This site provides a non-interactive view of several Maple worksheets that have been developed to insure the correctness of the definitions and the accuracy of the approximations for the software being developed to support the AP Statistics functions in the Casio FX 2.0.
HypTests.html displays the definitions and calculations for various hypothesis tests. Most of the examples are taken from the Casio User Guide, section 18-6.
(Last updated 7/31/99)Interval.html displays the definitions and calculations for confidence intervals. Most of the examples are taken from the Casio User Guide, section 18-7.
(Last updated 6/7/99)DIST.html displays the definitions and calculations for the various probability distributions. Most of the examples are taken from the Casio User Guide, section 18-8.
(Last updated 6/7/99)Erf.html displays the approximation procedures and accuracy calculations for the error function, the complementary error function, the inverse error function, the Normal probability density function, the cumulative Normal distribution, and the inverse cumulative Normal distribution. The approximation procedures for the error function and the complementary error function are based on the routines in Cody's specfun package. The approximation procedure for the inverse error function is based on Moler's example. It uses a low order approximation followed by two steps of the Newton-Raphson iteration. The approximation procedures for the Normal functions simply exploit the identities relating the error function and the cumulative Normal distribution. The error curves show that these approximation procedures are more than adequate for the FX 2.0 AP Stat development.
(Last updated 6/8/99)Beta.html displays the definitions, the approximation procedures, and the accuracy calculations for the Beta function, the Incomplete Beta function, the Student t probability density function, the Student t cumulative probability distribution, and the inverse Student t cumulative probability distribution.
(Last updated 7/31/99)Gamma1.html This provides an approximation procedure with better than 15 digit accuracy (relative error) for the gamma function. The procedure uses approximation 5239 from the book Computer Approximations by Hart et. al.
Lgamma3.html This provides an approximation procedure with better than 18 digit accuracy (relative error) for the logarithm of the gamma function. The procedure is based on the routine in W. J. Cody's specfun package.
IGamma.html displays the definitions, the approximation procedures, and the accuracy calculations for the Incomplete Gamma function.
Specifications for AP Statistics Calculations This report provides recommendations on the approximations for the AP Stat functions. The goal of these approximations is to provide robust, efficient approximations which are accurate to 15 digits.
(Last updated 7/23/99)Gallagher Report This report discusses the user interface design from a professional statistician's point of view for a calculator implementation of Two-way ANOVA and Multiple Linear Regression.
(Last updated 6/11/99)