Statistics have been bent since the word emerged in ancient Greek. The treasury of the Athenian Democracy was notoriously short in comparison with the figures of State, which is what the word derives from. The more modern use, the statistics part of science and mathematics, still causes much debate in terms of Hume’s concerns on induction and possible remedies to it. All this said, we can apply numbers to problem solving if we are honest about it. All statistics are practically fucked if we fail on honesty, or get what we allow in wrong, whether dishonestly or not. At theoretical levels, we know that what we formulate may later expose us a ‘Sorcerer’s Apprentices’ as all kinds of unexpected consequences flood in. However, we do not lack theory good enough for practice. We lack honesty in practice and the critical thinking that allows the ‘sums’ to be formulated right and effective spreadsheets to be created. One of the main problems in statistics practice is how easy it is to make it look as though a load of hard work has been put in hiding worthless data in a process designed to evade real enquiry. Most people lack the numeracy to spot sleight-of-hand in how the numbers are produced, and the critical thinking needed to spot how question selection and collection methods may be skewed to keep unwanted stories in the dark.
There is more of less no point in teaching everyone statistics in order to stop the abuse we see more or less every day. Most of the rubbish is probably produced by former students of people like me, and frankly, academe itself is hacking a lot of the stuff out. Nearly all the problems arise at a ‘pre-mathematical’ stage in analysis, some of which is to do with incompetence, but most of which is to do with fear of rocking the boat and losing jobs or pay for telling the truth, or even trying to get at the real problems. Most people in the business can see they won’t stay in a job or get the next contract if they produce the real, critical stuff. This is not even about fault finding a lot of the time – opportunities to open up new correlations that show just how positively we should value some programmes are regularly missed.
One of the first things to understand is how massively costly research is. This is the major barrier to people being able to do their own, and gives interests with money the opportunity to get “research” done skewed to their own interests. This system has been in place a long time and is now in vicious circle. Most academic research is frippery and their are bent people like Cyril Burt and JH Schon around making up “data”. The problem has been written about for years, one classic called ‘Silent Spring’ by Rachel Carson from 1962. I have papers from 1910. Even medical research is massively biased, and forensic “evidence” is far too often no such thing.
Bent police statistics could give us a great avenue to adventure solutions to a much wider problem.