When I was investigating the ‘Effect Size’ I found lots of criticism of significance testing on Social Science websites. Remember, this is once again, Social Scientists, often but not always Psychologists, criticising the way Mathematicians and Scientists do Statistics.

This is actually a fundamental part of the ‘Effect Size’ story as their failure to understand the significance testing procedure has led directly to the ‘Effect Size’ as they try to solve a ‘problem’ that isn’t really a problem, only a misunderstanding on their part.

It is also vital to recognise that the ‘Effect Size’ isn’t just another statistical method to choose from amongst many, it is the tip of the ice-berg of a completely different ethos. The people who advocate using the ‘Effect Size’ think that the whole way Mathematicians and Scientists do Statistics is wrong so they’ve decided to invent their own version. This has been mistakenly copied by people in Education like John Hattie.

In my next post I’ll be looking at the Maths of significance testing, but, what if you don’t know anything about Alpha levels or Type 1 and 2 errors, how could you judge? Well, a good place to start would be the mathematical credentials of the people making the criticism. So let’s have a look at the people who are criticising significance testing.

If we type in ‘Criticism of Significance testing’ into Google, the first ten results are –

http://community.dur.ac.uk/r.j.coe/teaching/critsig.htm – Number one on the list, our old friend Robert Coe, Professor of Education at Durham University

http://en.wikipedia.org/wiki/Statistics – A general article on Statistics by Wikipedia

http://www.cem.org/attachments/publications/CEMWeb037%20The%20Case%20Against%20Statistical%20Significance%20Testing.pdf – CEM, Professor Coe’s organisation publishing an article by Ronald P.Carver, Professor of Education and Psychology at the University of Missouri

http://errorstatistics.com/2012/12/24/13-well-worn-criticisms-of-significance-tests-and-how-to-avoid-them/ – Deborah Mayo, Professor of Philosophy at the University of Pennsylvania

http://www.johndcook.com/blog/2008/11/18/five-criticisms-of-significance-testing/ – John D Cook, Consultant in Applied Mathematics and Computing

http://www.uic.edu/classes/psych/psych548/fraley/ – R.Chris Fraley, Professor of Psychology at the University of Chicago

http://www.johnmyleswhite.com/notebook/2012/05/10/criticism-1-of-nhst-good-tools-for-individual-researchers-are-not-good-tools-for-research-communities/ – John Myles White, PhD student in Psychology

http://www.andrews.edu/~rbailey/Chapter%20two/7217331.pdf – Andrews University Education department. Authors, Jeffery Gliner, retired Professor of Psychology, Associate Professor Nancy Leech, PhD in Philosophy and MA in Counselling, George Morgan, retired Professor of Education

http://lesswrong.com/lw/g13/against_nhst/ – No information

http://www.ncbi.nlm.nih.gov/pubmed/17002771 – Authors – Dr Fiona Fidler, Environmental Science, background in Psychology and Philosophy, Mark Burgman, Environmental Science, background Zoology, Geoff Cummnigs, retired professor of Psychology, Robert Buttrose, background in Philosophy, Neil Thomason, historical and philisophical studies

And so it goes on, page after page of Psychologists, Philosophers and Education Professors critisicing the way Mathematicians and Scientists do Statistics.

So, you can judge for yourself the quality of the people criticising the way Mathematicans do Maths. Though this time we do seem to have a lot of Philosophers as well as Psychologists.

Now, this is important because, their mistakes in significance testing have led to the ‘Effect Size’, which has led to Education research being done incorrectly, which has an impact on real children in real classrooms.

In my next post, I will deal with the more Mathsy side of things. I will show that their criticisms of significance testing are baseless and just show their poor understanding of Statistics.

In general, it is, in my experience, unwise to assume that you can tell anything about people’s capabilities from their current jobs. Back in the 1980s, I actually taught Robert Coe on a PGCE course, and I seem to recall that he has a degree in mathematics…

Robert Coe does have a Maths degree, he is about the only person involved with the Effect Size who does, though I’d be interested to know how much of it was Statistics.

I have never encountered a mathematician using statistical significance. I have a degree in pure mathematics.

I think it depends on your definition of Mathematician Andrew. Most people who do Maths degrees do all three strands, Pure, Applied and Statistics, I know I did. Similarly for A Level Maths the kids do Pure, Mechanics and Statistics and we do teach significance testing in S2.

on P8 of his book, Hattie states that he uses effect size which is CONTRARY to the usual emphasis on statistical significance. Later on in p18 he states that if he used statistical significance his null hypothesis is very certain to be false. In other words, his work and findings would be useless!!!!