Opinion polls: politicians hail them when the figures are in their favour, express their pleasure at the results but say they are taking nothing for granted; those for whom the results are not so good dismiss the findings and say that the only poll that counts is the one on polling day. Both reactions are understandable and indeed both are right but opinion polls, treated with caution, can tell us a great deal and often more than our politicians perhaps would like.

First a few words of warning. Opinion polls are scientifically based (or at least they should be) but they are not precisely accurate. So one of the important factors in assessing opinion polls is the sample size (if it is over 1,000 people questioned then the poll is getting to a size where the result will reflect the voters intentions, if it’s just a journalist talking to 10 of his friends, then it won’t).

A poll of perhaps 1,500 voters (the sample size) taken to reflect the nature of the electorate, e.g. by sex, age and regional distribution, can produce a result with an error rate of plus or minus 3 per cent. So if the questioned asked was “If there was an election tomorrow, who would you vote for?” and the poll finds Candidate A receives 35 per cent support, he could be on anything between 32 per cent and 38 per cent of the vote at that time, given the statistical inaccuracy of an opinion poll covering that number of people.

The second warning is that an opinion poll is only a snap shot of voting intention, i.e., “Who would you vote for if the election was tomorrow?” The voter could well change his or her mind two days later.  So to get any real benefit from opinion poll assessment you’ve got to follow the results of the polls over a considerable period of time to spot the trends and that brings us to the third warning: to really assess opinion polls you have to look beyond the headline figures.

WHAT THE OPINION POLLS SAY: THE FUTURE’S NOT SO ORANGE?

So what of opinion polls in Kenya? What do they reveal about the likely result of the next election? The answer is by no means conclusive but it is interesting.

The headline figures taken over time have Raila Odinga on about 34 per cent of the vote if he were to stand for president. An Ipsos-Synovate poll published in the newspapers on 2 February had him on 31 per cent. Five days later a poll by Strategic Research put the figure at 36.9 per cent.

The same polls had Uhuru Kenyatta on, respectively, 23.9 per cent and 24 per cent. So if runs for president, or if he’s allowed to run for office, let’s put Kenyatta on 24 per cent of the vote or thereabouts.

That’s good news for Raila Odinga isn’t it? He’s about 10 per cent ahead of Kenyatta, that’s the headline figure. Well, yes it is but in answer to the question, no it may not be: remember, the next election will almost certainly be a two-round election unless one candidate gets over 50 per cent on the first vote which at present seems very unlikely.

WHAT THE OPINION POLLS SAY: SECONDS OUT, ROUND TWO

Raila Odinga, other than a boost for his poll ratings around about the time of the referendum, has been hovering at about 34 per cent support among voters for months. He has strong support from a sold one-third of the electorate but where are his votes to come from if he is to get over 50 per cent in a second round run-off, especially if that run-off is against Uhuru Kenyatta?

Try this not too scientific assessment.

First round over and Odinga (Luo) is on 34 per cent and Kenyatta (Kikuyu) on 24 per cent. All the other candidates drop out, just the two big guys left.

Where do William Ruto’s roughly 10 per cent of the vote go to? To Kenyatta.

Where do Kalonzo Musyoka’s roughly 10 per cent of the vote go to? Mostly to Kenyatta.

Where do Martha Karua’s roughly five per cent of the (Kikuyu) vote go to? To Kenyatta.

Where do Peter Kenneth’s roughly 2.5 per cent of the (Kikuyu) vote go to? All to Kenyatta.

What about Saitoti’s 2 per cent? Surely the Maasai will vote KANU, for Kenyatta?

The Luhya candidates, Eugene Wamalwa and Musalia Mudavadi could muster three per cent of the vote between them in the first round. Where will that go in the second round? Let’s give 1.5 per cent each to Odinga and Kenyatta.

So what does that add up to? It adds up to President Kenyatta being elected on anything up to 54 per cent of the vote in the second round, that’s what it adds up to. If he’s allowed to stand, of course…