What I hope to learn at IASE 2017

The theme of IASE 2017: Teaching Statistics in a Data Rich World

11–14 July 2017 the IASE conference is held in Rabat, Marocco. This is where statistics educators from all around the world gather to exchange views on how statistics might best be taught. Here are three things that I hope to learn at IASE 2017:

1. Which topics should we teach and which topics should we stop teaching?

I believe one of the functions of a teacher is to guide a path. When you are learning something new, things can be overwhelming. As a teacher, you can help students focus. But on which topics? In every course there is a finite amount of time (never enough). We need to make choices. So I am specifically interested in: what should we not teach?

What am I talking about? Here are two examples: let’s stop teaching hypothesis testing as our inferential method, or: let’s stop teaching parametric methods and focus on simulation instead.

At this IASE conference, I hope to learn from other educators how they make these choices. I am specifically interested in answering these questions with two types of audiences in mind:

a) first intro to statistics for high-school / university students who are not majoring in math or statistics

How can we introduce statistics as a course that is relevant and interesting for students who often say they hate statistics and think they just cannot do maths?

Which intro to statistics topics are essential? Which subjects are less relevant and could be dropped?

Should we explain Z-scores? Would starting with a simulation approach be a more gentle start, so that students can focus more energy on really grasping the key ideas? I’m very interested to hear your opinions on this!

b) basic numeracy for the general public

We believe that numeracy/quantitative literacy — meaning: the ability to work with and interpret quantitative information — should be an essential skill for all citizens. Numbers and statistics are all around us, in business, policy as well as in the news. Yet a lot of people do not have the skills to adequately interpret this information. We believe that just as people have fought for general literacy, it is now time to fight for general numeracy.

But which topics or skills are so important that we may expect everyone to master them? I believe the difference between correlation and causation should definitely be one of them. As well as knowing the difference between absolute and relative risk. What would be a general numeracy curriculum? (PS: I like to read Steen and Schield (statlit.org) contributions on this topic) This ‘curriculum’ should be so critical and small enough so that it can be taught to everyone.

Which basic numeracy/quantitative literacy topics/skills do you think every citizen should master?

2. How can we show the relevancy of data and statistics and make it less abstract?

A few years back, some of my friends were finishing their master thesis. When I asked how things were going, they said they really the content of their thesis. But there was one aspect about it, they just hated: the data analysis and statistics!

My friends found it abstract, unclear and they did not really see the use for them personally of mastering these skills. 
Now data and statistics are all around us, and in almost every domain, employers pay high salaries for people who can work and interpret data correctly. 
So I believe this is a failure on our part as statistics community: we fail to communicate well enough why, how and where statistics is relevant.

But what are good methods to attract the interest of students? How can we make them curious? How we can teach statistics and data analysis in such a way that students feel empowered by these skills?

3. How can we best assess whether students master topics and skills?

At the organisation I co-founded — I Hate Statistics, we always try to reflect on the specific learning goals that are being taught and assessed. For us this is particularly important as we connect each exercise to one or more specific learning goals — we call Knowledge Components (KCs).

We believe it is important to be good at assessing whether learners master specific knowledge components: it is a great opportunity for honest self-assessment and feedback on learning.

What we noticed though, and some readers might recognize, is that some knowledge components are easier to assess than others.

Assessing whether someone knows how to compute a Z-score is quite straightforward: just check if a learner gets the correct answer. If you generate the exercises based on variables (as we do), you are able to generate an almost infinite amount of practice exercises.

Assessing whether someone really understands what a p-value is and isn’t is much harder: what type of ‘conceptual’ questions can you ask? How do you make sure that a learner does not just memorise the answers yet fails to grasp the underlying principle?

How can we best assess ‘conceptual’ topics and skills? Which type of questions can we use?

These are the three questions I hope to find answers on. Please comment or email me if you have views on these questions or are interested in finding answers to these questions as well. I would really appreciate it!

Of course, I’ll probably discover additional question I don’t know about yet. In any case, I’m sure I’ll meet lots of interesting people and I am looking forward to meet as much of them as I can, and learn as much about statistics education as possible.

To IASE 2017!