2024 Sri Lankan Presidential Election
Who will be President (Part 3)?
The IHP Poll, Explanations, Clarifications & Serious Opinions
“You take the IHP Poll seriously?”, asked a friend, referring to the Presidential Election Voting Intention (PEVI) component, of the Institute of Health Policy (IHP)’s, Sri Lanka Opinion Tracker Survey (SLOTS).
“I take it seriously. But not literally”, I replied.
“Then, who will win the election?” , they continued.
This is my latest answer.
IHP & SLOTS
The Institute for Health Policy (IHP) is a research centre based in Colombo, Sri Lanka that focuses on health and social policy research.
IHP conducts the Sri Lanka Opinion Tracker Survey (SLOTS), which polls public opinion on various issues including consumer confidence, voting intentions, and views on the country’s direction. The lead investigator for SLOTS is Dr. Ravi Rannan-Eliya, who is also the Executive Director of IHP.
SLOTS is funded by the Velux Stiftung foundation, New York University Abu Dhabi, and the IHP Public Interest Research Fund. Past funders include the Neelan Tiruchelvam Trust, the Asia Foundation, the European Commission, the UK National Institute for Health and Care Research, and the Foundation Open Society Institute.
IHP MRP Presidential Election Update August 2024
The latest (and likely final) edition of PEVI, showed a 3-Horse Race between Anura Kumara Dissanayake (AKD), Sajith Premadasa (SP), and Ranil Wickremesinghe (RW) with 36%, 32%, and 28% of the vote respectively.
Estimates for SP and RW have a error margin of 4.4–6.4%, and those for AKD and NR — 2.2-2.4%. With these error margins the race looks even closer, with any of the three candidates having a reasonable chance of getting the most votes.
Thus, while a literal interpretation of this poll would lead to the simplistic and ignorant conclusion that “IHP says AKD is going to win”, a serious interpretation would be that the race will be very close.
[Note, in Who will be President? (Part 1), I speculated that the actual error of PEVI might be even much larger than reported.]
This “very close” narrative is further supported by other data included in the survey.
Momentum
The direction of recent PEVI results indicate positive momentum for RW, mostly at the expense of SP, and to a lesser extent AKD and NR.
If we were to extend the results one further step into an Imaginary September 2024 edition, RW moves into second place with 34%, with AKD still in the lead with 35%, and SP falling to third with 28%.
Lag
There are two reasons to speculate that the results of PEVI have a time lag; meaning the results of August 2024 actually represent July 2024, or even earlier.
Firstly, Multilevel Regression and Poststratification (MRP), the statistical model used, sometimes smooths effects by time. And so if August is very different from July, this difference might be subdued in the final results.
Secondly, in addition to surveys conducted in August 2024, the statistical model is influenced by previous results, going back to October 2021. It is not clear what the influence of this past history is, and until it is clarified, a lag effect cannot be ignored.
Interestingly, a lag effect might mean that our Imaginary September 2024 result might be closer to the final result, than the Official August 2024 result.
Turning Out
A serious problem that elections polls suffer from is whether the people surveyed will actually come out to vote on Election Day.
The 2019 Presidential Election had a turnout of 83.72%, the highest in any presidential election. Whether 2024 can beat that or even get close is an interesting question.
There are three types who are likely not to vote in 2024.
- Those who didn’t vote in 2019, and might be one of many who, don’t vote on principal.
- Those who voted for Gotabaya Rajapaksa (GR) in 2019. Many of these voters might be apathetic or disillusioned, and won’t vote in 2024.
- The Youth. This is a global trend. Despite being active on social media and Protest movements, younger voters seem less likely to turn out and vote on Election Day.
54% of “Did-not-Voters”, 46% of GR Voters, and 53% of Voters aged 19–29, intend to vote for AKD. And so, he is likely to suffer disproportionately more from any Turnout issues, compared to SP and RW.
Minority Vote
According to this poll, majority of both SL Tamils (55%) and Estate/Indian Tamils (53%) support SP, as do an overwhelming majority of Muslims (71%). This pattern is also clear from intention by province, with 55% of the Northern Province and 71% of the Eastern Province intending to vote for SP.
There is some speculation as to how this will change in the final days leading to the election.
For example, who will vote for Pakkiyaselvam Ariyanethiran, the former TNA MP, endorsed by several Tamil Parties as their “Common Candidate”?
Also, RW and AKD are campaigning fiercely in minority-majority regions. Will they manage to garner a last moment swing?
2nd/3rd Preferences
Despite all errors, momentums, lags, and other uncertainties at play, I very much doubt that any candidate will get anywhere near 50%. SP and AKD are facing headwinds, and RWs late momentum will likely not push him past his opponents.
And so, we face, for the first time, a 2nd/3rd preference count.
AKD vs. SP
As of now, the most likely candidates for the Top 2 slots are AKD and SP. RW, NR and all the other minor-candidates will be eliminated.
Those who placed 1st preferences for RW are probably more likely to place their 2nd preferences for SP, as opposed to AKD. And so, SP is likely to have a 2nd preference advantage. Similarly, voters who placed their 1st and 2nd preferences for RW, and some other eliminated candidate, are more likely to place their 3rd preference for SP.
AKD vs. RW
Assuming RW’s momentum continues, the other Top 2 scenario is AKD vs. RW. Here again, most voters with their 1st preference for SP, are likely to have RW as their 2nd preference.
And so, a 2nd/3rd preference scenario, is likely to favour SP and RW, over AKD.
An open question is whether many people will include 2nd and 3rd preferences. And of those who attempt to do so, how many would do so accurately, and thus not get their votes rejected. The answers will significantly depend on how educated voters are.
As Dr. Rannan-Eliya voices in the IHP MRP Presidential Election Update August 2024,
“…We urge civil society and the Elections Commission to intensify efforts to educate voters about using their preferences to ensure that every vote counts.”
Concluding Thoughts
On the one hand, I’ve said before and I’ll say it again: The IHP could be more transparent on their methodology, particularly on error computations. They could, for example, get independent statistical experts (say, from the University of Colombo, Department of Statistics), to verify their methodology.
On the other hand, it is the most credible and transparent poll on the 2024 Presidential Election, with publicly shared results. This is the first reason I take it seriously.
The second reason is that the IHP’s findings lead to many more serious questions.
Finally, a warning on those who take this poll and other polls literally.
By that I mean the sort of people who make ignorant and simplistic statements like “The IHP poll says AKD will win.”
Many of the same people go on to say stuff like “But that’s because the IHP supports its friends”. Interestingly, those who make these criticisms are silent when the poll supports their candidate, but them crawl out of the woodwork when the winds change.
These literalists can be safely ignored.
And if you are a Sri Lankan Citizen eligible to Vote on September 21st, please turn out and vote. Your choice, whatever it is, is a vote for democracy.
Thanks for Reading!
🟥🟧🟩
