Spreadsheet statistics.

Four spreadsheet templates are presented to carry out routine statistical analyses with relatively small amounts of data: two-sample Student's t test, one-factor analysis of variance with replication, two-factor analysis of variance with replication, and Wilcoxon's non-parametric two-sample T-test (Mann and Whitney U-test). The use of sometimes complex IF and logic (AND, OR) functions in these templates is a means of maintaining a neat output presentation by avoiding error values, selection of appropriate forms of the t-test, and iteratively assigning mean ranks to data in the non-parametric test. The advantages of the spreadsheet approach accrue through using a multi-purpose package rather than dedicated statistics programs, having usually unobstructed access to the program for modification, relative ease of programming spreadsheets, and ease of correction and alteration of input data. Disadvantages are the limited size of templates and the small number of templates available. Spreadsheets represent a declarative form of programming rather than procedural, with consequent advantages in specifying the problem to be solved, which may have particular application in teaching statistics and programming at undergraduate level.