KM 1966

“Some chatter on social media would have you believe that crime is out of control. The facts tell a different story.” No. The facts tell a MUCH bigger story than the handful you've chosen to present. That's the problem. You pick the statistic. You pick the definition. You pick the baseline. You pick what is included. You pick what is excluded. Then you tell the public: “Trust the data.” Let's actually look at your claims. You say violence is at its lowest level since the Crime Survey began. Fine. Say what it actually is: THE CRIME SURVEY. It is a survey estimate of people living in households. It is not a census of every crime committed in Britain. ONS itself says only around 36% of the headline crimes covered by the survey — excluding fraud and computer misuse — were reported to the police in the year to March 2026. Then: “Robbery is at a record low.” Which robbery measure? Because ONS specifically says POLICE DATA are its preferred measure of robbery because the Crime Survey finds relatively few robbery victims and its estimates can fluctuate. And what did the police data show? Personal robbery fell. Good. But overall recorded robbery rose by around 10%. That's rather different from simply tweeting: “Robbery is at a record low.” Then comes my favourite: “Crime, excluding fraud, is 78% down from its 1995 peak.” EXCLUDING FRAUD. You literally put the exclusion in the sentence and then use it to tell people how wonderfully crime is falling. Why exclude one of the biggest crime categories in modern Britain when you're making a sweeping statement about crime? The latest full ONS figures before today's release estimated around: 9.6 MILLION headline Crime Survey offences. Including approximately: 4.5 MILLION incidents of fraud. And fraud wasn't statistically falling — the number of fraud victims had actually risen. So “crime excluding fraud” is doing a fucking enormous amount of work in that sentence. Then: “Knife crime is down 11%.” Again — which definition? Because the Government's own knife-crime strategy explicitly EXCLUDES knife possession from its headline measure. Why? The Government says possession figures are more affected by policing activity. Fine. But perhaps mention what that excluded data says as well. Police recorded 30,216 offences of possessing an article with a blade or point in the year to June 2026 — reported as the highest level in that series. That doesn't cancel out the fall in knife-enabled offences. Both things can be true. THAT IS THE POINT. Show us both. Then: “Thefts including phone snatching have fallen by 20%.” Again, good. But shall we continue reading the tables? Police still recorded 519,516 shoplifting offences in the year to June 2026. Yes, that was down around 2% on the previous year. Say that too. But don't select “phone snatching down 20%” and pretend you've just described theft across Britain. And your “3,800 extra neighbourhood officers”? Your own Government describes that figure as more than 3,800 OFFICERS AND PCSOs being put into neighbourhood policing roles. That's not quite the same sentence as giving the impression you've simply recruited 3,800 extra police officers. Words matter. Definitions matter. And apparently they matter a great deal when politicians are choosing which ones to put in a tweet. Then we get: “We're tackling fraud with £250 million.” Lovely. Spending £250 million is an INPUT. It isn't an outcome. How much fraud occurred? How many victims? How much was lost? How much was recovered? How many offenders were identified? How many were prosecuted? THAT tells me whether it worked. This is why people don't trust you. And it isn't because they are too stupid to understand statistics or because they've been poisoned by “social media chatter”. People have watched governments do this for years. Change the dataset. Change the methodology. Change the terminology. Change the counting rules. Reclassify things. Put things under “unknown”. Put things under “not currently recorded”. Revise the figures later. Say two datasets aren't directly comparable. Then produce whichever percentage makes the press release look best. And the Government's own statistical guidance openly acknowledges that these things happen. Home Office immigration data, for example, explicitly contains “unknown” and “not currently recorded” categories. Its statistics are produced from live administrative systems and can subsequently be revised because of late cases, changes to cases, data cleansing and changes in categorisation. The Home Office even says that after moving from the old CID caseworking system to Atlas, some statistical categorisations were no longer available and methodologies had to be changed. And the Home Office has said it does not publish statistics on some subjects when it considers the information it holds isn't good enough to support public debate. Think about that sentence. THE STATE HOLDS INFORMATION. THE STATE DECIDES THE QUALITY ISN'T GOOD ENOUGH. SO THE PUBLIC DOESN'T GET A STATISTIC. Meanwhile an Independent Chief Inspector examining asylum casework reported in 2026 that decision quality had declined significantly and reached an “unacceptable level”. So spare me: “The facts tell a different story.” Give us ALL THE FACTS. Give us the favourable data. Give us the uncomfortable data. Give us the excluded categories. Give us the unknowns. Give us the methodology changes. Give us the revisions. Give us the recording-rule changes. Give us the raw tables. Give us the money. Give us the outcomes. And when a dataset isn't good enough to publish, tell us exactly WHAT isn't good enough and WHY. Then we can decide what story the data tells. Because you don't get to curate the evidence and then lecture the public about misinformation. Labour doesn't get that privilege. The Conservatives didn't deserve it either. I don't want your fucking narrative. I want the ledger.