Wentworth by-election: Why has Kerryn Phelps' lead over Dave Sharma gone down | Ben Raue | Australia news

Last night at 10 pm, while Wentworth's by-election coverage was over, independent candidate Kerryn Phelps was leading the poll, with 54% of the votes cast against Liberal candidate Dave Sharma and most commentators (not to mention the candidates and parties) had called the race at Phelps. However, when we woke up this morning, the race was much tighter. So, what has happened?

Phelps won a clear majority of the votes cast in local polling stations on polling day. She obtained 54% of the votes after the distribution of the preferences. Yet she did a lot worse by voting before the poll and by mail.

The votes before the vote (counted late last night) strongly favored Sharma, who managed 55% of the votes after the preferences of the four main voting booths. He did particularly well in Rose Bay, in the highly pro-liberal port suburbs, while Phelps only managed a slight majority in Waverley, although he won a large majority in the surrounding isolations on ballot.

We are still waiting for the rest of the mail ballots, but the first group strongly favored Sharma. He polled over 64% of those votes after preferences.

It is difficult to know how many postal votes remain to count. The Australian Electoral Commission has sent 12,788 postal voting files to voters, but not all will be returned. So far, 5,463 postal votes have been counted, and 1,266 more are waiting to be opened. There is more to arrive. If another batch of just over 3,000 postal votes was to favor Sharma as strongly as the first, it would put him ahead.

The Phelps campaign therefore hopes that relatively few postal votes will be cast or that the remaining votes by mail will be less favorable to Sharma.

Why were postal votes and pre-ballots so much stronger for the Liberal Party? It is normal to expect the Liberal Party to be successful in postal votes, but not 18% better. It is also unusual to see that pre-ballot votes deviate as much from voting on polling day.

It is possible that this reflects a late passage to Phelps in the last days of the campaign. It may also reflect the large Jewish population of Wentworth, some of whom are more likely to choose to vote before polling day and would likely favor Sharma.

None of this is particularly encouraging for Phelps, but there is another potential problem.

The analysis of polling day results suggests that there might be a problem in two or three polling booths, where Phelps received much lower preference flows than in other polling booths.

In most kiosks, Phelps received about 65% to 80% of the preferences of the other 14 candidates. Yet she received only 57% of the preferences at Bondi Beach and 49% at Bellevue Hill.

This could potentially be explained by a counting error at the end of a long day of work. The CEA always ensures that votes are counted again to check for errors. Therefore, if an error occurred, we would expect it to be detected in the next few days.

If these booths had more online preference feeds with similar booths, you would expect Phelps to get about 900 votes over Sharma. If that happened, it would become much harder for Sharma to take the lead in the mail ballots.

Finally, why was it so difficult to project the result accurately to Wentworth last night? It was particularly difficult because Phelps was a new candidate, with no candidate background in the electorate.

Election projections generally compare the results of the votes to the data of the same isolated voters in previous elections. This process allows you to eliminate bias in the first cab report and predict what the voting numbers will look like at the end of the count.

It was possible to make assumptions about the conduct of advance polls and correspondence, compared to voting on polling day, but the votes were astonishingly strong for the Liberal Party, which is the main reason for which the overnight forecasts have been replaced by a more cautious analysis today.