Doorknocking can be arduous work during political campaigns, particularly on a hot morning.
So before the 1998 Federal election, it was a welcome relief as I trudged the streets of suburban Penrith, to knock on one door and be invited in for a cuppa.
Sometimes, however, supporters of other political parties will be hospitable - to slow you down. A half-hour chat means that you will not talk to as many voters. I knew what Larry was up to, but I needed a break.
Over tea and biscuits, my host ruminated on a new political phenomenon; "me and the boys at the pub have been talkin', and we think that this Pauline sheila is alright".
It was just a straw-in-the-wind, which would, within three weeks, turn into a political cyclone. Without warning, Pauline Hanson's One Nation Party in the 1998 federal election, stormed the polls picking up more than one million votes across the nation.
The pollsters missed that one completely.
They also confidently predicted the end of the Howard government after only one term. On election night, the prime minister was putting the finishing touches to his concession speech, when his chief of staff burst into the room to tell him the government had been returned.
More recently, in Australia before the 2019 election, the Liberals were well behind in 50 monthly opinion polls on the trot, but Morrison won.
In the US presidential election, the pollsters were also wide of the mark. During 2020, the opinion polls consistently predicted that President Trump was behind by seven to nine percentage points and would lose the election by a wide margin - possibly a landslide. When the dust settled, the margin was closer to 2 per cent.
America and the world held their breath for 48 hours, as the two candidates slowly inched towards the magic 270 Electoral College votes required to win. Biden was just ahead throughout the count. The outcome looked like a toss-up.
Late on election night, I was in a texting exchange with two of my children, as we tried to plot the perilous Democrat path to victory through various combinations of states.
Why are polling predictions often so wide of the mark?
Opinion poll results are usually based on questions to quite a small sample of voters. This may be accurate if the sample reflects the makeup of the electorate. In the 2016 US election, they failed to predict the loss by Hillary Clinton, because the sample interviewed was too biased towards urban voters.
They missed the growing wall of "red" support for Donald Trump in the "rustbelt" states of the mid-west and the towns and rural areas of America. These voters became rusted-on Trump supporters, who stuck by him throughout his erratic presidency. They played a crucial role in shoring up his vote in 2020. The pollsters still need to do more work on their sampling technique.
Even if opinion polling is spot on, the "shy supporter" syndrome may come into play. When interviewed, many will not reveal their real voting intentions because they believe in a secret ballot.
The best way to change votes is face to face.
Others just tell lies. This was why both the Pauline Hanson and the Donald Trump vote was always understated. A lot of voters don't want to admit that they support extremist candidates openly.
Opinion polling will often fail to pick up a late change in sentiment. Towards the end of a campaign, after a lot of persuasive advertising and media commentary, voters become more focused on what the candidates have to offer. Just before election day, when many people consider the alternative government policies, they sometimes choose to stick with the 'devil they know'.
In Australia, this was a factor in the victories of Keating (1993), Howard (2004) and Morrison (2019). It may also explain why in the 2020 US election, Trump almost snatched victory from the jaws of defeat.
Finally, the other reason that Trump defied the opinion polls, and almost won the election, was because his Republican supporters had a better "ground game". Recently, I was shocked to read that the Democrats were relying on persuading the voters by electronic means and didn't plan to doorknock. Instead, they outspent Trump by hundreds of millions of dollars on TV and internet advertising, especially in the swing states.
In addition to electronic messaging, the Republicans also carried out an extensive doorknocking campaign. It almost got them across the line, making a mockery of the opinion polls. The best way to change votes is face to face.
Doorknocking seems old-fashioned in our high-tech world, but it really works, provided you don't stop and drink too many cups of tea.
It wins votes and has the power to change the election outcome, making fools of the pollsters and political pundits.
Often, when the final votes are counted, they have to eat humble pie.
Newcastle East's Dr John Tierney AM is a former Hunter-based federal senator for NSW
IN THE NEWS:
