Blue_green in comments in the CPRS polling thread asked:

Possum- It would be interesting if you could plot a correlation between ‘against cprs’ and ‘global warming is exaggerated’ for each demographic. I wonder if they are consistent or if some just want another climate policy (just not the CPRS).

That is a mighty fine idea! The results have some pretty profound implications.

Our only small problem is that we don’t have enough observations in the Morgan poll breakdowns to look at the relationship between generic global warming scepticism and disapproval levels of the CPRS legislation for any given age group or voting block – with only 3 periods of observation for any and every demographic cohort, it wouldn’t really tell us much.

Yet, what we can do is use those various cross-tab cohorts as observations themselves in a broader look at the national relationship between the CPRS and beliefs in global warming.

Morgan asks two questions in their CPRS and global warming polling series – the data that we’ll use here. The first is “Which of the following is closest to your view about Global Warming? Do you think: Concerns are exaggerated; if we don’t act now it will be too late; it is already too late?”, while the second question is a generic “do you approve or disapprove of the CPRS legislation” type affair. We have answers to these questions for three periods of time from the same respondent sample – August 2009, November 2009 and January 2010.

What we can do is not only look at whether the change in beliefs on global warming correlates to any changes in support for the CPRS, but we can also look at the size of any such effect.

Before we start though, we need to state the obvious – correlation is not causation without good cause.

Yet here, we have pretty good reason for cause. Before we even look at the data, we would expect that as the proportion of people that believe “if we don’t act now it will be too late” on global warming increases, the proportion of people that approve of the CPRS would increase as a result.

Similarly, we would also expect that as the proportion of the population that believes global warming “concerns are exaggerated” increases, the proportion of the population that disapproves of the CPRS would likewise increase.

First up, we’ll use just the result of the age breakdowns since it covers the the full spectrum of population and has no demographic overlap. Morgan runs cohorts that are 14-17 yrs, 18-24 yrs, 25-34 yrs, 35-49 yrs and 50+ yrs. If we chart the responses of each of these groups to CPRS approval rates and “if we don’t act now it will be too late”, we get:


The relationship here is statistically significant. As we can see, as the proportion of the population that believes “if we don’t act now it will be too late” increases, it walks hand in hand with an increase in the proportion of the population that approves of the CPRS.

It does so at a level where for every 1% increase in the proportion of the public that believes in the need for action now, the proportion of the population that approves of the CPRS increases, on average, by 1.0098% – let’s just call it 1% as well. This relationship explains 67% of the variation in the change in approval levels of the CPRS.

Using these demographic cohort results to run this type of analysis is really a second best option – a few little problems could pop up that increases the uncertainty involved. Ideally we’d use respondent level data and a different type of analysis – but since no pollster in Australia provides that, we have to struggle on and make do with the polling goodness we do have.

However, what we can also do (which is also a little unusual methodologically speaking) is increase the number of observations by using overlapping demographics. At the moment we have 15 observations – 3 each for the 5 cohorts. Rather than just use the age cohorts, we can also add results by party vote, by gender and by capital city/non capital city. We wouldn’t expect that to change the actual results of the analysis much, but if we were suffering from some large sort of pseudo-ecological fallacy (one of the potential dangers in this sort of pooled cohort analysis), we would expect a significant increase in the variation of the results.

After adding the new demographics, bringing us up to 36 observations for each series, we get (click to expand the charts from now on):


The significance still holds, although we now have an outlier in the middle-bottom of the chart. That result is the proportion of Coalition voters in January 2010 that approved of the CPRS on one hand and the proportion of Coalition voters that believed in the “need for action now” on the other – so there’s a fair bit of partisan interest folded in there on one side of the equation, explaining the outlier.

Importantly, we still have a significant relationship, it’s still an approximately 1 for 1 percentage effect between the two positions, our variation hasn’t exploded and our explanatory power has remained. In practical terms, the relationship looks pretty good. Why this is important, we’ll get to at the end.

Next up, we can do exactly the same all over again, but this time with the opposite position – measuring how the changing size of global warming scepticism affects the proportion of the population that disapproves of the CPRS.

Again, just using the age cohorts we get:


This too is a statistically significant relationship, but one where there is a slight over unity effect. For every 1 % increase in the proportion of the public that believes global warming is exaggerated, we get, on average, a 1.15% increase in the proportion of the population that disapproves of the CPRS. The changes in global warming scepticism explain 66% of the changes in CPRS disapproval rates.

Throwing in the extra demographics now, we get:


There’s what looks to be 4 outliers here, 3 of which I’ve identified by greens squares and the other being the observation at the top right of the chart.

Don’t be too distressed about that, it’s a result from Greens voters (insert joke about Greens voters being outliers here! 😛  ). It stems from Greens voters having registered ultra low levels of global warming scepticism over the three periods of August 09, November 09 and January 10 at 7, 11 and 13% respectively. Not only did they have ultra low levels of scepticism, but also moderate levels of disapproval because the Greens voted down the CPRS legislation on the basis of it not going far enough to curb carbon emissions. If we remove that Green cohort and redo the chart, we get:


Now what we find, is that the 4th outlier is not actually an outlier at all, but just appeared as such when the Greens cohort was skewing the regression line at the bottom end. The chart now much more closely resembles the original chart in terms of effect.

What we now find is that for every 1% increase in the proportion of the population that believes global warming concerns are exaggerated, we would expect to see, on average, a 1.2% increase in the proportion of the public that disapproves of the CPRS.

Effectively, 80% of the change in disapproval levels of the CPRS – from it’s low level of 24% disapproval in August 09 to it’s current high of 36% this month – came from the increase in global warming scepticism that occurred over the same period.

Around nine and a half points of the 12 point growth in disapproval levels of the CPRS legislation can be explained by the from growing global warming scepticism in Australia.

In a political nutshell – Labor is losing the ground war on generic global warming opinion in this country.

This has some pretty serious implications.

For a government to ever successfully act on carbon emissions, will require them to successfully counteract the denialists. This is something that the Rudd government has so far refused to do – no doubt, at least in part, because of the softness in the ALP vote they are picking up in both the 30-49 yr male demographic (primarily, but not isolated to non-capital cities) and their recent inroads into the over 50’s. Both of those cohorts are more likely to be climate change sceptics than any other.

But if a government, this government specifically, ever wishes to build a solid majority of democratic opinion behind any carbon abatement policy, they cannot afford to vacate the field of global warming public opinion like they have – we’ve seen how the dynamics of that plays out right here in this post.

They need to carry large majorities of people with them that believe global warming is a phenomenon that requires action. Support for carbon abatement policy, even a program as second rate as the proposed CPRS, is reliant upon the proportion of the population that accepts the weight of climate science evidence – even if vicariously.

The Coalition are getting slightly disproportionate political returns on their encouragement of scepticism – getting an increase of more than 1% disapproving of the CPRS for every 1% of the population they can convince on the political backchannels to become a climate change sceptic. A wink here to Bolt, a nudge there to Jones, a deliberate public outburst from a Coalition denialist every now and then (or in Barnyards case, 18 hours a day) is paying dividends,

As long as the government continues to allow this to go unchallenged, as long as they continue to vacate the field of the generic global warming debate, they will continue to find themselves on the wrong end of the change in public opinion metrics – making carbon abatement policy an increasingly dangerous political and electoral event.

The battle for climate change policy will not be won or lost on the public battlefield of the detail of carbon abatement policy, it will be won or lost on the size of the majority that believe in the weight of evidence of climate science. It will be won or lost on the numbers of people that the government can convince to believe in the data.

If you get one, the other follows like a loyal dog.