Skip to content

method

Who took part, against who lives there

A count of attendees says how many came. Comparing them with the census population by sex and age says whether the consultation heard the people the decision falls on, and names the groups it missed. How to make the comparison, and how to read it.

Takwimu editorial team, Editorial team5 min read

Most consultations, national or county, can say how many people took part. Far fewer can say whether those people resemble the population the decision falls on, and that is the question an opposing advocate asks first. The comparison is straightforward to make. Count the people who took part by sex and age, set those counts beside the county's population from the 2019 census, and name every group that is much smaller among participants than among residents, including any group nobody reached at all. The Supreme Court lists "inclusive and effective representation" among the components of meaningful participation in BAT (paragraph 96). This is how a county shows it.

Making the comparison

Two numbers are needed for each group: how many participants belong to it, and what share of the county's adults it makes up. The second comes from the census. The Kenya National Bureau of Statistics counted 987,653 people in Makueni in 2019, 489,691 men and 497,942 women, and Volume III's single-year tables give 562,420 adults aged 18 and over, of whom 288,490 were women. The county reference table has the same figures for all 47 counties. Women were therefore 51.3% of Makueni's adults. The first number comes from the consultation's own records, which is why each participant's sex and age group has to be captured when a view is collected, on the forum register, the memorandum form, the web form or the survey.

Take an illustration. Suppose a county consultation records 600 participants and 180 of them are women. Women are 30% of participants against 51.3% of the county's adults, a shortfall of about 21 points, and the statement says so in those terms. If nobody aged 18 to 24 took part although the census counts tens of thousands of them, the statement says that no one in that group was reached. A zero is a finding. It is never smoothed into a small percentage.

Why the counts stay unweighted

Survey analysts weight results so that a sample matches the population, and Takwimu does this for survey estimates. For a representativeness statement it must not. Weighting makes the sample match the census by construction, so a statement built on weighted figures would conclude that every consultation was representative, including one that reached four hundred men and eleven women. The comparison therefore uses the raw counts of who took part. Weighting has its place later, when a county wants to estimate what all residents think from a properly drawn survey, and the two jobs are kept apart.

Reading a gap

A gap does not by itself make a consultation unlawful. The Supreme Court held in BAT that "the fact that someone was not heard is not enough to annul the process" (paragraph 96). What a gap does is tell the county where to look while there is still time. If women are short, it may be the hour the forums were held or the venues chosen. If young people are short, the channel may be the problem, and a web form or SMS round can be added. Governors described one such pattern themselves: at the 2018 summit behind the Council of Governors' report on Makueni, they named "poor representation of middle-income groups" as a challenge (page 5). A county that measures the gap during the consultation can close part of it. One that measures it afterwards can still publish it, which is far stronger than having it produced against the county in court.

Small groups need care. A share computed from a handful of participants moves sharply when one person is added, so a statement reports the count beside the share, and a county should not read much into a group of a few people.

The level the figures support

The census reports by county and sub-county. It does not report by ward. The IEBC draws wards for elections, and no published table joins the two maps, so a comparison of who took part with who lives there can be made with census figures at county level, and at sub-county level where the consultation recorded which sub-county each participant lives in. A claim that a particular ward was fairly represented needs ward figures that the census does not publish. The reference table gives the number of wards in each county and the adults per ward on average, which is useful for planning forums. It is not a substitute for ward-level population figures. Reaching everyone covers how to plan the forums so that the gaps are small to begin with.

Where Takwimu fits

Takwimu produces a representativeness statement for each consultation, comparing the people who took part with the county's 2019 census population by sex and age on unweighted counts. It names any group the consultation did not reach, and it refuses a ward-level claim made on county figures, saying what would be needed to support one.

To see the statement for a consultation in your county, book a walk-through.

Takwimu Intelligence (TI) is a public participation system for Kenya.

Sources

  1. 1

    2019 Kenya Population and Housing Census, Volume I: Population by County and Sub-County · Kenya National Bureau of StatisticsTable 2.2, population by sex and county. Opened 29 October 2026.

  2. 2

    2019 Kenya Population and Housing Census, Volume III: Distribution of Population by Age and Sex · Kenya National Bureau of StatisticsTable 2.3, population by single year of age and sex for each county; the adults aged 18 and over are summed from it in the library's county table. Opened 28 October 2026 for that table.

  3. 3

    British American Tobacco Kenya, PLC v Cabinet Secretary for the Ministry of Health & 2 others, Petition No. 5 of 2017, judgment of 26 November 2019 · Supreme Court of KenyaParagraph 96, the components of meaningful participation. Opened 29 October 2026.

  4. 4

    A knowledge-sharing report on Makueni public participation model · Council of Governors, Maarifa Centre (2020)Page 5. Opened 29 October 2026.

More from the library

  • Deciding on every view, and saying why

    When a consultation is challenged, the document a court reads is the record of what was done with what people said. The Supreme Court described one that passed in October 2024: each proposal set out with whether it was accepted or rejected, similar views grouped by theme, and the whole report public.

  • Reporting back, and proving the report is genuine

    A consultation ends when the people who took part can find out what happened to their views. That means publishing the outcome where residents will see it, replying on the channels they used, and giving any reader a way to check that a printed report is the one the county issued.

  • Reaching everyone a decision affects

    A consultation is judged first on who could take part. That depends on a notice that says what is being decided and where to read it, documents people can understand, the languages and formats the Constitution names, and channels that reach residents who never attend a forum.