BE SURPRISED but not totally gob-smacked if Malcolm Turnbull loses his seat of Wentworth at next Saturday’s election – even if the Coalition wins.

Look at this way. One of the main reasons for the dramatic fall-off in Coalition support since the heady days of its huge winning margin shortly after Turnbull assumed the prime ministership has been the large disappointment in Turnbull himself.

Imagine, then, how magnified that might be in his own electorate?

Well imagine no more. Roy Morgan did some polling this week. Its main aim was to see how the Greens are faring in their best 20 seats. Co-incidentally one of those seats was Wentworth.

At the last election Turnbull won the seat with 63.3 per cent of the first-preference vote. All those chardonnay- and latte-sipping trendies along the well-to-do harbour and beach suburbs of eastern Sydney adored him. They liked what he said about marriage equality; the republic; public transport; and his sophisticated engagement instead of three-word slogans.

How much more disappointed and betrayed they must feel than the Australian electorate at large with the three words “Jobs and Growth” and little else?

One of the best measures of the personal popularity of a Member of the House of Representatives is to compare their Reps and Senate vote. In 2013 the Senate vote for the Coalition in Wentworth was 42.1 per cent. Turnbull’s vote was 21 percentage points higher. It was the best differential in the country. But that was then. This is now. This week’s Morgan polling in Wentworth had Turnbull’s (or the Coalition’s) vote down to 41.5 percent. It seems that the totality of his personal vote has evaporated and in Wentworth he is now seen as just another politician.

And where has that support gone, according to Morgan? Labor got 8.1 per cent of it taking them to 27.5 per cent. The Greens got 7.4 per cent of it taking them to 22 per cent. Together Labor and the Greens have 49.5 per cent of the vote in Wentworth. Others have 9 per cent.

If Labor gets the lion’s share of Green preferences and some from others, it is in with a chance.

Indeed, I looked at Sportsbet’s odds for Labor winning the seat. They were 21 to 1. Looked like a good bet to me.

But, and this is a big but, the sample size in the poll was quite small. Just 310 people in each of 20 electorates polled.

Those with little knowledge of statistics might think that a sample size 310 in an electorate of 150,000 is as good or better than the standard 1200 sample in the national electorate of 16 million.

Not so. The absolute size of the sample is the critical thing, not its percentage of the population.

An example will illustrate the point. Say, you have a population of four – two red and two blue and you do a sample of that population by randomly selecting two. The sample is 50 per cent of the population, but the chances of your sample accurately predicting the colours of the whole population is very poor – about 50 per cent. In other words, half the time you will get it completely wrong, saying the electorate is all blue or all red when it is in fact half and half.

The bigger your absolute sample the greater than chances of wrong results cancelling each other out and the greater the probability that the sample reflects the whole population.

It is one of the major difficulties for pollsters in single-member systems like ours. To get an accurate result you would have to poll the 30 most marginal seats with sample sizes of 1200 in each. That would cost a lot of money. So they don’t do it.

Overall the Morgan poll shows that the only hope of the Greens picking up more seats in the House is in the seat of Batman where the Greens primary vote is around 40 per cent – similar to that of the successful Green Adam Bandt.

But the trouble for the Greens in Batman and, indeed, in Bandt’s seat is that the Liberals are giving there preferences to Labor in those and all other seats, rather than the Greens.

That will almost certainly end the Greens’ hope for any seats outside Victoria. And it will make their chances of holding and winning seats in Victoria more difficult. However, the most recent State election results suggest the Greens will increase their vote in Victoria.

Overall, the Morgan poll showed significant swings to the Greens in all but a handful of the 20 seats polled. When we combine this with a poll giving the independent the lead in Indi; Bob Katter getting Labor’s preferences in Kennedy, Independent Andrew Wilkie’s strong position in Denison where he easily topped the poll in 2013, the strong showing of the Nick Xenophon Team in South Australia, and the closeness of the two major parties, it looks like a hung Parliament is very likely.

Sportsbet gives it one to 3.5 odds. I think it more likely than that.

And there is nothing wrong with a hung Parliament or a Senate where the government does not get its way. A hung Parliament will do what the voters want – force the major parties to pay more attention to policies for the general good rather than the sectional interests or pork barrels in marginal electorates.

Get used to it.

By the way, people often say that the punters get it right and the pollsters get it wrong. Not so. Rather, it is only after polling informs the punters that they start getting it right.

My guess is that, as the knowledge of the Morgan poll spreads during the week, the odds on Turnbull losing his seat will (pardon the pun) shorten.

www.crispinhull.com.au