Newsletter 12, June 1996
| Newsletter 12, June 1996
Quantitative Methods |
Since the new Windows Versions of SPSS (Ver 6.1.2) and Minitab (10Xtra) have become available on the networks at all sites, occasional problems have been encountered with insufficient memory. With this in mind, we advise that work should be saved regularly (modified data, output and graphs).
Once a graph has been satisfactorily completed and saved, then, to save memory, its window should be closed. (Both Minitab and SPSS allow you to open previously-saved graphs and work on them further if required.)
Although, unlike SPSS, Minitab cannot open previously-produced output, any output saved is in the form of a *.txt file. These files can be opened in any word processor that reads such a file, including Word, and so any required editing can be done from there.
If these steps are followed, your SPSS or Minitab process should keep a working level of memory available.
Although this advice is particularly directed towards students using the PCs in the IT Suites, it may also be applicable to staff working on their own desktop PCs.
Although we are not currently supporting Windows 95, it may interest some of you to know that we understand SPSS for Windows version 6.1.2 does not run under Windows 95. However, apparently SPSS version 6.1.3, although still a Windows 3.x package, does operate correctly with Windows 95.
If you want to acquire this latest upgrade, please contact Computing Services staff in the usual way.
Students often analyse data they have acquired as part of a project. However, unfortunately, in some cases, in order to do justice to the data quite complicated statistical analysis is required.
If any member of staff would like to discuss the outline of potential projects with me to ensure that they are relatively straightforward, statistically, please fell free to get in touch.
(It is always instructive to generate some artificial data of the kind you may expect to collect and attempt to analyse that.)
The Data Preparation Service hopes to acquire Windows-based data-entry software later in the year. This should make the data-entry process a little more straightforward.
The STEPS project brings together seven universities throughout the UK to develop problem based teaching and learning materials for statistics.
The materials produced are based on specific problems arising in Biology, Business, Geography and Psychology. The intention is for students to discover that statistical issues arise from these problems as important natural parts of the process of reaching conclusions. The role of the computer is to assist in the exploration of the problem and to provide support materials for the statistical ideas encountered. Graphical illustration plays a major role. A glossary of statistical terms is provided, and, where appropriate, problem modules can be operated in parallel with standard packages.
Two modules from each subject are currently available on the CITY server, in the STEPS Program Group within Windows. They are available to staff only at present, for evaluation puposes, but if they are found useful their provision for teaching uses is envisaged.
The details of each module are as follows:
The Case of Luddersby Hall
How can we trace the source of a salmonella food poisoning outbreak in a university hall of residence? Students are prompted through aspects of variable selection, hypothesis formulation and interpretation, and the calculation of appropriate test statistics. Contingency tables and chi-square tests are the principal analysis tools. The effects of different choices of response variable are explored.
Skinfold Thickness
Skinfold thickness is widely used as a physiological measurement and has a particular role in the construction of standards for growth of children. Sampling variability is explored in this context. The interpretation and use of confidence intervals are investigated, along with the relationship between sample size and precision. The idea of a reference range is introduced.
Pharmaceutical Behaviour
A Company is located next to a river and is considering introducing a new product. The student is led through an exploration of data sets that include expenditure on advertising, research and development and costs of pollution. Exploratory data-analysis techniques such as stem-and-leaf, box-and-whisker and dot plots are used and lead to consideration of wider issues. On completion, the student will have identified important features of single-sample data sets.
Pharmaceutical Environments
The same Company wishes to include information on the biological effect of pollution in their dissemination of the new product launch. The student is led through a comparison of related data sets using simple extensions of exploratory data analysis techniques applied in the previous module. On completion, the student will be able to compare important features between two or more data sets.
Pennine Rainfall: A Visual Approach
Data are available on annual rainfall over 50 years at different locations across the Pennines. Students are invited to investigate the differences in the characteristics of daily rainfall in relation to altitude and longitude. This involves considering the accuracy of the data and handling different units of measurement. The raw data are simplified and presented graphically as histograms or stem-and-leaf plots. Animation is used to illustrate the construction of these, including the “drag and drop” options for the student. Measures of location and dispersion are used to compare rainfall over the different sites.
Pennine Rainfall: A Numerical Approach
This is a follow-up to Module 1 on “Annual Rainfall”. A fuller investigation into rainfall at the four sites is made using means, medians, quartiles, box plots and standard deviations.
Exploring Dyslexia
Data are obtained on pre-school children and follow-up tests three years later. Scores from 34 children at nursery school were obtained in a variety of tests of vocabulary, motor skills, knowledge of prepositions and use of rhyming. At age 7, these 34 children were given a reading test and their reading ages compared with chronological ages to classify them into “poor” and “normal” readers. A comparison is made of the distribution of “poor”readers within the overall distribution and directly with “normal” readers using stem-and-leaf diagrams, histograms, box plots, means, hinges (quartiles), interquartile ranges and standard deviations. In this way, we try to identify which, if any, of the tests might be used as predictors of poor reading ability. This in turn is a possible indication of potential dyslexia.
Predicting Dyslexia
The background is the same as for “Exploring Dyslexia”. Only the tests identified as possible predictors are used. As in the module “Exploring Dyslexia”, poor reading ability is taken as an indication of potential dyslexia. Correlation and regression techniques are used to see how well any of the tests at age 4 predict Reading Age Deficiency at age 7.
Dr Paul Marchant