KEEPING THE MOMENTUM : SUSTAINING INTEREST IN
STATISTICS LECTURES FOR NON-STATISTICIANS
Angie Wade, Vicki Aldridge, Eirini Koutoumanou
Centre for Applied Statistics Courses (CASC), UCL Institute of Child Health & Great Ormond Street Hospital for Children NHS Foundation Trust,
London UK (Email contact: awade@ucl.ac.uk)
THE ‘BROKEN’ LECTURE
INTRODUCTION
We have developed an extensive programme of training in
statistical methods for non-statisticians. This includes iBSc and
MSc modules, as well as short courses for PhD students and staff in
our local university department and attached paediatric hospital.
Topics range from basic introductions in significance testing and
confidence intervals through to one-day courses on more complex
methods including GLMs, Bayesian analyses and Imputation
methods.
For all of our attendees, statistics is not their primary interest and
they often arrive fearful of the topic. It is seen as a necessary evil or
hurdle to be surmounted as quickly as possible.
In this poster, we show the elements of one of our typical ‘broken’
lectures. We identify how this approach improves student
comprehension and application of statistical concepts compared to
a traditional ‘flat’ format and give student comment on this.
AIMS
Our aim has been to get the participants involved, to make them
feel at ease with statistics and to leave with a proper understanding
of the methods rather than just knowing how to implement them or
which computer buttons to press.
We aim to do this by giving lots of opportunities for us to identify
where there are problems. We get to know the audience and give
them the chance to ask questions or voice concerns without the
need for everyone to hear.
Using sweets to
understand different
sampling techniques
in Uganda
The usual approach to teaching statistical methods is to have a lecture followed by practicals, which may be done within smaller groups. We
change this format into one long ‘broken’ lecture with breaks created by some combinations of:
• Refreshment breaks
• Voting pad sessions
• Interactive work with screen pens
• Data generation exercises
• Written individual practical activities
• Group practical work and feedback
Data generation and tea!
AN EXAMPLE: STATISTICS WITH R
A THREE DAY COURSE
In this course we aim to give the students a basic understanding of
study design, a thorough grounding in confidence intervals, p-values
and significance tests, plus make them proficient in using R to
analyse their data.
We enrol about 25 per class and this will contain a wide variety of
backgrounds and aptitude. Some will fare better with the R side and
some with the theoretical side of the stats. With 2 course facilitators
we aim to identify those with problems and help keep them on track.
With such a large amount to cover in the 3 days, and later
information building on earlier, it is vital that we recognise problem
areas early.
Working in randomly
selected pairs
Computer practicals with
helpers joining in
RESULTS AND SUMMARY
We find that the format used keeps both ourselves and the students
engaged with the learning process. Since we run many courses
many times a year, it may be difficult to sustain interest and
enthusiasm for ourselves, let alone the students, if using a standard
lecture followed by practical format.
Student assessments of the sessions are consistently positive. With
quotes such as:
“Thought the notes were good, availability of notes online and practicals.
Pace of the course good. Voting useful (knowing I wasn't the only one who
didn't get 100% of what was covered 100% of the time).”
In the three days of ‘Stats with R’ we use a mixture of:R demonstrations, R practicals, voting pads, analysis of generated
data, interactive pen, PowerPoint presentations, written short
practicals done alone, in pairs or in groups.
“The course was excellent and very well taught, I was very happy with the
way the lectures were taught and the use of the
response pads were a great way of making sure we understood.”
We have shown our ‘broken’ lecture technique to be valuable in
transmitting information and in sustaining the interest of nonstatisticians when learning statistical techniques.