their vote according to expectations.
Journalists often sensationalise and misrepresent, or simply misunderstand, polling, but this piece by the NT News is worse than most.
Posted on August 4, 2016, in campaigning, Chief Minister, Democracy, Election, Michael Gunner, Northern Territory, Opposition Leader, Voting and tagged journalism, market research, media, news reporting, Northern Territory Election, NT News, polling. Bookmark the permalink. Leave a comment.