How the Campaign Sorts What It Hears
Daniel Okafor · · 7 min read · More by this author
Every case that comes through a ward surgery is categorized at the time by the volunteer who takes it, from a fixed list of eighteen categories. Two years of that is 2,140 cases. Here is what we do with them, and what we are careful not to do.
The counts
The top three, together, are 46% of everything raised:
| Category | Cases |
|---|---|
| Childcare hours and cost | 419 |
| School transport | 286 |
| Housing waiting-list position | 271 |
| Parking and permits | 128 |
| Drainage and flooding | 119 |
The remaining thirteen categories are all below 6% each.
What this is good for
Three things, reliably:
- Prioritizing casework capacity. If childcare is a fifth of the load, the volunteer who understands the provider landscape should be on the rota every week, not every third week.
- Catching a change early. Drainage cases in the Kelvin catchment tripled in the six weeks after the September overflow, before any official figure was published.
- Arguing in committee. “Residents are worried about school transport” is a claim. “286 logged cases over 24 months, of which 212 relate to three routes withdrawn in 2021” is evidence.
What it is not good for
It is not a survey. It tells you about the people who come to a surgery, which is a specific and unrepresentative group: older than the district, more likely to be a long-term resident, more likely to have had a previous good experience of asking.
We know, for instance, that private renters are heavily under-represented in the casework relative to their share of the district. That is a fact about the surgeries, not about renters, and the fix is a change to how we run surgeries rather than a reason to discount what they tell us.
The category problem
Eighteen categories is a compromise. Too few and everything lands in “other”; too many and two volunteers code the same case differently.
We audit this: twice a year, fifty randomly chosen cases are re-coded blind by a second volunteer. Agreement currently runs at 84%. The disagreements cluster where you would expect — between “housing condition” and “housing allocation” — and we rewrote the guidance for those two after the last audit.
Publishing it
From this term, the anonymized category counts are published quarterly. If the top three change, the priorities change with them, and anyone can check whether they did.