Tuesday, February 19, 2013

What the opinion pollsters won’t tell you


By MIKE OLIEWO  ( email the author)

Posted  Monday, February 18  2013 at  20:00
In Summary
  • It is crucial to know the material day’s or week’s headlines or raging debates during the interview period. For example, a poll conducted soon after a rally in any neighbourhood by either Mr Kenyatta or Mr Odinga will derive varied responses because of publicity
  • It is crucial that the pollsters also reveal their design of polling in terms of ethnicity, gender, disability, religion and age-bracket. Which of the above categories think that either Kenyatta or Odinga is sympathetic to their status?
SHARE THIS STORY
 
 
 
0
Share

We are in the season of opinion polls and it is important to question the interrogator about reasons for the polling and the parameters used. Pollsters should table polling variables that explain whether Kenya is a society or a microcosm of societies within the same country.
There are salient issues regarding framing of questions posed by opinion pollsters which need to be critiqued. The pollsters should account for their questions and the context upon which respondents were interviewed.
Bourgeoisie preferences should be clearly indicated against what the down-trodden and the marginalised like, and how they relate with either presidential candidates Raila Odinga or Uhuru Kenyatta.
It is crucial to know the material day’s or week’s headlines or raging debates during the interview period. For example, a poll conducted soon after a rally in any neighbourhood by either Mr Kenyatta or Mr Odinga will derive varied responses because of publicity.
The spread of the question time and whether it was a one-day or month-long affair is also critical. Responses obtained on the day the media reported the shambolic TNA or ODM nominations would definitely be varied across the country.
Opinion pollsters should also disclose the formula employed to sample different classifications of their research targets. Which economic classes, political persuasions and socio-cultural inclinations were sampled? Did the interviewers differentiate between income levels, educational sophistication and cultural values?
It is crucial that the pollsters also reveal their design of polling in terms of ethnicity, gender, disability, religion and age-bracket. Which of the above categories think that either Kenyatta or Odinga is sympathetic to their status?
Let’s interrogate the statistics. What is the statistical variation between registered voters likely to vote and the registered ones who are unlikely to vote?
Since public schools will be closed from February 25 to March 7, employees who work far from their families may prefer to join them away from their previously registered polling stations, because the leave-period is enough to facilitate travel.
Secondly, there are Kenyans who registered but cannot vote because of stiff work schedules.
Thirdly, traumatic experiences by some Kenyans in the last elections might still be fresh in their minds and could make them hesitant to cast their ballots despite constant assurances from security organs.
The fourth category includes internally displaced persons and whether they have the same political preferences as Kenyans who live and work in their immediate neighbourhoods.
What follows are our cultural ways of life. A very huge percentage of Kenya is inhabited by pastoralists whose livelihood is derived from availability of pasture and water for their animals.
Where were they when the pollsters came calling and where will they be on voting day?
Similarly, from the census results and the just released voters’ register, what explains the disproportions between local demographics and their registered voters and which of the two figures were used by the opinion pollsters?
Mr Oliewo is director, Public Research & Development Consultants based in Nairobi. (mikeoliewo@yahoo.com)

No comments:

Post a Comment