MaltaToday survey post mortem: what we did right and what we missed

MaltaToday’s surveys had long been predicting a comfortable Labour Party victory in the election on Saturday and a higher abstention rate than five years ago.

The trend that had been developing over the course of the past five years consistently showed the PL ahead of the PN with strong margins.

On Saturday, this trend was vindicated although the extent of the PL’s victory was more pronounced than that indicated by MaltaToday’s last rolling survey published the Thursday before.

A post mortem of our polling shows that we did get some things right but we also went off mark on some key projections.

What we got right

High abstention rate: MaltaToday’s last survey predicted a share of valid votes of 87%, which represented a drop from the 90.1% recorded in the 2017 election. Although the predication was almost four points off the actual 2022 election outcome, which stood at 83.1%, the MaltaToday survey had from the start of the electoral campaign projected a lower share of valid votes cast. At one point at the start of the campaign, the survey was projecting a share of valid votes that would have reached 84%. Although a high abstention rate was expected nobody believed the election turnout would drop as low as 85.5% (leading to the share of valid votes of 83.1%).

Labour’s vote tally: The MaltaToday survey predicted the Labour Party would obtain 165,420 votes. We got this right with a slim difference of under 3,000 votes separating the survey prediction from the election result. On Saturday, the PL received 162,707 votes, which represented a drop of over 8,000 votes from the last general election. The survey correctly predicted that unsure Labour voters would fall in line and vote like they did five years ago, while those who said they will not vote stuck to their conviction. What we got wrong was the percentage result because this was based on a higher voting base as a result of overstating the number of PN voters (more about this below).

Third party share: The MaltaToday survey consistently polled the share of third parties at 2% or higher throughout the election campaign. The last rolling survey published on Thursday predicted that third parties would get 6,050 votes and a 2% share. The actual result was 9,308 votes and 3.2%, which fell within the margin of error.

*** *** ***

What we got wrong

PN vote tally and share: The last rolling survey predicted the PN would get 137,548 votes and a share of 44.5%. We were clearly off mark on this. The actual result saw the PN receive 123,333 votes and a voting share of 41.7%. The difference between our prediction and the election result was of 14,215 votes. It is evident that the assumption made in the survey that unsure PN voters would return back to the fold did not materialise. Throughout the electoral campaign, the raw survey results showed that uncertainty was higher among historic PN voters than it was among PL voters. In actual fact, this had been a recurring trend over the past five years. It transpires that uncertain PN voters were more motivated to stay at home or possibly even vote for third parties. This led to a collapse of more than 12,000 votes for the PN when compared to the 2017 election. MaltaToday’s survey did not pick up the extent of this collapse, which would have required more variables to be used to interpret the intention of unsure voters. It is evident that the mind-set of unsure PN voters was different from that of uncertain PL voters. This conditioned the gap and the percentage share of PL voters (see below).

The vote gap: The inability to correctly interpret the intention of uncertain PN voters automatically meant that the vote gap projected by the survey was less than that of the actual election result. MaltaToday’s last rolling survey on the Thursday before the election put the gap at 27,698. The election result delivered a gap of 39,474, a difference of more than 11,000 votes when compared to our survey. The difference is down to an overestimation of PN support.

The Labour share: The PL share of the vote in the election was 55.1%, while MaltaToday’s survey predicted 53.5%. Despite correctly identifying the number of votes the PL would get, the percentage in our survey was conditioned by a bigger voting base because the number of PN votes was overestimated.