All polls published ahead of the European Parliament elections, including the one commissioned by Times of Malta, suggested Labour was on track to win by a much bigger majority than it actually did. So what went wrong? 

Pollster Vincent Marmara says undecided voters appear to have upended predictions. Fellow pollster Morgan Parnis, whose company Esprimi ran the Times of Malta survey, thinks similarly.

Marmara explained in a Facebook post that he had used the same, reliable, scientific models as in the past, but the rate of undecided voters was exceptionally high in the run-up to polling day at 15% to 18%.

It appeared, he said, that undecided voters who at the 2022 general election had voted PN, had again cast their vote (for the PN) on Saturday, whereas many of the undecided who in 2022 voted Labour had this time either stayed away or voted for independent parties.

'Certain events' which unfolded near the end of the electoral campaign happened too quickly to be factored into the last opinion polls, he said in a clear reference to the arraignment of Joseph Muscat.

Nonetheless the polls were accurate in projecting three EP seats for each of the main parties.  

He also pointed out that there were major instances abroad when opinion polls were off, as in the case of Brexit. 

Parnis agreed in comments to Times of Malta that the primary issue appeared to be the undecided voters.

“For the question 'if a general election was held tomorrow', we had predicted a 6.9% gap which is closer to the actual result,” he observed.

Parnis said his team would be looking at sample sizes and the frequency of surveys but he noted that other polls carried out more frequently and with larger samples were also significantly off.

“Understanding these discrepancies will be crucial for improving our methodologies in future polling efforts,” he said.