James Ball, data journalist, the Guardian
Click here to load reader
-
Upload
joelmgunter -
Category
Documents
-
view
1.549 -
download
3
Transcript of James Ball, data journalist, the Guardian
![Page 1: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/1.jpg)
DON’T KILL THE audience
WHY JOURNALISTS NEED DATA –AND data needs journalists
![Page 2: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/2.jpg)
statsitisA terrible new diseases – statsitis – affects 1 in 1,000 people. It is always fatal. Thankfully, it’s treatable.
The treatment always cures you if you’ve got statsitis. But if you haven’t and take the treatment, it’ll make you REALLY ill.
There is a test, 95% accurate, that will tell you if you have it.
You take the test. You test positive. Who wants treatment? What’s the chance you’ve got statsitis?
![Page 3: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/3.jpg)
USE AND ABUSE OF STATS
![Page 4: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/4.jpg)
REALLY?How much does a slice of bread weigh?
…what about a pizza base?
…what about cheese?
…meat?
…tomato?
![Page 5: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/5.jpg)
use and abuse of stats
![Page 6: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/6.jpg)
REALLY?INCOME: £2,000 / month
OUTGOINGS: £2,050 / month (including rent, debt, etc)
DEFICIT: = £2,050 - £2,000 = £50 / month
DEBT: £20,000
I cancel my £20 / month gym membership. Do I need to cut another £30, or £19,800?
![Page 7: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/7.jpg)
use and abuse of stats
![Page 8: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/8.jpg)
REALLY?MEDIAN FULL-TIME INCOME (UK): £26,000
TAX ON £26,000: £3,905
A significant % of people don’t work, or work part time.
So how, exactly, are we paying £4,000 each for pensions?
![Page 9: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/9.jpg)
USE AND ABUSE OF STATS
![Page 10: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/10.jpg)
REALLY?Highest ever TV audience: 1.1bn
Actual (est) wedding audience: 300m
....we all do it sometimes.
![Page 11: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/11.jpg)
REMEMBER THIS?A terrible new diseases – statsitis – affects 1 in 1,000 people. It is always fatal. Thankfully, it’s treatable.
The treatment always cures you if you’ve got statsitis. But if you haven’t and take the treatment, it’ll make you REALLY ill.
There is a test, 95% accurate, that will tell you if you have it.
You take the test. You test positive. Who wants treatment? What’s the chance you’ve got statsitis?
![Page 12: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/12.jpg)
Here comes the science partA 95% accurate test will get it wrong 5 times out of 100.
So for 1,000 people, 5% x 1,000 = 50 will test positive. Of those, only 1 is ill.
So if test positive, you have only a 1 in 50 = 2% chance of having the illness.
What does that mean in real life?
![Page 13: James Ball, data journalist, the Guardian](https://reader037.fdocuments.us/reader037/viewer/2022100507/559bb8b71a28ab85538b47bf/html5/thumbnails/13.jpg)
Who else gets this wrong?The example just given was based on a real study performed on a group of professionals.
The first time this was carried out (Eddy, 1982), 95% of professionals over-estimated by a factor of ten
They thought a positive result meant a 75% chance of having the illness. In reality, it was 7.7%
They were all doctors.