Propaganda failure

The WorkChoices campaign – for and against – must be the most expensive in Australian political history. But how effective was it?

Last month I argued that perhaps the government’s ‘fairness’ test change and subsequent advertising helped ease concerns among Liberal voters. But the polling data I have been analysing over the last week for a Policy article I’ve been writing on WorkChoices tells us something, I think, about the limits of political propaganda.

All along, the polling on for or against questions about WorkChoices has been stable. Three Morgan Polls between July 2005 and April 2006 found the proportion of voters against the reforms varied by just 0.8%, after taking out those not expressing a view. Just under three-quarters of Morgan respondents with an opinion were against WorkChoices. In five ACNielsen polls, of those offering a for or against opinion, the proportion against varied from 69% to 74%, a very similar result to Morgan.

This suggest that the outline of WorkChoices triggered reactions based on stable aspects of public opinion, and nothing anyone said or did after that changed the basic yes/no position of the electorate.

On more specific aspects of WorkChoices, we do see opinion changing. Two of the three pollsters that asked about the reform’s personal impact found that fewer people over time perceived negative effects, with no clear trend on positive effects. The questions asked as part of Sensis consumer research were most interesting, because they have two surveys after the PM’s backdown on the ‘fairness’ test. That took the ‘negative’ group below 30%, when it had been mid-30s in most surveys. ACNielsen found a steady drop in those thinking that they would be ‘worse off’ in three polls between October 2005 and March 2007, from 31% to 21%.

Newspoll was the exception here – it found the ‘worse off’ group stable (moving within the margin of error) between October 2005 and March 2007. This is perhaps because it only asked people with jobs this question.

Overall, it suggests that the ACTU scare campaign, based on melodramatic scenes of ‘unfairly’ treated workers, didn’t scare anyone who wasn’t already scared, and did not stop some Sensis and ACNielsen respondents realising that they had nothing to fear after all. Except for the very first Sensis survey, ‘no real impact’ or ‘no difference’ was always the most popular answer.

A key message in the government campaign was that WorkChoices would be good for jobs. Like the ACTU’s personal risk campaign, this message failed to persuade. Indeed, the Newspoll question tracking this issues shows that those thinking WorkChoices would be bad for jobs rose from 39% in late 2005 to 48% early this year.

In my article, this is the result I am most struggling to explain. Though public understanding of economics is poor, the coincidence equals causation logic of DIY economics would still, on the basis of very strong jobs figures, point to WorkChoices not being bad for jobs, and perhaps even being good for jobs. While WorkChoices’ opponents were trying to explain away jobs growth by attributing it to global economic conditions, I don’t think that they were even seriously arguing that it would be bad for the number of jobs. How did the public come to a conclusion that was hardly put in the public domain?

The only explanation I can think of is that they extrapolated from greater power to sack that there would be net job losses. Perhaps employees know of large-scale feather-bedding that they think would come to an end, but it’s hard to believe that this practice is now common, except perhaps in parts of the public sector.

Whatever the explanation of that result, WorkChoices is a case study in the limits of opinion manipulation. Except perhaps in the last two Sensis surveys, there is no evidence that any of the advertising had a net positive effect (and there it may reflect objective changes) for the advertiser, and both the ACTU and the government suffered public opinion setbacks on key messages.