Peter Obi should be mindful of the Bradley Effect
I was on the Nigeria Politics Weekly some weeks back, and during our discussion about the coming elections, I said that while it is true that Peter Obi is ahead in all polls conducted so far, we need to keep our eyes peeled for the Bradley Effect.
Someone sent me a message on Twitter to ask what that means, so this column today serves as a brief explainer.
For 20 years from 1973, a Blackman, Tom Bradley, was Mayor of Los Angeles. He was an excellent mayor, which is why he lasted so long. In 1981, he ran for governor of California against a White dude. Because by all accounts, Bradley was an outstanding mayor and popular, he was the favourite to win the governorship contest.
In short, all the polling agencies had Bradley defeating his Republican opponent by double-digit margins. The polling was so one-sided that the major media outlets cast their headlines, and given that those were the days before the internet, they went to press with those headlines, and one newspaper even began to distribute. Then the California INEC (whatever it is called) announced the results. Bradley lost.
But the elections were validated, and Bradley’s opponent was sworn in as governor, so it was left to people like me, political scientists, to find out what happened.
Here is what happened — in the build-up to the elections, many White voters gave pollsters incorrect responses about how they would vote so they would not be accused of being racist. This had the effect of inflating projections for Bradley. Another thing that was noticed was that there had been a large number of “undecided” voters in the polling before the elections. The vast majority of those “undecideds” voted against Bradley.
Thus, the Bradley Effect was born. In plain English, the Bradley Effect says that because of social desirability bias, respondents to pre-election polls give incorrect responses to pollsters because they are afraid of sounding bigoted. However, their actual vote will come out in the sanctity of the voting booth.
Remember back in 2016, when I suggested that Trump would win and people belly-ached? That was the Bradley Effect. We saw the Bradley Effect in the recent elections in Brazil, where Bolsonaro was predicted to lose handsomely to Lula. Well, that didn’t happen.
From my point of view, when you have many “undecideds” in a contentious election, that election has been set up for the Bradley Effect.
This brings us to Nigeria. Could we be set up for the Bradley Effect?
The first poll that came out regarding the 2023 elections in Nigeria was the ANAP Foundation/NOI Polls survey, which showed that Peter Obi led Bola Tinubu and Atiku Abubakar by 7 percentage points. Premise Data Corp conducted the second poll to emerge on behalf of the American media firm, Bloomberg. The overwhelming majority of respondents to that poll, 72 percent, said Peter Obi was their man. In comparison, Bola Tinubu had four and a half times fewer supporters at 16 percent, and Atiku was a further 7 percentage points behind.
Last week, NexTier Advisory, an Abuja-based think-tank, published its poll, which stuck with rural communities, and showed Peter Obi in the lead at 41 percent, while Atiku Abubakar and Bola Tinubu were 14 and 20 percentage points behind him, respectively.
While I think all of the polls show a general trend, and I suspect that the major parties’ candidates have seen private polls that show similarities hence the overwhelming focus of their supporters on the “third force” candidate, there are certain things I’d like to point out.
First, the ANA/NOI and Bloomberg/Premise polls were phone/online, meaning they sampled mainly the middle class. NexTier’s on the other hand, with its focus on rural Nigeria, appears to have covered that gap. However, the question needs to be asked – what was the percentage of “undecideds”? The average number of “undecideds” in the polls is 32 percent, which is a large enough proportion to move the elections away from the frontrunner in the polling.
I find it hard to believe that in an election which is, in many respects, existential to Nigeria, that one in three prospective voters is “undecided”. Are the many “undecideds” going to plunk for one of the two established party candidates on election day?
Are we being set up for the Bradley Effect here in Nigeria?
Only time would tell.
Nwanze is a partner at SBM Intelligence