William M. Grove's Downloadable Programs/Source Code

For more types and other versions of taxometric software (written by Leslie J. Yonce and/or Niels Waller), visit Niels Waller's Taxometrics Home Page. Note: versions of Yonce/Waller taxometric programs, rewritten by Grove and available for download here, are renamed. Example: Yonce/Waller program "MAXCOV.prg" or "MAXCOV.s" or "MAXCOV.r" after Grove rewrote it, became "MAXCOVwmg.r". This is to prevent mixing up versions. Program source code filenames here all end with ".r". Extensive comments at the top of Grove's source code files make clear who the program author is, what the date of the last revision is, how program works, etc.

All taxometric programs from Yonce/Waller/Grove run in R (version 1.6 or later); this is available under the GNU Public License for Linux, Mac, and Win32 at The Comprehensive R Network. Some of the code may run in Insightful Corp.'s S-Plus ^TM language and environment, but this is not guaranteed; significant code tweaking and/or debugging may be necessary if you want to run the code in S-Plus. Such use is not supported by me (i.e., do not contact me with problems related to use in S-Plus.)

For bug reports and questions about Grove's versions of taxometric routines (e.g., "MAXCOVwmg.r"), or any other routines on this page, contact William M. Grove. For bug reports and questions about Yonce/ Waller routines, contact Niels Waller.

To download code in Internet Explorer:
right click on link, then choose "File Save As...":

The MAXCOV, MAMBAC, and MAXEIG routines were written by Leslie J. Yonce and/or Niels Waller, and rewritten by Will Grove. This was to get more consistent style, thorough inline ocumentation of code, a few bug fixes, and change in call syntax (which makes routines easier to use in Monte Carlo simulation study runs&emdash;where routines are called noninteractively and all user results are passed back in the form of data structures rather than on-screen displays).

• MAXCOVwmg Taxometric Procedure written by Yonce, amended by Waller, rewritten by Grove.
• MAMBACwmg Taxometric Procedure written by Yonce, amended by Waller, rewritten by Grove. Only Grove version delivers full parameter estimates..
• MAXEIGwmg Taxometric Procedure written by Waller, rewritten by Grove. Waller & Grove versions estimate only the P (taxon base rate) parameter.
• MAXSLOPE Taxometric Procedure by Grove. Implements procedure described in his pubs 75, 114.
• MCLUSTwmg routine. R code wrapper for MCLUST() library function, which is (in part) Gaussian finite mixture model. See MCLUST documentation for details.
• Sample code to extract various statistics from output of MAXCOVwmg, MAMBACwmg, MAXEIGwmg, MAXSLOPE, and MCLUSTwmg (i.e., parameter estimates needed for consistency tests, plus statistics likely to be summarized in Monte Carlo simulations)
• Sample code to calculate certain basic non-graphical consistency test statistics, for a five-variable data set to which MAMBACwmg and MAXCOVwmg have been applied.
• Sample code to calculate a statistic for testing whether Bayes posterior probabilities of taxon membership have U-shaped distribution, obtained by fitting two-parameter beta distribution to these probabilities and then quantifying inverse of min(shape1, shape2), where these are the two shape parameters of beta distribution.
• Sample code to calculate two kinds of statistics for measuring the resemblance between an empirical MAXCOV (conditional XY-covariance) curve and one expected under conjecture that there is a taxon. Shape and height statistics are returned. See code for explanation of how these values are justified and calculated. Statistic giving shape resemblance to curve predicted from single common linear factor model is returned also.
• Sample code to calculate two kinds of statistics for measuring the resemblance between an empirical MAMBAC (conditional Mean Above Minus Below A Cut) curve and one expected under conjecture that there is a taxon. Shape and height statistics are returned. See code for explanation of how these values are justified and calculated.
• Sample code to calculate two kinds of statistics for measuring the resemblance between an empirical MAXSLOPE dy/dx (conditional slope) curve and one expected under conjecture that there is a taxon. Shape and height statistics are returned. See code for explanation of how these values are justified and calculated. Statistic giving shape resemblance to curve predicted from single common linear factor model is returned also.
• R function code to calculate disattenuated kappa coefficient for agreement between classifications from two taxometric (or mixture model) programs. Implements procedure described in .
• Mathematica code to perform symbolic and numerical integration for publication <"Correction and extension of Wainer's (1976) 'Estimating coefficients in linear models: It don't make no nevermind'" (Zip file)
• Fortran 90 code to perform Monte Carlo numerical integration for publication <"Correction and extension of Wainer's (1976) 'Estimating coefficients in linear models: It don't make no nevermind'" (Zip file)