An opinion poll is a survey of public opinion from a particular sample.
The sample and questions are designed to indicate the opinion of a larger group, for example the nation.
Some general comments on opinion polling and opinion polling in relation to constitutional change follow these initial comments.
In a nation obsessed at the political and media level in polling, it is worth at this point introducing some cynicism about polling.
The nation’s eminent psephologist, Malcolm Mackerras, once read out a definition handed to him by an ACM supporter.
It went something like this: “An opinion poll consists of the answers of those willing to respond to uninvited questions put without notice on matters on which the respondents have not had the time to consider.”
...from intial polling to the actual vote...
Before we come to our conclusions on polling on a politicians' republic, we should bear in mind that polls taken before a debate on a referendum proposal will normally record significantly support than during the referndum.
The trend line indicates that support for a vague undefined republic is at the time of writing, as a percentage, only in the low forties.
Because the people will have the opportunity to hear both sides, it is likely to fall even further at the actual vote.
This happened in 1999 even with a highly biassed mainline media and a wealthy Yes campign supported by twothirds of the politicians.
This will be exacerbated by the precise question which must introduce a model. Many hitherto Yes voters opposing the model chosen will then prefer the constiutional monarchy.
This is the reason why republicans prefer an intial plebiscite or plebiscites. They are even divided on the number of plebiscites.
At the present time polling and other evidence suggests fifteen conclusions:
1. Since the 1999 republic referendum, there has been a long term decline in support for a vague undefined ( politicians’) republic, currently between 35% to 48% by one poll which is consistently higher than the others;
2. Polling continues to indicate a bell shaped curve revealing lower support among the young and continuing strong opposition among the aged. In the latest poll, the Morgan Poll in 2011, support for a politicians' republic among the young was at 31% and among new immigrants (28%);
3. Support is strongest among inner city voters especially middle aged males and supporters of the Greens;
4. Once a republican model is announced as the preferred republic, the Condorcet principle espoused by psephologist Malcolm Mackerras applies and support falls further ( that is a significant number of republicans always prefer the constitutional monarchy over the opposing model);
5. Interest in republican change is generally weak. Those who describe themselves as strong supporters were, according to the April 2011 Newpoll, down to 25%. Among the young this was 20%. The contrasting experiences of ACM and the republicans in calling for public demonstrations supports a conclusion that many more monarchists are strong supporters of their cause than republicans.
6. The latest poll on direct election ( by Morgan polling) indicates no greater support for this than there is for a vague undefined republic;
7. As with any other polling, occasionally a "rogue" poll going against the trend will emerge, as with the 2009 UMR poll released at the time of the tenth anniversary of the referendum;
8. Another referendum on the 1999 model would be overwhelmingly defeated and a referendum on a model involving the direct election of a President would also be defeated ( republican Professor Craven says the defeat would be greater than in 1999);
9. A referendum delaying change until the end the reign would be defeated overwhelmingly;
10. If a plebiscite were to be held, it will be weighted in favour of a Yes vote. This will be done carefully designing the question. This will be done by taxpayer funded specialists, aided by substantial taxpayer funding including provision for “education” and “information”, probably little or no public funding for the No case, possibly no Yes/No booklet, and with strong support from about two thirds of the politicians and from the mainstream media.
11. Experience indicates that support for the affirmative case falls significantly between the announcement of a proposal and the actual vote. This is because the voters have then had some opportunity of hearing both sides of the debate.
12. Those who are uncommitted in a poll tend to move to the No case, or in the poll do not reveal an intention to vote No. This is because the republican camp has been successful in suggesting the monarchist case is old fashioned, dated, etc.
13. Polls taken now indicating opinions at some future date, say, the end of the reign, are clearly unreliable.
14. That even if the Prime Minister and Leader of the Opposition were to support the Yes case ina referndum, this will not ensure success, as was demonstrated in 1967. But if there is no No case, i.e. the Parliament unanimously supports the referendum, this can significantly help the Yes case. It may be that the support of an unpopular Prime Minister and/or government may harm the Yes case. This was said to be one of the reasons why Paul Keating chose not to put a referendum on a republic.
15. The theme of any referendum on a republic will probably be around the proposition that only a politicians’ republic can deliver an Australian Head of State. To counter this, constitutional monarchists will need to be as well organised and as disciplined as they were in 1999.
...general comments on polling...
Opinion polls do not claim absolute accuracy and will usually indicate a margin of error. They are not predictions as to the future but an attempt to measure opinion at the time of the poll. This applies to views about what may or may not happen at the end of the reign. These are views held now, not one swhich will emerge at the end of the reign.
There can be errors or a bias in taking the sample. For example a telephone survey excludes those who do not have landlines. Some people will be reluctant to answer, or may give an answer they think the questioner wants.
By looking at trends from different polls taken over time, differences can be neutralised.
...the right question?..
Opinion polls can be biased in formulating questions. This can be unintentional.
The question may vary considerably from the referendum question. A referendum necessarily involves agreeing to a specific republican model. But some polls purporting to measure voting attitudes in the 1999 referendum ignored this and tested support for some vague undefined republic.
But in questions concerning constitutional change certain words can mislead.
For example, there is a debate between republicans and constitutional monarchists over the meaning of Head of State, and the question to be answered in the referendum may not even use that word.
“ Do you think an Australian should be Head of State instead of The Queen ?” assumes we do not already have an Australian Head of State, which is a principal point in issue in the debate.
This is important. In the 1999 referendum, the Yes case used the argument that only in a republic could we have an Australian as Head of State nine times, more than any other.
Even asking whether Australia should become a republic assumes we are not already a republic, albeit a crowned republic
...have they heard both sides? ...
When referendums are announced, it is common to find polling indicates strong public support. But this can change after the public has heard both sides.This was exacerbated in the nineties because the mainstream media supported the republican movement. At the same time the media thrives on conflict and even a biassed media is forced to allow the other side to be heard at least partially.
In the early stages of the campaign in the nineties the public had not really heard both sides of the debate. They had heard more by the time of the referendum.
Isolated polls should be treated with caution. The trend in polling from different pollsters over time is a better indicator. It is particularly unwise to rely on one poll which goes against the trend.
In 2009 the republicans released a poll to coincide with the tenth anniversary of the referendum. This indicated that 59% support for “a republic.” This went against all the trends and was what may best be called a “rogue poll”, which, we hasten to add, suggests no impropriety.
In Australia the best known pollsters are:
- Newspoll - published in News Limited's The Australian newspaper
- Roy Morgan Research - published in the Crikey email reporting service
- Galaxy Polling - published in News Limited's tabloid papers
- AC Nielsen Polling - published in Fairfax newspapers
Although less well known, UMR has also conducted polls on this issue. Its polls have always found substantially more republican support than any of the others.
Essential media is a new pollster more associated with the unions, without this resulting in any bias.Its political polling produces results broadly in line with the other polllsters.