Well done, Graham, you got the WA result right. In your view, as I understand it, the result confirms the theory, outlined in earlier OLO essays, that governments don't win, oppositions lose. That is, Geoff Gallop does not deserve any credit for good economic management and winning the vote, because it was Colin Barnett who lost the election, by running on the "wrong" issues and being a generally unattractive person.

I have expressed my scepticism about the superficiality of this theory in previous forums on the basis that it lacks any explanatory value. It might be useful to dig a little deeper than the easily recognisable "kitchen sink" issues that are usually the subject of opinion polling, and the "oppositions lose" framework that does not tell us anything useful, and try to learn something from the WA result that could be applied in the future.

For example, in a previous OLO forum I raised two factors, almost unique to WA, which I thought might affect voter opinion in that state: the first is electoral malapportionment, implemented and supported by the conservative parties, that makes WA the only state in Australia without "one-vote-one-value"; and the second is the substantial and very obvious right-wing bias in the only daily newspaper in Perth, the West Australian.

Both these factors ostensibly advantage the WA Libs, but for the first time in my memory, they were extensively covered in the mainstream national media in the two or three weeks leading up to polling day, and the commentary was predominantly negative. Could it be that Western Australians are finally fed up with the conservative bias that infects their electoral system and their local daily news, and expressed their dissatisfaction at the ballot box?

I would be interested to see any post-election polling that addresses these issues, and whether party affiliations show any correlation. Voters might be more intelligent than political parties and pollsters give them credit for. Congratulations Geoff Gallop.

Grace, I don't think I've ever advanced the idea that "governments don't win, oppositions lose", and in fact I can't think of anyone who ever has. My general approach is the reverse - Governments lose elections, oppositions don't win them. But that doesn't mean that everything an opposition or government do during an election campaign is irrelevant. In this election the government campaign was competent, and therefore not particularly interesting to write about.

I can say pretty categorically that the issues you raise are not ones that concerned voters, at least in terms of casting their votes. If they had been you would have thought that electors would raise them when asked what the issues were for them in voting, and they didn't.

The malapportionment is one that has now been embraced by the government as well as the opposition, so I would have thought, irrespective of the polling, it couldn't be an issue. The bias of the West Australian may have annoyed some voters, but I think worrying about newspaper bias is an elite preoccupation - most voters get their news from TV.

Graham - thanks for clarifying your theory. I expressed it inexactly, but I don't think precisely in reverse.

You go on to say, "I can say pretty categorically that the issues you raise are not ones that concerned voters, at least in terms of casting their votes. If they had been you would have thought that electors would raise them when asked what the issues were for them in voting, and they didn't."

Did the opinion polling to which you refer actually ask people to give a view on these particular issues? I am assuming from what you have said that your conclusions are based on "focus groups" rather than direct questioning.

However, such supposedly free-ranging group discussions are still likely to be based on what people have heard and read elsewhere, and be swayed by group consensus, so that familiar topics like economic management, leadership, health and education, and of course the issue du jour, the canal, are likely to "frame" the discussion to the exclusion of less obvious factors.

That is, even though people did not independently raise the issues of electoral malapportionment or media bias, this does not mean that they were not persuasive. Hence my interest in post-election polling, where we are more likely to find out the reasons why people voted the way they did.

Grace,

If people do not raise an issue it is highly unlikely to have been persuasive. The research to which I refer is our own qualitative research which you can access from http://www.ozelections.com. It is based on one online focus group and 136 qualitative surveys.

The sort of research you are suggesting appears to centre on asking people whether the issues that you nominate were persuasive for them. You might get a positive response, but if you did it is most likely to be because you made the issue an agenda item for the respondent, rather than it being the real issue. How could something be "persuasive" for a voter if it is not top of mind? You can't be persuaded by something you hadn't thought of.

BTW, the canal was an issue, but as a negative for Barnett. It was mentioned four times when we asked about election issues - three times negatively for Barnett. When we asked for hesitations about voting for Barnett it was mentioned 14 times. Of course it framed the discussion, but I hope you're not suggesting that politicians' promises should be ignored in terms of the discussion.

Graham - thanks for referring me to your research site, I was not aware it existed. However, as you know doubt expect I would say, I am not persuaded that one small focus group with Janet and Fred etc, and a limited questionnaire with 136 responses really tells us very much, beyond giving some sort of indicator of the result. But I am more interested in the "why" rather than the actual outcome.

You say, "How could something be "persuasive" for a voter if it is not top of mind? You can't be persuaded by something you hadn't thought of".

Well, Janet and Fred, and 136 others might not have thought of the issues I suggested, but that does not mean many others did not. I am not convinced that your sample was really representative (of opinions, not outcome), particularly given the questionnaire respondents were self-selecting, as I understand it. Further, an issue can explode into consciousness a couple of days before polling day, if given enough media attention. If your polling ran over a period of weeks, then it would have undervalued any late surge on other (new) issues.

You say, "Of course [the canal] framed the discussion, but I hope you're not suggesting that politicians' promises should be ignored in terms of the discussion." Not at all, Graham, and that is not my point. I remain interested in any post-election polling that might show why voters elected Gallop (not why voters rejected Barnett, which is clearly of more interest to you).

Here's a further query. I read somewhere else (correct me if I'm wrong, as I can't source it) that Liberal Party polling suggested Gallop was viewed as "indecisive". I am not sure what this means and how it was derived, but it is a loaded concept. Was this the reason Barnett decided to go all out with the improperly costed canal project, to make him look "decisive" by comparison. If so, it would appear to have been a major mistake by the party strategists

Grace, you don't seem to grasp how qualitative polling works versus quantitative. 136 separate responses is more than enough to get a feel for what people are thinking. In an election campaign, to be absolutely sure of the strength of feeling, and what groups it primarily affects, you would follow qual up with quantitative polling, but you can tell a lot without doing that.

If out of 136 people no-one mentions media bias, then you can be absolutely sure it wasn't an issue. They may be self-selecting, but so what? They actually select from a group - internet users - which is likely to be more attuned to this as an issue, rather than less attuned. If a reasonable percentage had mentioned it, then it would deserve further investigation, but no-one did, so you can eliminate it.

It's also unlikely that the issues would have changed significantly over the course of the campaign - they hardly ever do. Don't forget this analysis contains tacit knowledge gleaned from researching maybe ten other election campaigns over the years, as well as strategising a number.

I don't disagree with your suggestion re: the canal and Gallop's alleged indecision. In fact it was contained in my analysis before the election, and I did a blog piece on it this morning at http://ambit-gambit.nationalforum.com.au/archives/000516.html. Noel Crichton-Brown also wrote some sensible analysis of it as an issue over at Crikey!

Textor Crosby are getting the blame for it, but I'd be surprised if they would have made a mistake like that.

BTW, the reason Gallop won is because he hadn't done enough wrong to lose. The research shows that people weren't terribly enthusiastic about him, but they had no reason to switch. Gallop actually ran his campaign against Barnett, more than for himself.