Here’s a 5 minute film I did on Newsnight last week, about how politics needs better data. Specifically, it’s about how politicians misuse statistics, how we can stop them, and how we can generate better evidence on what works, and what fails.
If you’re interested in more on this topic, well… there are some good examples of dodgy government stats on this site, and in the chapter on Dodgy Government Statistics in my latest book. For more about randomised trials of government policy, here’s a Radio 4 documentary I did on them, here’s a Cabinet Office paper I co-authored on them, and they crop up in this that I did for the Department of Education. If that’s not enough then I’ll write some more.
Roy Watson said,
February 16, 2015 at 11:13 am
Nicely themed piece, since the very opening statement is at best misrepresentative (if it’s meant to be “This is the Oxford University debating society [misleading pause]…”) or simply wrong (if it’s meant to be “This is the Oxford University Debating Society…”). The point you make about data is reflected here: the barest additional element of contextualisation would make all the difference (two words, making it “This is the Oxford Union, the university debating society…”).
Mark Thompson said,
February 16, 2015 at 11:17 am
Very interesting piece Ben and I basically agree with everything that you say in it.
However what you did not really touch upon is why politicians will not engage with evidence based policy. It’s because (they perceive that) it undermines their raison d’etre. They (or think tanks) come up with whizz bang policies. The politician then “sells” these both to his/her party and then eventually to the wider electorate. It’s a function of the way politics works that this ends up being done by a politician claiming that his/her idea is the best one and will work brilliantly. They wouldn’t get very far in our system if they were equivocal and said the truth which is “maybe it will work better maybe it won’t, we’ll have to see”. In fact new policies would be unlikely to get beyond internal discussions if their “champions” were honest about them.
So whilst randomised controlled trials for policies would be brilliant I think we would have to have some pretty fundamental changes to our political structures and ways of operating before they are likely to gain any significant traction.
Richard Ponsford said,
February 16, 2015 at 1:06 pm
Hi Ben,
Some fantastic ideas here, it’s brightened my mood this Monday morning. Do you have any suggestions on tangible actions I could take to contribute to your suggestions becoming more likely? For example, petitions to sign, groups to join, meetings to attend, donations to make etc? For context, I’m an evaluator at an international charity.
Thanks,
Richard
Jody Aberdein said,
February 16, 2015 at 3:04 pm
Is there a place for a sort of ‘good stats’ kite-mark, that signifies a bona fide statistician thinks the analysis is ok? Perhaps run by the RSS, similar to the little crystal mark the plain english campaign do?
R W Smith (Dr) said,
February 16, 2015 at 4:15 pm
Interesting video… but I would have thought that you, of all people, would have stopped those morons who keep on ruining things with ridiculous background noises. Please don’t let yourself join them!
If what you say is interesting, it doesn’t need added noise, and if it’s not interesting no amount of noise will make it so.
I’m disappointed.
Tom Campbell-Ricketts said,
February 16, 2015 at 7:25 pm
A very nice piece, well done Ben.
I feel a major part of the problem with getting people to accept that politics must rationally be evidence based is the prevailing belief that science has no tools to solve problems of human affairs. The politician is thus assumed to have insight that the scientist can not tap into.
Little thought is given, however, to what amazing mechanism allows the politician to glean such knowledge, the like of which, rational analysis cannot attain. Does he cast bones? Or receive wisdom through dreams or voices?
So prevailing is this crazy culture, that many prominent scientists seem to have climbed hapilly on board – even to the point of encouraging it in their advice to politicians: this is the data, but here my job ends, and you must take over.
I wrote about this at a little more length, in a piece that seems relevant enought to link to without feeling too guilty (hope you agree):
maximum-entropy-blog.blogspot.com/2014/02/practical-morality-part-2.html
Cheers,
Tom
Jody Aberdin said,
February 16, 2015 at 8:38 pm
What about some kind of statistical quality mark, like the plain english crystal mark. It could be run by the RSS, and prominent parties / institutions would effectively have to subscribe to its services to retain credibility.
Ben S said,
February 17, 2015 at 6:07 am
Politics needs better data, but politicians don’t want it; ask David Nutt.
Cat said,
February 20, 2015 at 1:50 pm
You’ve been BBC’d!
www.youtube.com/watch?v=aHun58mz3vI
I’ve never seen you talk in that manner before.
Deirdre Toomey said,
February 22, 2015 at 4:47 pm
Quote Disraeli ” There are lies, damned lies and statistics. And he was a novelist by trade.
David Pollard said,
March 17, 2015 at 11:30 am
Perhaps you could take a look at the NHS misuse of statistics and data capture. For example, from a little while ago:
www.england.nhs.uk/statistics/2014/07/03/gp-patient-survey-2013-14/
With response rate of less than 35% this reckons that “85.7% of patients rate their overall experience of their GP surgery as good”
The latest survey technique is to send a text to patients shortly after they visit s GP surgery to ask them to score on a scale of 1 to 5 if they would recommend the surgery to family and friends.
Aaaarggh. How does this improve medicine?
Kenny Collins said,
April 14, 2015 at 3:14 am
I find this video frustrating in the extreme. Every single point Ben Goldacre makes is clear, rational, and urgently needed. In the USA, we have the OMB to do precisely the kinds of studies Goldacre suggests. But our current politics makes Goldacre’s suggestions sound like something quaint from the 1800s. We have a growing cabal of legislators at the national level who hold the government hostage by refusing to negotiate any legislation at all. Government shutdowns and threats of government shutdowns are becoming the norm. These representatives pander to their constituency by demonizing anything and everything associated with the government. In their rhetoric, if the OMB recommends something they happen to disagree with this week, then the OMB is obviously wrong, and likely part of some conspiracy.
I came here from Goldacre’s two TED talks which I find poisonous. They are specific to Big Pharma, and accurately point out the dysfunctions there. But TED made no effort to promote the talks as specifi to Big Pharma. The growing American audience for something called “Bad Science” will grab at anything to be skeptical of everything scientific.
In the US, I see only one way of reducing the momentum of our anti-intellectuals: reversing our “Citizens United” Supreme Court ruling. Among representative democracies, the USA has a unique artificial built-in two-party system. Third parties never gain traction, because they never get a voice, because they never have any realistic chance of getting in at any level. That polarizes us into two camps, and neither of them have any incentive to cooperate. Coalitions are a 20th century thingy.
Ron Warrick said,
February 10, 2016 at 9:30 pm
In theory, the original design of the government in the United States would have automatically accomplished much of this by allowing the separate states to pass their own laws for most purposes. Other states could adopt those policies from among the states’ best practices. With the death of federalism and the advent of the Federal Reserve banking system, there is nothing to stop toxic policies from being applied nationwide by the national politicians.