What a cynical bunch some of you Kenya Forum readers are!

There has been some criticism of the latest opinion poll published by Infotrak Research & Consulting, the results of which suggested that Raila Odinga and his running mate Kalonzo Musyoka have 51 per cent support from voters (see Kenya Forum posting, January 14, ‘Latest Opinion Poll Survey: Raila Odinga And CORD Could Win In First Round’).

Some people have challenged the nature of the poll and its findings but then those coming second or lower in such surveys tend to debunk poll findings. Others have alleged that Infotrak might be biased in favour of Raila Odinga. Surely not?

A Kenya Forum reader sent us the following information which we believe to be correct…

WHO OWNS INFOTRAK?

INFOTRAK OWNERS, MANAGERS AND STAFF:

• ‘Informal founder’ & Group CEO: Angela Ambitho
• General Manager: Tobias Odhiambo
• Operations Manager: Julius Ochieng
• Co-Owner: Jerry Okungu
• Senior Editor: Clinton Ojiambo
• Head of Communications: Janet Oyongo
• Senior Researcher: Adams Oloo
• Administrator: Philip Odhiambo
• Chief Accountant: James Opiyo
• Driver: Fidel Onyango
• Messenger: Harriet Oguda

Is there a common denominator? Kenyan readers will understand…

EQUAL OPPORTUNITIES?

For what it is worth the Kenya Forum actually thinks the methodology and results of the Infotrak poll stand up to scrutiny but when it comes to the company’s recruitment policy equal opportunity employment it is not!

WHO OWNS INFOTRAK?

The Kenya Forum would however, like to know the answer to one or two questions. Angela Ambitho is listed as the ‘Informal founder’ & Group CEO of Infortrak Research & Consulting and Jerry Okungu as a ‘Co-Owner’. Who then is the ‘formal’ owner? And who is or are the other co-owner or owners?

UPDATE, January 21, 2013

The Kenya Forum received these thoughts from a US political scientist regarding this posting and the report ‘Latest Opinion Poll Survey Results’ published on January 14,which we now pass on to our readers.

In its most recent poll on the Presidential election in Kenya, Infotrak found that Raila Odinga had a substantial lead over Uhuru Kenyatta. This advantage, 51% to just less than 40%, was of course, front page news. Obviously the Odinga camp was very cheered by this result, whilst the Kenyatta side minimized its importance, suggesting it could have been a rogue poll, or just badly done.

Some criticized the poll on the basis that no information was presented on a number of items and that the lack of information about refusals (those who refused to answer) and the way they were handled was troubling. This is a good point.

It would be useful to enquire further about how professionally this poll was done and learn something about that by going past what the newspapers reported and checking the Infotrak page for further details. Then we can compare the information that we find against what we would hope to know about the manner in which a proper poll is conducted.

To begin with, we must remember that taking a very precise poll in Kenya is a real challenge. The population is growing rapidly, making the generation of a proper sampling frame a difficult thing indeed. A mobile population is also a problem.

That said, there are some aspects of the lack of disclosure that are troubling, beyond the question of non-responders.

One thing that would be useful to know is more about the demographics of the poll. Infotrak gives us only rural/urban and education demographics. It would really be useful to know the gender proportions and something about income levels, since we know from 2007 that these factors were important, and they would reflect on the representativeness of the sample in any case.

Readers should be aware that pollsters report a margin of error for their polls, a factor that indicates how likely the poll is to be a rogue one. Much of this has to do with the polling frame (in this case people who claim to be registered voters), but it also it affected by the representativeness of the poll, i.e., the extent to which it reflects the composition of the population being polled.

Since we have official statistics to compare the poll composition with the actual national and regional populations distributions, it would be very useful to see these reported by Infortrak. They have them but we don’t see them. It is easy to be suspicious about the representativeness of the poll as a result.

And as a final note, the more lopsided a poll, the less confidence one can have in the result. As a rule a poll with as large a gap as this one would be unlikely to have a margin of error as small as the one reported. In other words, the advice here is “caveat emptor”!