UniServe€Science News: Newsletter of the Science Software Clearinghouse Vol. 3, March 1996

Teaching data analysis techniques

Les Kirkup

Introduction

Fundamental advances in the physical sciences rely heavily on the insights that carefully designed experiments can provide. Much insight is gained by extracting, evaluating and analysing data from experiments. We want our students to develop confidence and facility in handling experimental data early in their undergraduate careers. Equally, we recognise that teaching data analysis techniques represents something of a challenge. In the first place students often perceive the study of such topics as `outside' the mainstream of the courses "I want to be a chemist - I don't want to play with numbers!". Additionally, much material available for teaching analysis techniques describes contexts which have little to do with science and, due to the overseas origin of the material, sometimes reveals a distracting national `slant'.

Development of an integrated teaching package

With the support of a CAUT grant, our aim is to re-evaluate the teaching of data analysis techniques and, in particular, to create a package which exploits contemporary media available in the most appropriate manner for improving teaching and learning. For example, a short video can be an excellent way of introducing a topic and for offering context rich situations in which particular analysis techniques are used. By contrast, a video is generally less appropriate when the minute details of the derivation of some pivotal equation are to be discussed.

A very important analysis technique used in the physical sciences and the first to be treated through this work is that of `least squares' for curve-fitting data points. The various elements that make up the package (entitled Where do you draw the line?) were trialed with students in 1995. The package consists of:

* A video offering an overview of the technique of least squares beginning with an illustration of the comparative worth of qualitative and quantitative knowledge in the physical sciences. The technique is described graphically and non-mathematically. The video is suitable for use in the classroom as well as for self-study by the student.

* A student booklet expanding upon and consolidating details presented in the video. In particular the booklet provides necessary background information and introduces unweighted and weighted least squares as well as the role of residuals in establishing `goodness of fit'. The booklet also acts as a signpost to reference material which will assist in understanding specific points raised and is a resource of useful equations and worked examples. The use of spreadsheets for least squares analysis (with examples and exercises) is also dealt with at some length.

* An instructors booklet containing suggestions for use of the package as well as answers to most of the problems/exercises appearing in the student booklet.

* A disc for PCs containing examples of spreadsheet layouts (using Excel 5) for least squares analysis as discussed in the students booklet as well as data in ASCII format that may be read by any spreadsheet.

Conclusion

As scientists we find the analysis of experimental data an absorbing and rewarding occupation. We wish for our students to experience similar dividends when analysing experimental data and to communicate in a more compelling manner the role that data analysis plays in the professional life of a physical scientist. Accordingly, we have developed an innovative mixed media package sympathetic to the needs and interests of physical science students. Though the elements mentioned here have been trialed, it is too early for a thorough evaluation of responses to the innovations. First responses from students and staff to the developments have been favourable, but a long term evaluation of the impact on learning and teaching in this important area remains to be done.

Les Kirkup
kirkup@phys.uts.edu.au

Les Kirkup is a Senior Lecturer in the Department of Applied Physics at the University of Technology, Sydney.

It is anticipated that the package will be available for evaluation by May 1996.