By Gabriel Oguda
Ever since IPSOS released that study that said CORD supporters are poorer than their JUBILEE counterparts, there is one bigot who has been camping on my inbox asking me to justify why I disagree with the findings. He says I am just an opposition sycophant who should be given the contempt card just like â€˜your clueless Babaâ€™.
I have since blocked him, because he couldn’t agree to engage in a civilised debate. For the benefit of everyone, I want to tell you why I think that study should not be taken seriously. And I want to address the issue of poverty vis-a-vis political affiliation.
I have gone to the IPSOS Kenya Twitter page and they havenâ€™t bothered to share with us their sampling methodology. Without that, I can only make the assumption that they did clustered random sampling, which just means that they first divided Kenya into Counties upon which they allocated every county a certain quota of respondents. For example, of the 17 constituencies that constitutes Nairobi County, 9 of them are occupied by TNA MPs – more than half of the total, yet the Nairobi Governor is a CORD guy.
So, did IPSOS categorise Nairobi County as a CORD or JUBILEE zone based on the fact that the Governor is a CORD chap? How many people from Nairobi County fell into the overall sampling frame? And which constituencies did they come from? What if three-quarters of their Nairobi respondents were derived from Nairobi constituencies that were sympathetic to JUBILEE?
These questions will never be answered by IPSOS, because there is only one study that does poverty mapping in this country and it is called the Kenya Integrated Household and Budget Survey (KIHBS). And you know what, there is no question on political affiliation in the KIHBS questionnaire. Simply because; (i) political affiliations change overnight, and (ii) there is no proven correlation between political affiliation and consumption patterns (for example, there is no scientific evidence to show that those who support Raila Odinga eat three plates of deep fried chips per day, while those supporting Uhuru Kenyatta drink 100 bottle-tops of kumi kumi per hour.
But that is not even where the fundamental flaw is.
Research has proven that reported income is not the best way to measure poverty, because humans tend to underreport or exaggerate their poverty, or wealth, status based on prevailing circumstances.
Let me give you an example. If the government announced that from today henceforth the allocation of the Constituency Development Fund will rely on three indicators (poverty, marginalisation and terrain), and my MP wants Seme consttuency to be allocated more money than Mandera West, he will simply call my mother down in the village to inform her that if a group of data collectors knocked on her door collecting CDF data, let her tell them that her children have long abandoned her and they donâ€™t send her anything, in cash or in kind, and that she only relies on goodwill from neighbours and the church to survive.
She must also say that she has never gone to school and that she has more than 8 children all of whom are currently roaming around the local market scrapping the barrelâ€™s bottom for a hard-fought meal. Because,while analysing poverty-related data, the chaps at the Kenya National Bureau of Statistics (KNBS) assign what they call â€˜weightsâ€™ to poverty indicators.
For example, oneâ€™s level of education has a bearing on their general poverty (a primary school dropout is, averagely, poorer than a university graduate), the same to number of children (a household with 7 kids is, averagely, poorer than a household with 3 children), the same to geography (a household in Jimo village is, averagely, poorer than another household in downtown Nairobi.
If everyone in Seme constituency answered the poverty mapping questions like my mother did, then Seme constituency will be considered the poorest electoral area in the entire country. Everyone knows that is not the truth, and that is why the KIHBS questionnaire goes further than â€˜opinion pollingâ€™, they actually ask that they come home and confirm that you donâ€™t have any domestic animal the way you claim, and that your kitchen hasnâ€™t seen any fire for a week the way you responded when they asked you that questions earlier.
This is exactly what happened during the 2009 National Census where results from certain constituencies were nullified.
The Commission for Revenue Allocation (CRA) went ahead and used this data to allocate CDF funds to constituencies for the 2013/2014 FY. There is still a court battle going on these contested results, although the then Planning Minister, now Kakamega Governor, Hon. Oparanya, said that these results had since been harmonised and there should be no cause for alarm.
There is this question that IPSOS asked that was basis of this ruckus: â€œDoes anyone in your household ever go to sleep hungry?/Does everyone in your household always get enough to eat?â€
Of course! If you asked me that question with the knowledge that if I answered to the affirmative I would be enrolled to a social protection program, of course I would say that I sleep hungry every day. Everyone would. And that is why that question shouldnâ€™t be analysed on its own. Every thoroughbred researcher will tell you that whenever they design data collection tools (quantitative or qualitative), they always sneak in one or two questions for triangulation purposes. These are questions that help the interviewer catch any inconsistencies in the responses being given.
The KIHBS 2 Survey started last week and it will run for one full year, thatâ€™s a staggering 12 months of nothing but data collection. Yet IPSOS have the balls to go into a site for four weeks and come out with skewed data and they want us to accept them?
I want to know which Ethical Review Body sanctioned this study. These people give research a bad name.
Gabriel Oguda is a policy researcher and political consultant.