MaltaToday electoral survey | PN and Labour lose points, Greens at 2.4%

A PROFILE OF THE UNDECIDED VOTERS | Methodology and full data [Google Drive]

The MaltaToday survey shows a small dip in support for both parties and a surprising 4-point increase in undecided respondents just a week before the election.

On the other hand, the percentage of respondents who refused to reveal their voting intention has remained the same as last week. 

Normally, due to increased pressure on voters by political parties, it is this category which increases before elections and not those who are genuinely undecided.

This is an indication that despite an uneventful week following weeks of scandal-mongering, a number of respondents could be rethinking their vote. 

This is reflected by the decrease in support for the two major parties. The survey was held before and after the debate between the three political leaders organised by the Times of Malta and before Friday's Xarbank debate.

Support for AD rose slightly to 2.4%, which is the highest registered in the past year of MaltaToday's surveys. 

For the first time, AD is attracting a small but consistent shift from both the PN (2.2%) and the PL (1.3%) and a relatively high percentage of new voters (8%) and under-35-year-old voters (4%).

The survey's sample had practically an equal number of respondents who voted PN or PL in the last general election. This means that the sample matches the 2008 general elections result. First-time voters - which account for 29,000 (19,000 of which were under age in 2008) - are under-represented in the sample.

The percentage of people who intend not to vote has decreased slightly from last week. Joseph Muscat continues to enjoy a very high trust rating of 44% against Gonzi's 30%.  

A profile of the undecided

One in every four respondents is either undecided or would not reveal who he or she intends to vote. This percentage would rise to 29% if non-voters are included. Since none of the two major parties surpass the 50% mark in the survey, the role of these three categories could be crucial in determining the outcome of next week's general election.

The survey indicates that PN voters outnumber PL voters in all three categories.

In fact, while 26% of undecided voters declared voting PN in 2008, only 15% of presently undecided voters declared voting PL.

Among non respondents - the majority of which did not reveal how they had voted in 2008 - 10% declared voting PN while 7% declared voting Labour in 2008.

Former PN voters account for 39% of present non-voters.

None of those intent on not voting this time around are former PL voters.

In view of this, the PN would increase its support by 7% if it recovers all its past voters who are presently undecided, not replying or intent on not voting. 

Labour would only increase its support by 2.8% if it does the same.

Yet despite the predominance of PN voters among these categories, the PN would still be trailing Labour by 7.8 points if it recovers all these former PN voters. 

Even recovering these voters at such a late stage is not an easy task for the PN. 

80% of those intent not voting do not trust either Muscat or Gonzi. Moreover, Muscat is more trusted than Gonzi among both the undecided and non-respondents. While 20% of the undecided trust Muscat, only 11% trust Gonzi while 22% trust neither. 

One of the most curious aspects of this survey is that while non-respondents have not increased as one could expect a week before the election due to increased pressure on voters by political parties, the number of undecided has shot up by 4 percentage points. 

This increase in undecided voters comes in the wake of a decrease in support for both major parties. Interestingly, 43% of these voters declared which party they had voted for in the past election. This indicates that a large chunk of these voters are genuinely undecided. 

This component could be vital in any attempt by the PN to cut the margin but could also shift to AD especially if they are convinced that the greens have a realistic chance of making it to parliament. One factor considered by PN-leaning undecided voters is whether the PN has any realistic chance of winning. It is also possible that former PN voters who are still undecided but trust Muscat more than Gonzi could shift to Labour, thus increasing its majority.

The margin of error

One important consideration in any survey is the size of the margin of error. A margin of error occurs in any survey and depends on the size of the sample used, the greater the sample the lower the margin of error. 

In this case, the margin of error amounts to +/-3.7%. To decrease the margin of error to +/-2% one would need a sample of 2,200 respondents.

This means that Labour's support in this particular survey could be anything between 36.5% and 43.9%. On the other hand, support for the PN could be anything between 24.5% and 31.9%. 

This means that if one were to put Labour's support at a minimum of 36.5% and the PN's support at a maximum of 31.9%, Labour would still be leading by 4.6%.  

But one has also to consider that the lowest gap registered in the past year of surveys been of 9 points.

