A survey of Kenyan voters exiting polling places for the March 4 election shows a statistical tie between Uhuru Kenyatta and Raila Odinga, with both candidates receiving about 41 per cent of the vote.
The so-called exit poll was conducted by two US political scientists and included about 6000 Kenyan voters. The results captured on this video link were released at a Washington think-tank forum on Thursday.
The exit poll showed Mr Odinga receiving 40.9 per cent of votes and Mr Kenyatta 40.6 per cent, with support for other candidates accounting for some of the remainder.
Nearly 12 per cent of the Kenyans included in the survey refused to indicate for whom they had voted in the presidential race.
The results represent "a statistical tie" between the two top vote-getters due to the survey's margin of error, the pollsters said.
One of the pollsters, Harvard University vitiating professor James Long, acknowledged under sharp questioning from the audience that the margin of error could be as high as three per cent.
But even taking account of various statistical uncertainties, Prof Long said, "There is no reasonable assumption that gets either candidate to 50 per cent."
For example, the contention that Mr Kenyatta benefited from a comparatively much higher turnout in Kikuyu-dominated parts of the country was shown through the exit poll to be "a myth," Prof Long declared.
He and his colleague, Prof Clark Gibson of the University of California at San Diego, declined to identify precise causes for the discrepancy between their survey's results and the outcome certified by the Independent Elections and Boundaries Commission.
The official result showed Mr Kenyatta winning slightly more than 50 percent and thus avoiding a second-round runoff.
The pollsters have "no evidence of people stuffing ballots" or buying votes, Prof Long said.
"Why should somebody believe the results of the election are invalid?" a member of the audience asked.
Prof Long suggested, "Because people broke the law when they counted votes."
Prof Long said at the outset of his presentation at the Johns Hopkins University School of Advanced International Studies that exit polls are "immune" to ballot stuffing or technological breakdowns.
Data from the survey also indicated that both Mr Odinga and Mr Kenyatta enjoyed overwhelming support from their respective ethnic groups.
Mr Kenyatta received 83 per cent of the Kikuyu vote, according to the exit poll, while Mr Odinga got 94 per cent of the Luo vote, the survey indicated.
Among Kalenjin included in the poll, 74 per cent voted for Mr Kenyatta and 11 percent for Mr Odinga.
For Kamba who took part in the survey, the breakdown was 63 per cent for Mr Odinga and 12 per cent for Mr Kenyatta.
Overall, "there seems to be some loosening of ethnic identification" with particular candidates, Prof Long suggested.
In response to a survey question asking participants to name the most important issue behind their choice, most cited either the economy or employment.
Very few voters identified tribalism, land or the cases before the International Criminal Court as the most important issues in the election.