Opinion polls reward analysis
AT the weekend the Sunday Vision published results of an opinion poll that showed that the FDC was performing strongly in urban centres against the NRM. But what is the significance of this poll?
AT the weekend the Sunday Vision published results of an opinion poll that showed that the FDC was performing strongly in urban centres against the NRM. But what is the significance of this poll?
The first problem with opinion polls in Uganda is that it is just too expensive to do random sampling in rural areas. A properly conducted American survey in the late 1990s confirmed to donors that the NRM had a decisive lead in the rural areas but cost over $250,000.
Therefore virtually all opinion polls in Uganda merely sample attitudes in towns and peri-urban areas.
Surveying remote trading centres is the nearest researchers can get to a rural survey and yet attitudes in these centres have more in common with the towns than the country.
The second problem is that an opinion poll is a snapshot in time. It will be strongly influenced by events at that particular moment.
The Sunday Vision poll appears to have been influenced by the events surrounding Besigye’s arrest. The poll might have looked very different if it had been conducted in early October, or if Universal Secondary Education had been announced without Besigye being arrested. Next week the results might be different again.
Nevertheless, despite the limitations of opinion polls in Uganda, they still reward analysis.
Why did Museveni apparently perform more strongly in Kampala than in other towns? Is this because people in the capital have more to lose and witnessed the riots? Had NRM done more mobilisation in Kampala?
What is the significance of a majority of NRM supporters saying they opposed civilians being tried by the court martial? Is this why NRM under-performed nationally?
Newspapers and political organisations should conduct more regular opinion polls so that they can follow political trends; assess the margin of error in different polls; and above all attempt to determine what the Ugandan public is thinking.
The first problem with opinion polls in Uganda is that it is just too expensive to do random sampling in rural areas. A properly conducted American survey in the late 1990s confirmed to donors that the NRM had a decisive lead in the rural areas but cost over $250,000.
Therefore virtually all opinion polls in Uganda merely sample attitudes in towns and peri-urban areas.
Surveying remote trading centres is the nearest researchers can get to a rural survey and yet attitudes in these centres have more in common with the towns than the country.
The second problem is that an opinion poll is a snapshot in time. It will be strongly influenced by events at that particular moment.
The Sunday Vision poll appears to have been influenced by the events surrounding Besigye’s arrest. The poll might have looked very different if it had been conducted in early October, or if Universal Secondary Education had been announced without Besigye being arrested. Next week the results might be different again.
Nevertheless, despite the limitations of opinion polls in Uganda, they still reward analysis.
Why did Museveni apparently perform more strongly in Kampala than in other towns? Is this because people in the capital have more to lose and witnessed the riots? Had NRM done more mobilisation in Kampala?
What is the significance of a majority of NRM supporters saying they opposed civilians being tried by the court martial? Is this why NRM under-performed nationally?
Newspapers and political organisations should conduct more regular opinion polls so that they can follow political trends; assess the margin of error in different polls; and above all attempt to determine what the Ugandan public is thinking.