Statview 512+. The Professional, Graphic, Statistics Utility

Statview 512+ is a powerful statistical analysis tool, written for the Macintosh computer, that combines the strengths of high-level statistical packages with the simplicity of the Macintosh interface. The program brings an extensive range of statistical functions into the hands of researchers and students who might otherwise not have access to this level of statistical computing. It is easy to use, well documented, and, most important, is numerically precise. Although Statview is not primarily designed to serve as a data manager, the key to the program lies within its access to data. Its intuitive combination of graphics and dialogue boxes allows the specification of variables for analysis in a remarkably simple way. Data sets are displayed in spreadsheet-like windows that show the variable names as column entries and individual records as rows. In order to specify a particular variable for analysis, one merely selects the variable of interest as "X" and proceeds to choose the appropriate statistic. For procedures that require more than one variable (such as linear regression) the dependent variable is designated as "Y," and independent variables are labeled with a subscripted "X." Statistical procedures may then be applied with a click of the mouse. The output may be displayed in one of several formats, including a tabular layout or a more intuitive graphical representation. The data window also facilitates the most powerful feature of Statview 512+: interactive data analysis. By manipulating data in the data window, one can immediately see the effect of the change on the results. Variables may be included that were omitted in a previous analysis, or outlier observations may be deleted from the analysis. New results are calculated immediately and replace previous output. This feature is particularly useful for selecting appropriate regression models. Although many statistical packages claim to perform analyses interactively, Statview 512+ sets the standard by which to measure all others. The range of statistical procedures supported by Statview 512+ is sufficient to satisfy most biomedical data analysts. Descriptive statistics include computations of the mean, standard deviation, kurtosis, and skewness of a group of observations and also permit the description of the distribution of a variable around its mean. Statview also provides frequency distributions for both continuous and categorical data. Comparisons of two means can be accomplished with either a paired or unpaired t-test, and more than two means may be tested in analysis of variance (ANOVA) models. Statview performs well …