Statistical Procedures for Students

Presumably software houses are running out of superlatives and acronyms for their products, so it's no surprise that a workaday little statistical package for students of business studies from Sapphire merely receives the accolade of Superstat.

This package allows the user to create data sets either as disc files or as data statements - the program being written in Basic using either ungrouped or frequency data format.

The data can then be edited and a variety of statistical procedures applied to them, including the usual descriptive statistics - mean, median, range, standard, deviation and so on - together with histograms and cumulative normal distribution (ogives) for frequency data.

Under the heading of Forecasting, the package can apply a variety of time-series an smoothing techniques to sequential data.

Random deviates from a variety of statistical distributions can be generated, and the chi square goodness-of-fit test is available for testing a variety of null hypotheses, though surprisingly not that of the uniform distribution. Correlation and simple linear regression can be calculated for pairs of data.

The package has a number of good features, but with any interactive statistics program, especially one intended for students, ease of data input and editing is absolutely vital.

There really are too many short-cuts in the programming here. For example, there's no facility for editing an item of data as you type it in - hitting any non-numeric key, gives an error and you start the item again. Also there's poor format ting of data when it's listed on the screen for checking.

Where there is more than one screenful of data you have to wait until all the screens have been presented before you edit an item (which could have been the first one).

Other drawbacks include limited data sets - 200 obser vations or 70 pairs - and lack of a graphics dump facility (you have to add your own PROCedure).

However I did like the various graphs that were produced, and the program does bring together an interesting set of statistical procedures which business studies students should find very useful.

Robert Sykes