DATELINE: December 06, 2000
Real Statistics for Excel
By Charles Seiter
When Microsoft bought a statistics package from a third party and stuck it into Excel over a decade ago, it was met with disdain. Excel statistics miscalculate standard deviations for large numbers; its t-tests flub missing-value calculations; the function TINV and NORMINV produce errors at extreme values; and the regression routine gives misleading error messages for collinear data. With simple data sets, these problems don’t show themselves; however, Excel statistics isn’t adequate for even a first year statistics class.
Data Desk/XL, from Data Description, fixes the statistics problems in Excel by providing a separate menu of graphics and statistical functions that work with data in Excel spreadsheets. Data Desk/XL delivers basic summary statistics, chart functionality from boxplots to rotating 3-D plots (see “Around They Go”), simple and multiple regression and 1- and 2-way ANOVA, confidence intervals, and the standard assortment of z-tests and t-tests. Also included is a complete menu of tests for quality-control measurements.
The graphics with this add-in are especially helpful — they allow you to lasso groups of points and scale axes with sliders, just as in the complete Data Desk package. All tests and graphics are directed from clear and intuitive dialog boxes, accompanied by a clear and thorough help system. While Data Desk/XL is missing a few advanced functions such as nonparametric tests, it is an excellent tool to accompany first-year statistics. Data Desk/XL is conveniently available by download from
(free 30-day trial; $50 for permanent activation). It supports Microsoft Excel 5.0 (including Microsoft Office 98) and later.
Around They Go
This rotating 3-D plot takes data directly from your Excel worksheet and provides a complete set of Data Desk palettes for detailed investigation.
Note: When you purchase something after clicking links in our articles, we may earn a small commission. Read ouraffiliate link policyfor more details.