Support for AD could be anything between 0% and 6.1%. In AD's case, the range in the margin of error includes the both the prospect of a remarkable success and a complete wipe out. Still, the last seven electoral surveys have indicated that AD has seen its support increase from 1.5% to 2.4% - which would represent an increase of 60% in its support.

PN losing 11% of its former voters

The survey shows the PN losing 9.2% of its 2008 voters to the PL and a further 2.2% to AD. This means that the PN is losing 11.4% of its 2008 voters to the other two parties. In last week's survey the PN was losing 10.4% to both parties.

The percentage of PL voters in 2008 now shifting to the PN amounts to just 1.3%; which is the same amount Labour is losing to the greens. 

For the first time in the past year the PL is losing more than 1% of its voters to the greens. In total, the PL is losing 2.6% of its 2008 voters to the other two parties.

Over the past week, the net swing between the two major parties has increased from 5.8 points to 7.9 points.

This week's survey sees the PN retaining less of its 2008 voters than last week. 

In fact, the percentage of votes retained by the PN has decreased from 72.4% to 70%. This is corresponded by a 2-point increase in the number of undecided Nationalist voters.

Labour has also seen a slight decrease from 92% to 89% in the percentage of former PL voters who intend voting the party again. This is also corresponded by a 2-point increase in undecided former Labour voters.

In this survey, AD retains 50% of its 2008 voters and loses a third of its former voters to the PL.

The party compensates for these losses by attracting a swing from both major parties as well as an amount of new voters and non-voters in the 2008 election. Among new voters, AD reaches its highest level of support: 8.3%.

If AD manages to retain its undecided 2008 voters it would reach 2.8%.

Significantly, the PN is not compensating its losses to Labour by making gains among new voters, as was clearly the case before the 2008 election. In fact among this category of first time voters the PL enjoys a 13-point lead over the PN.

The PN is also trailing Labour by 24 points among another pivotal category - those who did not vote in the 2008 general election. 

The 2008 election had seen the lowest turnout since 1971 with the number of non-voters increasing by 9,000 over 2003 levels. 

The survey indicates that 38.2 % (down from 47% last week) of non-voters in 2008 will be voting in March and the overwhelming majority of these will be voting for Labour.

Support for AD in this category has risen to 4.8%. Significantly the survey shows a higher level of indecision than last week among this strategic category of voters. Indecision among this segment has shot up by 8 points.

New voters, which were pivotal to the PN's victory in 2008, are also shifting towards Labour. 

This category has been quite volatile in its voting intentions with the PN leading Labour over most of the past year but Labour has gained an edge over the past weeks.

Younger voters more undecided

The survey shows both major parties registering losses among younger voters. While the PN loses 10 points, the PL loses 7 points over last week. AD retains its 4% in this category.

The decrease in support for both parties is corresponded by a sharp rise of 17 points in undecided young respondents. In fact, nearly one in every three respondent aged under 35 years of age was either undecided or refused to reply. A further 4% are intent on not voting. 

Over the past week, both major parties have managed to retain the same levels of support in the other age groups. 

Labour now enjoys a 15-point lead among 18 to 34 year olds, an 11-point lead among 35 to 54-year-olds and over 55-year-olds. AD has seen an increase in support among over 55-year-olds where its support doubled over the past week. But AD loses some support among 35 to 54-year-olds.

Muscat retains trust lead

In the trust barometer, Joseph Muscat enjoys a 14 point lead up from 12.5 points last week.

The survey suggests that just one week before the election Gonzi has still not recovered one of the greatest hurdle for his re-election, which is Muscat's higher trust rating. This is a complete reversal of the situation in 2008 when Gonzi was more trusted than Alfred Sant.

Interestingly, although 9.2% of PN voters in 2008 would now vote for Labour, 11 % of PN voters in 2008 now trust Muscat more than Gonzi.

This suggests that a small category of former Nationalist voters prefer Muscat but are hesitant on switching party allegiance.  But this factor could also make it harder for the PN to reduce the gap.

Only 1.3% of Labour voters in 2008 prefer Gonzi to Muscat. This is equivalent to the percentage of voters shifting from the PL to the PN.

Methodology

The survey was held between Monday 25 February and Thursday 28 February. A total of 1,002 respondents were randomly chosen from telephone directories and contacted by telephone.  Of these 700 accepted to be interviewed. Results were weighed to reflect the age and sex balance of population. The survey has a margin of error of +/-3.7%.