Ejisu By-election: An Analysis by Global Info Analytics

A day after Ejisu poll was published, many analysts are pondering over the numbers and others are perplexed about the poll which sampled 1351 voters from 19 out of 28 electoral areas of the constituency. The sample of the electoral areas were drawn from heavy, middle and low density populated areas of the constituency, thus ensuring that areas covered are representative of the constituency and allowing us to carry out analysis of the rest not polled using Multilevel regression with poststratification (MRP) analysis to carry out deeper understanding of the constituency.

In the special Ejisu poll, the NPP’s candidate, Kwabena Boateng, leads the poll with 50.6% and Kwabena Owusu Aduomi, an Independent candidate, has 47.5% of the votes. The NDC is not contesting the by-election. The poll’s margin of error is 3.42%. As the gap between the candidates is 3.1%, which is withing the poll’s margin of error, the race is statistically tied, meaning any of the candidates has a chance of winning the election on Tuesday.

The poll shows that 70% of the respondents were NPP voters, 9% were NDC, 15% were floating voters, 2% other parties and 4% did not disclose their party affiliation. However, the most significant danger for the ruling party is that 36% of their own voters intend to vote for the Independent candidate, a former MP for Ejisu. The poll also shows that there is massive coalition against NPP in the constituency as 92% of NDC voters intend to vote for Aduomi in the absence of NDC candidate. Again, 66% of floating voters intend to vote for Aduomi, 95% of other parties voters will also vote Aduomi and even among those who declined to disclose their party affiliation, who are a proxy of NPP voters in the constituency, 60% intend to vote for Aduomi.

Our model suggests that if the NPP is able to hold their nerve and get at least 70% of their base to vote for Kwabena Boateng, then, it is likely the NPP could win the election. However, if support Kwabena Boateng among NPP voters falls to near 60%, while Aduomi hold his nerve among other parties support or even better, he could deliver a seismic shock to the ruling party on Tuesday and by that altering which party becomes majority if he decided to caucus with the NDC. However, this is unlikely in my view but he could then become most powerful MP in the august house.

The NPP has all to do on Tuesday and their ability to increase turnout while maintaining 70% support from their base for Kwabena Boateng will be crucial. Anything short may not be good enough. If the NPP wins by a whisker, it could dent a huge blow for them in the region and could embolden other independent candidates to seriously consider the idea in many constituencies.

Now on the presidential poll and matters arising, I have had a queries it. There are suggestions that the poll is confounding because the Independent candidate doing so well but Alan Kyerematen not matching his numbers. The answer is simple. Presidential election is not parliamentary election so, we should not expect people voting for Aduomi to automatically vote for Alan. NDC voters are also voting for Aduomi and will certainly vote for Mahama and not Alan. Alan support, which is 13%, was largely from NPP.

The Ejisu poll is also confirming a pattern observed nationally, which is DMB is losing some of the base to Alan Kyerematen. Data from 2020 elections shows that Nana Addo won Ejisu with 82% of the votes but DMB is currently doing 70%, representing a drop of 10%. The poll also shows that 9% of NPP voters in Ejisu intend to vote for Alan, 2% for Mahama and 88% for Dr Bawumia.

In summary, Ejisu by-election could go either way but at the moment NPP has the edge.

Get full polling questionnaire here: https://1drv.ms/b/s!AjkwSuYBHbHhgZMtFcR5upYD43hxdQ?e=AEJfJm

Get the full poll report: https://1drv.ms/b/s!AjkwSuYBHbHhgZMYHWwSu3e0zrc8Ag?e=olbOTd