The design of a user friendly statistical software package for the IBM personal computer and compatibles written in Turbo Pascal

Abstract

The design and features of a user friendly statistical software package for the IBM personal computer and IBM PC compatibles (PC VSTAT) is described. The software package, written in Turbo Pascal, requires an IBM Personal Computer or IBM PC compatible computer with either two floppy disk drives or one fixed (hard) disk and one floppy disk drive, and a line printer to obtain hard copies of the output. The most important feature of this statistical software package is the use of data vector files as storage for the experimental observations. The use of PC VSTAT does not require any previous programming experience and uses a menu format for selection of options. A number of special programming techniques were used to facilitate user input and to make the software package user friendly. PC VSTAT permits the user to name, create, delete, and edit data vectors, list data, sort data, manipulate data mathematically, produce frequency tables and histograms, perform linear regression and correlation, perform nonlinear (quadratic and cubic) regression, produce scatter plots, generate random sampling sequences and generate random numbers, and perform a number or statistical hypothesis tests such as t-tests, F-tests, Chi-Square tests and Kolmogorov-Smirnov tests. Output may be displayed on the screen or directed to a line printer. PC VSTAT uses algorithms to compute the probability values for the corresponding values of the test statistics. User acceptance has been excellent with much positive feedback concerning the speed at which calculations are completed (a feature of the Turbo Pascal programming language) and the friendly interface that exists between the user and the computer via the menu format. This package is well suited for the user in an introductory statistics class and for the analysis of research data which requires basic data manipulation, descriptive statistical measures, simple hypothesis testing and goodness-of-fit testing.