This page describes the changes in Release 22.
1. Highlights
● 21 new procedures
● 6 new functions
● random permutation tests for the fixed effects in a REML analysis (VPERMTEST)
● Tobit linear mixed model analysis on data with fixed-threshold censoring (TOBIT).
● new procedures (and menus) to follow-up the analysis of a generalized linear mixed model by GLMM: display further output (GLDISPLAY), save information in Genstat data structures (GLKEEP), form predictions (GLPREDICT), plot residuals (GLPLOT), and do permutation tests to assess the fixed model (GLPERMTEST)
● equivalence, non-inferiority and non-superiority tests (TEQUIVALENCE, TTEST)
● standard curves with common nonlinear parameters across groups fitted by new procedure RCURVECOMMONNONLINEAR to provide s.e.’s for linear parameters (unavailable from FITCURVE)
● analysis of variance with nearest-neighbour adjustments using a Papadakis covariate (APAPADAKIS)
● Markov models for rainfall data (RFFAMOUNT, RFFPROBABILITY, RFSUMMARY)
● 2-dimensional scatter plots with confidence, prediction and equal-frequency ellipses (DELLIPSE)
● reshaping a data set with classifying factors for rows and columns, into a reorganized data set with new identifying factors (RESHAPE)
● functions to check the type of a data structure (GCONSTANTS, TYPE)
● function to copy values from a table into a variate (TVECTOR)
● functions to find the position of the maximum or minimum value in a data structure (MAXPOSITION, MINPOSITION)
2. What’s new
2.1 Directives
None
2.2 Procedures
APAPADAKIS analysis of variance with an added Papadakis covariate, formed from neighbouring residuals.
ARCSPLITPLOT adds extra treatments onto the replicates of a resolvable row-column design, and generates factors giving the row and column locations of the plots within the design.
BPCONVERT converts bit patterns between integers, pointers of set bits and textual descriptions.
DELLIPSE draws a 2-dimensional scatter plot with confidence, prediction and/or equal-frequency ellipses superimposed.
FWITHINTERMS forms factors to define terms representing the effects of one factor within another factor.
GETTEMPFOLDER gets the location of the folder used by Genstat for temporary files.
GLDISPLAY displays further output from a GLMM analysis.
GLKEEP saves results from a GLMM analysis.
GLPERMTEST does random permutation tests for generalized linear mixed models.
GLPLOT plots residuals from a GLMM analysis.
GLPREDICT forms predictions from a GLMM analysis.
GLRTEST calculates likelihood tests to assess random terms in a generalized linear mixed model.
RCURVECOMMONNONLINEAR refits a standard curve with common nonlinear parameters across groups to provide s.e.’s for linear parameters.
RESHAPE reshapes a data set with classifying factors for rows and columns, into a reorganized data set with new identifying factors.
RFFAMOUNT fits harmonic models to mean rainfall amounts for a Markov model.
RFFPROBABILITY fits harmonic models to rainfall probabilities for a Markov model.
RFSUMMARY forms summaries for a Markov model from rainfall data.
TEQUIVALENCE performs equivalence, non-inferiority and non-superiority tests.
TOBIT performs a Tobit linear mixed model analysis on data with fixed-threshold censoring.
VPERMTEST does random permutation tests for the fixed effects in a REML analysis.
VSUMMARY summarizes a variate, with classifying factors, into a data matrix of variates and factors.
2.2 Functions
GCONSTANTS provides type numbers of Genstat data structures.
LEVELS forms a variate containing the levels of the factor.
MAXPOSITION finds the position of the first instance of the maximum value in a numerical structure.
MINPOSITION finds the position of the first instance of the minimum value in a numerical structure.
TVECTOR copies values from a table into a variate, allowing margins to be included or excluded, and a list of its classifying factors to be specified to define the order in which the values are taken.
TYPE gives the type number of a data structure.
3. What’s changed
Most of the changes are compatible with Release 21, the previous release. There are a few commands, however, where new options or parameters have been inserted into the existing lists. These may cause problems in statements where option or parameter names have been omitted or abbreviated (see Section 1.7.1 of Part 1 of the Guide to the Genstat Command Language for details). To avoid any difficulty, the name of the option/parameter after the new option/parameter should be given explicitly, and not abbreviated to fewer than four characters.
Any command, where changes in Release 22 may cause incompatibilities in existing programs, is marked in Sections 3.1 and 3.2 by the symbol †. The full details are given in Section 3.4.
3.1 Directives
FDELETE now supports the * and ? wildcards in the file name.
SET now allows you to control whether seconds are included with the time12 and time24 date representations.
SPLOAD now allows you to control whether the directory details are included with the name of the file in the catalogue.
SUSPEND has a new option MINIMIZE to minimize the console window.
3.2 Procedures
†AGRAPH, †AUGRAPH, DTABLE and †VGRAPH now allow you to specify details of the axes, and to control whether the factor names appear in the key descriptions, plot titles and page titles.
APPEND now allows values from mixtures of variates, factors, scalars and texts to be appended into a compatible new vector.
†ASPREADSHEET and AUSPREADSHEET can now save LSDs of means.
†GLMM can now display fitted values, residuals and deviances, and provides more control of the underlying REML algorithm.
IMPORT has a new XLSCONTENT option to specify what to import from an Excel XLSX file.
LSIPLOT has new parameters to specify the symbol to be used to plot the estimates, the colour and size of the symbol, and the size of the labels.
PLS has a new parameter FTEST to save the results from the Osten F test.
†MANNWHITNEY has a new print setting, hodgeslehmann, to draw attention to the fact that it is the Hodges-Lehmann estimate that is used for the difference in locations of the two samples. The estimate can be saved by the new HODGESLEHMANN parameter.
†RNEGBINOMIAL can now fit the terms individually, and provide confidence intervals for the estimates.
RPERMTEST can now permute the individuals from binomial observations. You can also save the summary and accumulated tables.
R2LINES now has a METHOD option, to allow a different optimization method to be tried in case of convergence problems. Also, the default number of grid lines for the initial search has been increased from 30 to 100.
TTEST can now do equivalence, non-inferiority and non-superiority tests.
†VFPEDIGREE can now form the inverse pedigree matrix, so that you no longer need to call VPEDIGREE yourself afterwards.
3.3 Functions
In previous releases, the table functions TSUMS, TTOTALS, TMEANS, TSD, TSEMEANS, TSKEWNESS, TKURTOSIS and TVARIANCES would generate a missing value for the result if there was a missing value in any of the cells that contribute to the result. Conversely the corresponding summary functions (SUM, TOTAL etc) have always ignored missing values. The table functions now ignore missing values, other than that the result will be missing if all the contributing cells contain missing values.
3.4 Incompatibilities
AGRAPH, AUGRAPHand VGRAPH
Options KEYMETHOD, PLOTTITLEMETHOD, PAGETITLEMETHOD and USEAXES inserted before SAVE.
ASPREADSHEET procedure
Options LSDMEANS and LSDLEVEL inserted before SPREADSHEET.
AUSPREADSHEET procedure
Options LSDMEANS and LSDLEVEL inserted before SPREADSHEET.
GLMM procedure
Parameter EXIT inserted before SAVE.
MANNWHITNEY procedure
Parameter HODGESLEHMANN inserted before LOWER.
RNEGBINOMIAL procedure
Option POOL inserted before NOMESSAGE, and option PROBABILITY inserted before SEAGGREGATION.
VFPEDIGREE procedure
Option SEX inserted before UNKNOWN.