Statistical Project Work:

Since it is the ability of the statistician to apply his subject in practice which makes him a statistician rather than a mathematician, and the ability to cope with real world situations develops from professional experience and a maturity of outlook, the inclusion of project work as part of the statistics curriculum is seen as important by the Department of Statistics at the University of Pretoria. It is compulsory in the final (third) year of the undergraduate study.

Project work comprises a combination of statistical theory and practice. it involves the planning, execution and presentation of a practical statistical investigation under the guidance of a tutor but performed independently by each student. Studies of a theoretical nature or which would be based on artificially (e.g. computer) generated data are not acceptable.

Students are randomly assigned to tutors, in the ratio of one tutor for approximately five students. The exercise proceeds through the following stages:

Students are expected to select their own topics for the investigation according to their personal interests. This provides a more stimulating approach and an increased motivation for the exercise. Indirectly students are thereby also forced to establish relationships between natural phenomena and statistical models. The allocation of topics from a reserve list is discouraged. Projects related to hobbies or to contentious campus matters seem to be very popular. Students are also expected to work individually rather than in groups.

In order to discipline students in a precise formulation and constructive planning of their projects, a memorandum containing details of the intended investigation must be submitted beforehand to the relevant tutor for comment and approval. Comments at this stage are directive and aimed at assisting each student to formulate clearly a meaningful and feasible investigation within the framework of the available time and facilities. When both parties are satisfied that the project is realistic, approval is granted for the exercise to proceed. Discussions between tutor and student continue during the investigation.

Students collect their own data, thus exposing them to the difficulties often encountered in practice. They must also satisfy themselves of the credibility and relevancy of the data for the purpose of the investigation. Our experience has been that students’ attitude towards project work is very often negative at the outset, but that enthusiasm usually builds up while the investigation is in progress. After the first obstacles have been surmounted they become so involved that considerable capital investments are made voluntarily and even holidays are sacrificed to do the field world. In one instance a student personally interviewed 220 housewives randomly sampled from an area of about 400 kin², truly a major operation! All analyses and interpretations thereof have to be performed by the students personally. This serves as a test of their ability to cope with the posing of a specific problem, choosing and performing a suitable analysis and interpretation of the results.

The whole project must be summed up in the form of a project report, of approximately 20 to 30 typed A4 pages. Students normally find this a difficult task, but it is necessary in order to develop the skill of technical communication. Preciseness is very important. Phrases such as "estimated", "on the average", "significantly", "projected", etc. have a statistical connotation and must be used in their correct scientific context. The report should be technically complete without becoming a thesis. One copy of the report goes to the departmental library. We have had some very well prepared and even colourfully illustrated reports from students taking a great pride in their work. Very frequently it is the less able student, as far as theoretical class work is concerned, that puts a greater effort in project work and produces a better end product.

Each student is given 10 to 20 minutes to present his/her report in the class, followed by a short discussion. The obvious objective is to develop the student’s ability to converse on a technical issue and to defend the manner in which the investigation was handled.

For the evaluation of project work a marking scheme is employed with the following loading:

10% for the choice of the topic, particularly in respect of originality, practical and analytical feasibility and depth of the subject.
15% for the planning of the investigation, taking into consideration the preparation of the memorandum, formulation of the statistical objectives tailor-made methods (e.g. in questionnaire design) and general organisation.
35% for the execution of the investigation, which comprises the field work, analysis of the data and interpretation of the results and, in general, the total personal effort of the student.
30% for the report in respect of exposition, style, language and technical care.
10% for the class presentation with emphasis on explicitness, correctness of scientific expression, self-confidence, and enthusiasm.

Under the above marking scheme 60% of the marks are awarded for statistical proficiency and 40% for communication of the results of the exercise.

The value of a project exercise does not lie in its confirmation of an anticipated outcome. Projects which fail to meet with anticipations present a challenge to the student (and very often also to the tutor!) to find reasons for the unexpected result. A student who happens to choose such a topic certainly has a very valuable lesson in the objectivity of statistics.

The tutor benefits from the exercise in that his scientific insight, knowledge and experience is also enlarged. This is particularly valuable to younger lecturers.

In our department, project work has already contributed as much as 15 per cent of a student’s final mark. In addition a prize is also awarded annually for the most deserving project.

From a variety of projects already undertaken by students in our department, we quote the following interesting titles:

* Patterns in the thefts of motorcars in a suburb of Pretoria

* The impact of advertisements over television and the radio

* Students’ rag: Is it worthwhile?

* An analysis of biorhythms

* The psychological strain on gold mine workers

* The ability of people to estimate weights, areas and volumes

* Differences in the personalities of smokers and non-smokers

Dolan, 0. (1979). Learning Statistics Through Project Work. Teaching Statistics,1, 34—41.
Griffiths, I. D. and Evans, B. E. (1976). Project Work in Statistics. The Statistician, 25, 117—123.
Kanji, G. K. (1979). The Role of Projects in Statistical Education. The Statistician, 28, 19—27.
Scott, J. F. (1976). Practical Projects in the Teaching of Statistics at Universities. The Statistician, 25, 95—108.