SURVEY, for Windows 3.x and above, is available from SPA

When I first received my copy of this package and saw the SPA label I held a great deal of optimism as I have always enjoyed using their well thought out school material. I was not to be disappointed. Survey needs Windows 3.1 or later to run and installed very easily on Windows 95. The package runs as three separate programs, Survey Designer, Survey User and Scatter which need to be accessed seperately from their icons and would demand some knowledge of working in a Windows environment from pupils. Once inside the programs the large clear and simple toolbar is very friendly and would be a good introduction for pupils not familiar working in this computer environment. In each program however there are discrete menus which a teacher, or more advanced pupil, could use to fine tune the program..

The first program, 'Survey Designer', is what it says, a means to set up a sheet on which data could be collected. Whether this would satisfy the National Curriculum requirement on designing a collection sheet is doubtful as the only big decision is whether to use words, single numbers or groups in the left hand columns. The number format option allows discrete numbers or groups to be used. The groups can described in different ways; "0-9" or "0 and less than 9" but sadly there is no option to use inequality notation which is restricting for more advanced pupils (Note: in fact you can; Survey Designer allows you to use any text in describing grouped data; you simply type the separated text you want in the "Text" box in the number format option, so you can have "10 to 19" or "10 - 19" or "10 and <=19" or "10<=L<=19" or whatever, Ed). There are options to assign particular colours or icons from a library given or pasted in from other programs (although the latter is not that easy to scale correctly), and these are used later on to draw charts. This is something I am sure will appeal to pupils. The recording sheet can be printed at this stage so that pupils can go around and collect their own data, though some Network Managers might want to curb the multiple copies option.

In The 'Survey User' option the data is easily entered by clicking in the tally column to create bar & gate tallies. As well as those created in 'Survey User' there are a bank of useful recording sheets for data such as pupil's heights. If not enough space is allowed the program starts to use (10)'s to fit all the tallies in which was slightly messy. The real meat of this program is however in producing suitable charts. Like many who have watched pupils produce very pretty but totally useless graphs with spreadsheet packages, I was keen to see how technically sound the graphs in this section were. When using groups 0-9, 10-19 etc. the bar chart leaves gaps and this worried me, but once the group descriptors are changed to 0.0-9.9 etc the gaps disappear. This I think would make a very useful teaching point about drawing such charts and the need for different graphs in different situations. Again the program is sensible in its choice of charts for qualitative and discrete data, and frequency polygons are plotted against mid-marks. Although some 'silly' charts are possible with continuous data (e.g. 'stick' diagrams plotted against mid marks) the program is generally thoughtful in the use of bar charts. The pie charts have a minor problem in overwriting of labels when sectors are small but the use of a colour scheme chosen in Designer would give students the feeling that they had ultimate control over the appearance of these charts. The printout is disappointing - charts that looked attractive and large on the screen are small and undistinguished as hard copies. I think pupils looking for charts to go in projects would feel let down by these. (Note: users have control over the printed size of their chart by simply changing the size of the chart window, Ed.)
The 'averages' which are produced are sound, in particular with grouped data the 'modal-class' is given rather than attempting to calculate a precise value. What would be nice from a teaching point of view, and to encourage good practice in report writing, is some attempt to show the working in calculating these values.
The last program 'Survey Scatter' is a bit of an anomoly in that it functions completely seperately from the other programs in the suite. Two variables can be entered in columns against individuals names and this can easil;y be extended to more variables. It has a useful tool whereby the teacher can set sensible ranges for values in each column so that pupils cannot enter clearly incorrect values. The scatter diagrams it produces use sensible scales with an option to include (0,0) on any graph. This avoids the problem on so many packages where some manipulation of scales is needed to give a useful graph. There are I think two major weaknesses with this program:-

1. It is not possible to carry out any operations on individual columns as a spreadsheet might do. This program cannot calculate, for example, the mean of a particular variable and the data would have to be re-entered in Survey Designer/User to do this. This singling out of bivariate data as separate from single variable work is clumsy and unhealthy from a teaching point of view.
   
2. There is a 'Statistics' option which give the correlation coefficient and an equation of a line of 'best fit', which it is then possible to draw on the graph. The calculations are way beyond the scope of the pupils it is clearly aimed at and one must view with some horror the production of these figures, as it were, from thin air. This program I felt was almost an afterthought, it didn't tie up with the rest of the suite both from the transfer of data aspect and from the teaching outlook.