Encouraging more people to think like a statistician underpins much of the work that the Royal Statistical Society does. It is also a goal of Chris Wild, a professor of statistics at the University of Auckland, whose work in statistical education has attracted interest far beyond his home country of New Zealand. His department’s free-to-access, massive online open course (MOOC), An Introduction to Data Analysis, was first launched in October 2014 and was so successful that it’s now running for a second year.
The course aims to ‘teach students how to find the stories told by statistical data’ as well as understanding the limitations of data and how to avoid being misled. It uses visualisation to aid understanding and throughout the eight weeks (3 hours per week) students can go from very little experience with statistical ideas and concepts to someone who ‘thinks like a statistician’.
This visual element, according to Wild, is important to help understand complexities within data. ‘The key difference between this course and traditional introductory statistics is harnessing visualisation and software purposefully designed to accelerate how far beginning students can get with the exploration of multivariate data sets (iNZight), and other visualisation-based software to accelerate conceptual development (VIT),’ he explains. This approach seems to have gone down very well with the online students, according to feedback. ‘I don’t think that previously I’ve ever experienced students talking of “falling in love” with statistics,' he says.
When tasked with the opportunity to build a statistics course by his university, Wild says he ‘immediately saw it as a chance to build something new, a prototype that tried out some fairly revolutionary ideas.’ It also contrasts sharply with common statistical teaching practices, which tend to focus on the theory. ‘ I think that current approaches to lower level statistics are long passed their use-by dates,’ he continues. ‘They lack ambition and proceed far too slowly. Ten years ago Hans Rosling showed us that ordinary people can understand much more complex data stories than we would ever have believed possible – this long before the current hoopla over big data and data science.’
Over the last two decades, much of Wild's work has focused around the teaching of statistics. However, he didn't start out specialising in statistics education. 'I had just been a typical statistical academic, one who took his teaching responsibilities reasonably seriously,' he says. It wasn't until 1996, when he was asked to supervise Maxine Pfannkuch’s PhD, that he was introduced to statistics education research. 'It wasn’t really a supervisor/student relationship', he remembers. 'We were both learning together'. Their 1999 International Statistical Review discussion paper 'Statistical Thinking in Empirical Enquiry' now has more than 830 citations on Google Scholar and is still growing strongly.
He then co-led the University of Auckland's first-year statistics teaching team to a national Tertiary Teaching Excellence Award in 2003 and was President of the International Association for Statistics Education (2003-5). His paper (written with Pfannkuch, Regan and Horton), 'Towards more accessible conceptions of statistical inference', was read to the Royal Statistical Society in 2010 before being published in its Series A Journal in 2011.
Wild hopes that in the future, statistics can inspire as well as inform. 'We need to find ways to convey much more of what the data world has to offer and do it much more quickly, to find ways to awaken the imagination and fertilise it with intimations of what might be possible, and to provide valuable skills for extracting value from data,' he says.
He also believes that more effective teaching is important for the profession itself. 'Teaching unconsciously sends marketing messages to the next generation of decision leaders about who we are and what we good for. The most important future decision-leaders that we teach are students sprinkled amongst the masses in introductory statistics who will not continue with statistics. In these future decision-leaders, our small-vision education can feed “irrelevant people” perceptions.
'My course is an attempt to do something about these things. It may only be baby steps but at least it makes a serious start.'
Chris Wild's free-to-access, massive online open course (MOOC), An Introduction to Data Analysis, runs until December.