Case Study: Statistics Laboratory

Peter Petocz, Mathematics and Computing Sciences

Previous mode of presentation (4 hrs Lecture and tutorials) not addressing students needs.

"I felt the experiences the students were getting were not congruent with what one does as a statistician - this couldn't be done in a lecture format."

Subject extended to include 2 hour Laboratory component. "it was important to have regular sessions where students actually did statistics - meaning sitting with a computer and working."

Importance of using and communicating data.

"It involves students in playing with data - they are analysing, finding pictures of data, using data to come up with decisions and writing about them."

 Each laboratory takes 1 - 2 sets of data.

"Students are asked introductory questions and carry out analysis, it's guided exploration of a set of data and often open ended."

Encourage reflective writing.

"students hand the laboratories in (ten questions) - some have to be encouraged to write more, others to write less - they need to write something reflective rather than just doing a complete screen dump."

Simple marking scheme: 0, 1, 2, 3.0 = Not handed in1 = Incomplete

"With such large classes I'm not so interested in the marking as in seeing that they actually do it and giving them some guidance as to the level that they're working, making suggestions, keeping them focused."

Constructive feedback

"Generally tutors simply decide between 2 and 3 - I advise tutors which specific questions need to be looked at - they give a simple mark and write a couple of constructive comments."

"I prefer students to get a small amount of feedback for each of the ten labs, rather than using only one or two pieces of work where students get a large amount of feedback."

Flexibility in choosing laboratories and mode of attendance.

"They can choose to do or not, plus they're only marked on the best eight out of ten, this avoids confrontations about sickness."

"There's also flexibility of choosing not to do the lab during lab times. They can buy a package and do them at home, or can stay in the evening."

Problem-based approach.

"Most statisticians used to define statistics as a branch of pure mathematics. I like to define it as Numerical Detective Work and this needs fieldwork practice."

Reduced lecture hours.

"by utilising tutors."

Able to provide input to both tutors and students.

"I have time to go from group to group, posing relevant questions to encourage students to think in a different way."

Stay closer in touch with students.

"I like to see the way they're thinking, this is very transparent when you can observe them working at a computer."

 Become familiar with individual students

"I prefer to work with people, answer their questions rather than lecturing to them - you get to know the majority of names and faces."

 Learning about 'real life' communication.

"not just a numeric answer, but giving an explanation."

Replace something normally done in a lecture with something that is 'value added'.

"Work out a scheme for reducing your lecturing from 2 to 1 hour a week. Prepare outline notes/study guide."

"Something which relieves you of the responsibility of 'saying everything' in the lecture."

Demonstrate your professional role.

"use the freed-up space and time to show students what you do as a professional - laboratory, role-play, case studies, etcetera."