Essential Report redux and the perils of push polling

The pdf file of the most recent Essential Report has now appeared on their site and it’s in full working order, allowing us to go through it properly.These questions ran off a sample of 1028, giving us an MoE that maxes out around the 3.1% mark.

The first question has been asked twice previously, giving us a nice series that demonstrates how opinion has been changing here.

If the Opposition continues to refuse to pass the Government’s legislation for the Emissions Trading Scheme to address climate change, do you think Prime Minister Kevin Rudd would be justified in calling an early election?

What a difference 4 months makes!  I was surprised by the April result because Australians have never been particularly keen on elections, but even if we assume that the April numbers were slightly overcooked on the YES side, the trend away from the CPRS legislation justifying an early election is substantial.

I suppose the biggest question is why? From June to last weekend, there’s an argument that because the renewable energy component got the go ahead, that satisfied the small numbers of people that were likely to have changed their view over that period (I say likely, because the change between the June and August periods was not statistically significant). Yet the substantial jump between April and June suggests a more generic change in public opinion was afoot.

On the cross-tabs, Essential said:

Labor and Green voters were more likely to support the calling of an early election if the Opposition continues to block the Emission Trading Scheme legislation (42% Labor and 42% Green), while Coalition voters were more likely to oppose an early election (54%).

That doesn’t really help us much either here, looking more like partisan self-interest than anything else.

Over the next 12 months do you think economic conditions in Australia will get better, get worse or stay much the same?

(click to expand)

The post-budget period of May-June really changed the trajectory of economic expectations – something we’ve also witnessed with Newspoll Standard of Living poll as well as the more orthodox consumer confidence surveys that litter the landscape of the nations business pages. If we break those totals down into their respective components, we see that most of the change has been from the “little worse” to “little better” column suggesting that we aren’t ecstatic about the outlook, but have changed from marginally gloomy to marginally optimistic.

The cross-tabs repeat the same pattern we regularly see here on questions of economic outlook, where supporters of the government of the day are generally more optimistic than Opposition supporters:

Labor voters were more likely to think that economic conditions in Australia will get better over the next 12 months (64%), while Green and voters were more likely to think they will get worse (31% Green, 28% Coalition).

People aged 55 years and over were more likely to think economic conditions will get better (61%), while those aged 35 – 44 years were more likely to think conditions will get worse over the next 12 months (31%).

Essential asked a further follow-up question on the economy focused on just the nature of the slowdown.

Do you think the Australian economy is over the worst of the slowdown resulting from the global financial crisis and is starting to improve or do you think it will still get worse before it starts to improve?

Unsurprisingly, the “Will get worse” response comes in identical to the “Total Worse” response in the first question, with some of the “Stay the Same”ers moving across into the optimistic column. The cross-tabs repeated the same partisan pattern:

Labor voters were more likely to think that the Australian economy is starting to improve (71%), while Coalition voters were more likely to think that things will get worse before it starts to improve (28%).

Which of the following do you think should be Australia’s next Ambassador to the USA?

The cross-tabs said:

Coalition voters were most likely to support Downer being appointed as Australia’s Ambassador to the USA (28%) and Costello (20%), while Labor voters were more likely to support a career diplomat (22%) or Kim Beazley (18%). Green voters were also more likely to support Beazley (20%).

Respondents aged 55 years and over were more likely to support Alexander Downer (28%), while those aged 45 – 54 were more likely to support Kim Beazley (18%).

South Australian voters were more likely to support Downer (20%), Victorian voters were more likely to support Costello (17%) and Western Australian voters were more likely to support Beazley (23%). Only 2% of NSW voters support Carr as Australia’s next Ambassador to the USA.

Some people say that the Liberal Party is so divided and without direction at the moment that regardless of what you think of the Rudd Labor Government, the Liberals are just not prepared at the moment to take on the difficult task of governing Australia. Do you agree a lot, a little or disagree a lot or a little with this statement?

The cross-tabs said:

Labor voters were more likely to agree with the position that the Liberal Party is not prepared at the moment for the task of governing Australia, while Coalition voters were more likely to disagree with this statement (53%).

However, 36% of Coalition voters agreed with the statement that suggests the Liberal Party is too divided and without direction that it is not prepared for the task of governing Australia.

Respondents aged 45 – 54 were more likely to agree with the position that the Liberal Party is not prepared for the task of governing Australia (78%), while respondents aged 55 years and over were more likely to disagree with this position (31%).

The issue of push polling was raised with this question – but there’s a really big misconception over what a “push poll”  is. A “push poll” isnt actually a real poll, it’s a piece of deliberate political propaganda and marketing where the facade of “being a survey” is what’s used to hook respondents into listening and being fed the propaganda. The purpose of a push poll isn’t to elicit a response from the survey respondent, it’s sole purpose is to feed (usually) negative information – usually some smear – into the minds of listeners.

“Push polling” is an exercise in marketing rather than polling. The way a usual “push poll” works – which let it be said, is a pretty rare thing in Australian politics – is that large numbers of calls are made to some constituency that have been identified as susceptible to some particular issue or set of values. The push pollster pretends to be a fair dinkum market research organisation looking for answers to questions, but instead of asking a rather long set of questions (which is what proper surveys usually do), they ask a short set of questions usually framed like “Would you be more or less likely to vote for the Labor Party if you knew that they were going to abolish funding for all private schools“.

The answer never gets recorded,  the point of the exercise isn’t research – it’s rumour spreading. That’s a push poll.

Where the lines start to get blurry is when political parties and their pollsters launch a round of “message testing” on a population larger than a focus group. Focus groups are generally small – half a dozen to 25 or so people in a room getting psychologically probed. From focus group work it might become apparent that a political opponent – let’s say Malcolm Turnbull – has regular themes said about him by the focus group members. For instance, the people in such focus groups might regularly suggest that they are concerned about Turnbull’s political “judgement” .

If the Labor Party thought that line might work as an angle of political attack, but they weren’t sure – they could ask their pollsters to essentially expand their focus group work into a few hundred people via telephone polling. So the pollsters might go out and ask some innocuous political question followed by something  like “Moving on now to the leader of the Opposition, Malcolm Turnbull.Do you believe that Malcolm Turnbulls political judgement is very good, good, neither good nor poor, poor or very poor?“. They might then follow up with another question asking “Does Malcolm Turnbull’s judgement make you more likely to vote for the Coalition, less likely to vote for the Coalition or does it make no difference on the likelihood of you voting for the Coalition?“.

Message testing like that can verge very close to push polling – but the difference is always in the intent. If you are trying to change someone’s opinion by the exercise itself, it’s push polling. If you are trying to ask someone’s opinion to see whether a political angle resonates,  it’s message testing.

But questions like the one above that Essential Media asked are quite simply neither.

Sometimes questions can’t be perfectly balanced. If a pollster is interested in finding out how perceptions of Coalition disunity are affecting their political support level – there’s a number of ways to try and reduce or minimise any leading of the respondent into a given answer, but because the topic itself is relatively negative for the Coalition, it’s unavoidable that the question might lead to a slight theoretical response bias. Good pollsters will minimise as much as is humanly possible any theoretical endogenous response bias caused by the topic itself – but often it can never be theoretically removed completely.

The way Essential asked the above question is perfectly valid – they were asking about a negative issue for the Coalition, but it wasnt a “have you stopped beating your wife” type catch 22. There were responses available for the full spectrum of beliefs.