WHAT COULD HAVE GONE WRONG WITH THE POLLS?
5 PLAUSIBLE EXPLANATIONS
Since last Saturday’s Governorship Election in Ekiti State and the announcement of the election results in favour of the candidate of the All Progressives Party (APC), I have received a number of calls from some friends and associates, seeking to understand what happened with our predictions from the pre-election polls conducted prior to the election. Some of the calls were from partners and senior friends whom we have worked with in the past and can attest to the quality of our work, and the level of professionalism we deployed in conducting our polls and surveys. There has been a barrage of questions- What went wrong? Why did you guys miss this one? Was your methodology faulty? Did you guys miss anything out in the analysis? Did you not consider the political influence and federal might as important variables in your analysis? And so on….
First things first, I wish to state categorically that the two streams of pre-election polls conducted by NOIPolls were NOT funded by any individual, organization, group or political party. They were simply in-house decisions to feel the pulse of Ekiti residents regarding voters’ preparedness and intention for the gubernatorial election. Secondly, NOIPolls wasn’t the only organization that conducted and released polls prior to the election. Our first poll was conducted between 18th and 23rd June 2018, and it put the candidate of People Democratic Party (PDP) ahead of the APC candidate by 8 percentage points. Our 2nd poll was conducted in the week of the election between the 9th and 12th July 2018; and again, it put the PDP candidate ahead of his APC counterpart by 25 percentage points. Then there was a poll by CrowdForce, released on 1st July 2018, which fell right within the margin of error of our 2nd poll and also predicted a win for the PDP candidate, with about 26 percentage points.
With the election done and dusted, and the announcement by the Independent National Elections Commission (INEC) declaring the candidate of the APC winner of the election and Governor-Elect (with 51 percent of total valid votes cast) over his opponent, the candidate of the PDP (with 46 percent of total valid votes cast); the question on the lips of most of our followers has been- What went wrong with the pre-election polls and predictions?
Let me commence by stating, in my personal opinion, I believe strongly that the discrepancy witnessed between the pre-election predictions and actual result released by INEC, has more to do with politics and the conduct of the election, than with the polling methodology. I wish to state categorically that the methodology applied in conducting the pre-election polls were in line with our typical rigorous and robust polling processes, and we can stand boldly to defend the outcomes of the polls. However, it’s worth noting that all over the world, results of public opinion polls have the tendency of been swayed by major happenings and events in the polity; and in the case of Ekiti, there may have been some events building up to the election day, which may have swayed public opinion.
From the perspective of an experienced pollster, I can say there are really no silver bullets to explaining what went wrong. However, I’ve had time to reflect deeply on the questions with my colleagues; and come up with 5 plausible explanations for the observed variance between the poll predictions and the actual result.
Here are my personal thoughts and I wish to share them, in no particular order, as follows:
Ekiti Voters simply changed their minds
My first plausible explanation is that Ekiti voters may have decided to change their minds on who to vote for on the D-Day for reasons best known to them. Although there have been several postulations in the media regarding what may have caused this sort of flip-flop; the Civil-Servants Factor, the non-payment of salaries and the high-handedness of the seating governor may have been some of such factors responsible for the discrepancy.
Alleged Electoral Malpractices
The media has been awash with reports of alleged electoral malpractices that trailed the 2018 Ekiti state governorship election. There have been reported cases and allegations of vote-buying, rigging and ballot box snatching at some polling units, leading to the cancellation of some results; and these may have contributed to the observed variance between the pre-election prediction and the actual result.
Heavy Presence of Security Personnel
Prior to the election, the Nigerian Police had announced the deployment of 30,000 police ocers to Ekiti state. Similarly, there were reports of heavy presence
of the Military and Civil Defence personnel in the state; and incidents of dispersing of crowds with the use of teargas canisters It is instructive that these may have restricted movements and discouraged some voters from coming out to exercise their franchise, in fear of not wanting to be harmed. As the post-election data from INEC revealed, only 405,861 persons were accredited to vote, out of the 909,585 registered voters and 667,270 voters that collected their PVCs in the state; indicating over 55 percent of registered voters who for some reason decided not to participate in the process.
Alleged Bullying and Intimidation of Party Agents
Another plausible explanation important for consideration involves the alleged cases of bullying, intimidation, and arrests of party agents and loyalists in some parts of the state. There have been some news reports in both the traditional and social media citing incidents of bullying, intimidation, and arrests of party agents, stalwarts, and loyalists. Again, these may have discouraged some intending voters from coming out to exercise their franchise. Anecdotal evidence and amateur video footages have captured some of these incidents.
Parties Re-Strategizing with Pre-Election Poll Result
Finally, one factor that cannot be ruled out as a plausible explanation is that all the pre-election poll results were released to the general public, and all political
parties and their strategists had access to the results of the polls on our website and social media platforms. Consequently, I strongly reckon that the results of the two pre-election polls conducted by NOIPolls and the one by CrowdForce may have served as pro-bono consulting and used by APC strategists to re-tool their strategies and recalibrate for victory. Besides, there are theories that PDP may have become complacent, is that all the polls predicted their victory by a wide margin.