Tuesday , September 26 2017
Home / ARTICLES 2008-2015 / Client leaked Synovate poll

Client leaked Synovate poll

By Mubatsi Asinja Habati

An opinion poll by the research firm, Synovate Uganda, published by The New Vision newspaper has generated debate about who commissioned it. Virginia Nkwanzi is the Country Manager of Synovate Uganda. The Independent’s Mubatsi Asinja Habati spoke to her

The New Vision published an opinion poll claiming it was conducted by Synovate on behalf of the opposition. Did you conduct the poll?

Synovate did the poll but we did not do it for the opposition. We are well-known pollsters conducting polls every year with or without elections. Our polls can be about social, economic and political issues. It was not that we were doing the poll because of politics but we are an information services company, and we are in the business of providing information. Therefore these opinion polls are an issue of information gathering to give the public knowledge. So we did not do the poll for the opposition. We did it as Synovate and usually when we do our polls we always ask different people if they can buy the questions. We structure our questions basing on what is happening. So as a company we have been monitoring the campaigns and we know what is going on. The poll had more questions than those talked about in the press.


What were your findings?

You saw them in the papers. Those were correct results. Even if anybody asked us to defend that research I would definitely do it.

So you stand by those results that said Museveni would have got 67% and Besigye 19% of the vote if the election was conducted in December last year?

Definitely yes. It is not questionable at all. That is our report. Opinion polls are not about politics, they are about perceptions.

How did your poll leak?

Because we collect information and write reports so many people come to us looking for this information. It could be that some of the companies that got this information from us, the information got out of their hands but it did not leak from us because I have so many other questions that would have been of interest to people and they were not published. So the press people picked what was the heated topic at the time.

Have you made predictions on Ugandan elections in the past?

We have done all the years. Whenever there is an election we do a poll.

How accurate have they been?

They have been very close. We produce the results based on when we do them. This particular poll was done in December when campaign rallies had just started but now if you did another poll, maybe the results would be different. That means the timing of the opinion poll matters and of course the same methodology has to be maintained for easy tracking. Again there is the factor of changing perceptions, if for example, we did a poll in February towards the election, the perceptions might have changed. That is a poll will predict the right results if done when it is close to elections and if the elections are free and fair. We have always conducted previous polls a month or two to the election and within that they may predict the actual picture in terms of winning percentages for each candidate. Our predictions have always been close by 5 percent more or less in the final elections results.

How reliable can your opinion polls be given that Uganda’s elections are sometimes characterised by rigging as the Supreme Court has twice confirmed the rigging?

Where they say there is rigging, I have not seen somebody come out outrightly saying I caught so and so rigging. They just say ‘we suspect there was rigging because this and that happened’. Can they tell us to what extent the rigging affected the results? If they can say that without rigging the results of the poll would have been this percentage, then we would be able to justify that our poll was not spot on. We are talking from the research point of view. What if the rigging was about 10 votes or 1000 votes, to what extent does this affect the result?

But this would alter the final results.

Yes it would, but to what extent? One percent cannot be compared to 10 percent even in our data we always say at least an error of plus or minus 2 is acceptable. However, if there is an error of 10 percent then that is wrong and the data may not be reliable.

Don’t your opinion polls bias voters in favour or against particular candidates or political parties?

Sometimes they do. People will always like to go with the majority. People are influenced by a number of factors and in different ways. It is not that poll results will move all voters to one candidate. There are those who lose hope and decide maybe to vote for a winning candidate, others stick to their candidates. It depends on what strategy a candidate takes to convince people. Whether the results portray a picture one does not want or a picture they want, it is important that they take them positively. If they are positive it shows what they should do to maintain that position or even do better. If the results show negative perceptions, it should awaken them on what strategies they should use to improve that. A person who is not scoring highly in these polls should use it as opportunity to go out there and say how do I improve my stand.

Leave a Reply

Your email address will not be published. Required fields are marked *