This column is written by a journalist. We rank 22 in a list of 26 in an Australian Readers’ Digest survey published this week about Australia’s most trusted professions.
Politicians are at the bottom, just above them are car seller and real-estate agents. Lawyers are on 21. Ambulance officers, fire-fighters and assorted medical and help-giving professionals are on the top.
The poll sampled 1500 Australians in a population of 20 million. Pollsters were not rated among the professions. Perhaps that is because the vast majority of the population, asked about the accuracy of a poll of a mere 1500 people in a population of 20 million, would laugh at the poll’s obvious absurdity.
And they would be wrong.
The problem with polling is not so much whether the sample is representative of the whole, but whether the opinions extracted are worth anything.
A sample of 1500 people, if selected randomly, will provide a chillingly accurate assessment of the opinion of 20 million people’s often-worthless opinions.
The accuracy of polling can be determined fairly easily. Good pollsters run some simple demographic questions with every poll, such as age group, sex, postcode and the like. If the sample matches census data on the demographic questions, it is a sign of a good sample. Half female, spread of postcodes, spread of ages and so on. Even the age question is has been crafted to remove bias: invariably the age groups avoid the sensitive 30, 40, 50 and 60 by asking age ranges 25-34, 35-45 and so on because of the tendency of people to understate their age just after they hit a round decade.
Once you have got a random sample of 600 you can be 95 per cent confident that you have got the population’s opinion within three per cent. Is does not matter much whether it is a survey across 20 million Australians or just, say, Belconnen. You will still need a sample 600 in both cases for that degree of confidence. The sample does not have to go up proportionately with population size.
Of course, the sample has to be randomly selected and conclusions drawn only according to the question asked.
But don’t the pollsters always get it wrong, I hear you say.
No, that is an unfortunate myth based on decades of election polling. Election polling has many pitfalls, especially in Australia.
The main pitfall is not the skewing of the sample, but the skewing of the population. The major political parties aim for middle-ground approval. They change their behaviour and policies to this very end. If, on their own internal polling, they appear to be out on a limb, they usually move back to the centre. Small wonder there is rarely much between them – a 55-45 split is as much as you will ever get in an Australian election. Such a tiny gap increases the chances a poll will be on the wrong side of it, but not by very much.
The polling of these narrow political contests is about the only polls that are put to the test. A poll of 600 taken on Tuesday has its accuracy verified (or not) on the very next Saturday when the whole nation votes and all the votes are counted. Pollsters’ credibility is at stake on the very polling which is most likely to get the wrong result (even if only by a very small margin). But the perception of error remains.
Further, in the Australian system, the election winner is determined by seats won, not most votes won. In 1990 and 1999 the winner got fewer votes than the loser. Again, it makes it harder for the pollster.
And elections are run in a very volatile climate with intense media bombardment that is likely to change people’s voting intention backwards and forwards quite often — more chance for the pollster to get it wrong.
And when the pollsters get the election result wrong, the media has a post-election, finger-pointing field day. Unfortunately, for pollsters they are judged in this harsh environment and the public believes, wrongly, that they get it wrong most of the time.
If pollsters were polled on their trust-worthiness, they might come below journalists or even politicians.
But on 60-40 or 70-30 events (like Coke vs Pepsi) pollsters would get it right every time, though no-one is likely to verify such polling with a nation-wide vote or census question.
The professional-trust poll is almost certainly right. People do think journalists, politicians, car sellers and real-estate agents are scum and that ambulance drivers and nurses are saints. However, it does not mean they are scum or saints.
The professions thought of as scum in this poll are the ones that do things in public or do things that people can easily expose. Houses and cars often fall into disrepair after purchase and the agent or seller is blamed. Journalists’ work is in the public for all to see and pick holes in publicly. Politicians’ work is all public. The courts are open to the public.
One the other hand, people in professions who help individuals in private circumstances are in a different position. Their competence is hard to verify so they are regarded as saints. Their mistakes are quietly buried. Sure, there is the odd medical negligence case, but nowhere near the scrutiny that a politician gets.
Fire fighters (and ambulance officers) are seen dashing about with sirens or in the media at the firefront. The bulk of the time, though, they are hidden away. Rail-track workers, on the other hand, are seen as slackers. They toil for hours only to rest on their shovels for the few seconds that a train passed – the very time that people can see them slaking.
It may be that “good” people are attracted to “selfless” professions, but my guess is that nurses, pharmacists, ambulance drivers and the like have a roughly equal number of incompetent, selfish, caring, competent, screwball, generous, idiotic, sagacious, witty, dull, wonderful and deplorable individuals as any other professionals. No matter. What maters is perception.
This is an accurate poll – chillingly accurate. It describes what people think about the trustworthiness of each profession. But it is not a gauge of the actual trustworthiness of each profession. They are entirely different things.
The masses who were polled are simply not a good judge of trustworthiness. Why do I say that? Because the very profession which the masses rate last on the trustworthy stakes – politicians – is the only profession whose ranks are filled solely by the vote of the masses themselves. The people actually voted in these untrustworthy individuals – proving that the people are not very good judges of trustworthiness.
What a delicious paradox.
It makes one proud to be a journalist. There is no shame in being rated poorly by the same masses who are gullible enough to vote for people who in the next breath say are untrustworthy.
I would be nice to think that the polls are inaccurate. Alas no. They provide an accurate description of the folly and ignorance of the vast ruck of people being polled.