Insights
Original analysis drawn from the FOI releases we track — not just a republished feed.
- Bigger requests get refused more - except in policing, where they don't
Across the dataset as a whole, refusal rate climbs steadily with how many sub-questions a request has - simple requests refused 16% of the time, six-part ones 31%. Policing is the one niche that bucks its own trend at the complex end.
- Does pushing back on a refusal actually work? We can only partly answer that
2.2% of every release we track is a genuine internal review - a requester formally challenging the original decision. What that review actually decided is legible in only a small fraction of cases, which is itself worth being honest about.
- Does a slow answer mean a refusal? Two different kinds of 'no', moving in opposite directions
Cross-referencing response speed against refusal and exemption language across 455 policing FOI exchanges reveals a split, not a straight line: quick, clean 'not held' dismissals pull one way, slow substantive fights pull the other, and the middle - not the slowest cases - is where refusal is most likely.
- How long does it actually take to get an answer? Less than a quarter within the legal limit
We extracted real request and response dates from over 450 policing FOI exchanges. Only 22% came back within the statutory 20 working days, and response times have gotten dramatically worse since 2014 - a trend that also shows up, more mildly, in official government statistics.
- Openness comes as a PDF: what format says about how usable this data really is
86% of every attachment across the FOI documents we've indexed is a PDF; genuinely machine-readable CSV is 2.5%. And which legal exemption gets used varies sharply by topic - counter-terrorism and firearms requests get stopped very differently to misconduct and immigration ones.
- The Friday effect: more releases go out that day than any other, and refusal rates go with them
One day in seven is Friday, but a fifth of every FOI release we track publishes then - and Friday releases carry refusal or exemption language at a meaningfully higher rate than the rest of the week.
- Police forces are the likeliest body we track to say they don't hold what you asked for
We read the full text of over 5,000 policing FOI responses, not just the titles. The single most common outcome isn't a refusal under any exemption — it's 'we do not hold this information', and it happens to police forces more than anyone else we track.
- Which organisations actually say no: a league table, not a vibe
Ranking every organisation with enough archived text to measure fairly: the Serious Fraud Office refuses or cites an exemption on 90% of what we can read, the Home Office on 1.8% - despite the Home Office publishing nearly 24 times as much. Refusal rate and publishing volume aren't the same thing.
- GOV.UK barely covers your local police force — so we went and got the data ourselves
GOV.UK's FOI feed almost entirely skips the 43 territorial police forces. Here's why, and how this tracker closes that gap using individual requests from WhatDoTheyKnow.
- The nine-month gap: what a stall in FOI disclosure logs looks like
Between January and September 2023, our policing & justice dataset recorded almost no new FOI releases at all. Then the backlog arrived in batches. Here's the pattern, from the data alone.
- Immigration and asylum is the single largest FOI topic we track — here's what's actually in it
164 FOI releases about immigration and asylum, the large majority from the Home Office. What the questions people ask actually look like, and how that's shifted over two decades.