OPINION: Research World International Opinion Polls are uncredible

By Besi Ndereya

On 6th/05/19, Research World International (RWI) during a public event at Hotel Africana presented results of their most recent opinion poll: “Perceptions of Ugandans towards socio-economic and political issues in the Country.” The strategic purpose of this poll was to “measure public opinions and perceptions on the social economic and political aspects in Uganda.” This exercise was conducted across 60 districts with a total sample size of 2,042 between 12th April 2019 and 25th April 2019.

The methodology used is alleged to have been a Quantitative study approach based on a multi stage stratified random sampling design. As is practice world over, Opinion polls are usually designed to represent the opinions of a population by conducting a series of questions and then extrapolating generalities in ratio or within confidence intervals. A poll that “leans” to one side or the other relative to other polls is not necessarily wrong; it may be that the other polls are wrong. However, a “biased” poll is one that election results show to be wrong. In this piece, I will attempt to deconstruct the conclusion some are drawing on “decline of Museveni’s support” as well as the subtle intent to replace Dr. Besigye with Mr. Kyagulanyi as “Opposition king pin” in scheduled 2021 general elections.

I will show this subtle intent to be negatively biased and therefore wrong. The authors among many set out to gauge peoples’ views about governance, economic situation of the country and views about elections and voting intentions. The sample composition was 51% female, 65% rural and 71% with education level ranging from “No education to part secondary education”. Also 60% of sample identifies as self-employed, 18% not employed, 12% full time employment, 7% part time employment. Just 2% identified as students. Within a reduced strata of 1137, 64% of sampled size identified as being NRM supporters, 18% as FDC leaning whereas only 10% supported Mr. Kyagulanyi’s “People power”. It is from this reduced strata of 1137 that Research World International at the behest of a mystery funder then conclude that support for H.E. Museveni has declined to 32%. An irony because they fail to explain how 64% of 1137 which approximately 728 respondents is support NRM and yet do not support Museveni.

Thus a ‘bias’ against Museveni. No amount of extrapolation can wish away this clear “bias”. Only a national election can show this “bias”. From this same strata of 1137 respondents, RWI asked “Do you support the idea that the opposition should front one joint presidential candidate?” The results are as follows: 39% said yes, 38% disagreed. 15% are not sure if it be a good idea or not while 8% made no comment on the subject. This to me can only mean that there is an even split between those in support of this idea or against it AND those 23% who are undecided. It is at this point that RWI, and I suspect at the behest of a mystery client, now attempt to manipulate the poll in favor of replacing Dr. Besigye with Mr. Kyagulanyi as the official opposition “king pin” at the next elections. They subtly now ask this same strata a leading question: “If yes, who should be the joint opposition presidential candidate?” Though it seems direct enough a question, it actually is ambiguous.

It is a kin to the ambiguous question asked in a 1992 poll commissioned by the American Jewish Committee. It was phrased thusly: "Does it seem possible or does it seem impossible to you that the Nazi extermination of the Jews never happened?" This implies that only the 39% who said yes to the question of fielding a single opposition candidate can answer. The remaining 61% of the sample is excluded. This now reduces the strata from 1137 to approximately 443 persons. Of these 443 persons, RWI then concludes as follows: Mr. Kyagulanyi 42%, Dr. Besigyye 24%, Hon. Mao 5%, Maj.Gen (Rtd) Muntu 3% and Hon. Jimmy Akena 1%. Results based on just one third of a sample are often inconclusive and certainly uncredible.

Yet RWI casually dismisses this statistical principle. In light of the questionable sample size of 2,042 respondents drawn from 60 districts and also in light of the ambiguity of the questions especially in light of attempting to statistically alter national support for president Museveni, this poll wilts and withers away. I am confident that the 2021 Presidential Election results shall without a doubt vindicate me by proving this poll to be wrong.

The writer is a Communications Assistant at Government Citizen Interaction Center (GCIC), Ministry of ICT and National Guidance.