Release 24: new features

This page describes the additions and changes to procedures and directives in Release 24.

1. Highlights

• 23 new procedures
• 4 new functions
• Analysis of variance for censored data (ATOBIT, AUTOBIT)
• Analysis of censored regression analyses with Normal, gamma and negative binomial data (RGTOBIT, RNTOBITRNBTOBIT)
• REML analysis of field trials using a two-dimensional spline model (V2DSPLINE)
• Design and analysis of sets of n-of-1 trials (N1ANOVA, N1DESIGN, N1PLOT, N1SIMULATE, N1TTEST)
• Negative binomial probabilities and random numbers (CLNEGATIVEBINOMIAL, CUNEGATIVEBINOMIAL, EDNEGATIVEBINOMIAL, PRNEGATIVEBINOMIAL, ELNEGBINOMIAL, EUNEGBINOMIAL, GRANDOM)
• Gamma distribution expected values (ELGAMMA, EUGAMMA)
• Improved default settings for the symbols, colours and line-styles used in graphs
• At the start of each job, the server now loads the default graphics environment that is specified in the computer’s Registry, and that Registry setting can be accessed and defined (GET, SET)
• Heatmaps (DBANDCOLOURS, DHEATMAP, FHEATCOLOURS)
• Graphical displays of data in a table, so that you can create PDF reports that include graphs (DPRINT)

2. What’s new

2.1 Directives

None.

2.2 Procedures

ATOBIT uses the Tobit method to perform analysis of variance with censored data.

AUTOBIT uses the Tobit method for analysis of variance of an unbalanced design with censored data.

DBANDCOLOURS plots a band of colours, for example, to provide a key for a heatmap.

DEQUATION plots an equation on a graph.

DHEATMAP plots a heatmap.

DPRINT creates a graphical display of data in a table.

DSCOLOURS selects colours for graphics using colour brewer palettes.

ELGAMMA calculates expected values of the lower parts of gamma distributions.

ELNEGBINOMIAL calculates expected values of the lower parts of negative binomial distributions.

EUGAMMA calculates expected values of the upper parts of gamma distributions.

EUNEGBINOMIAL calculates expected values of the upper parts of negative binomial distributions.

FHEATCOLOURS forms colours from the values in a variate to use in a heatmap.

FMARGINALTERMS determines the marginal terms for the terms in a model formula.

N1ANOVA does an analysis of variance with data from a set of n-of-1 trials.

N1DESIGN forms designs for sets of n-of-1 trials.

N1PLOT displays summaries of data from a set of n-of-1 trials.

N1SIMULATE simulates data for a set of n-of-1 trials.

N1TTEST performs t-tests with data from a set of n-of-1 trials.

RGTOBIT uses the Tobit method to fit models to censored gamma data.

RNTOBIT uses the Tobit method for regression with censored data.

RNBTOBIT uses the Tobit method to fit models to censored negative binomial data.

TURNBULL calculates the non-parametric maximum likelihood estimator (NPMLE) of the survival function for interval censored data, also known as the Turnbull estimator.

V2DSPLINE models spatial trends in a field trial using a two-dimensional spline model.

2.3 Functions

CLNEGATIVEBINOMIAL cumulative lower probability for a negative binomial distribution.

CUNEGATIVEBINOMIAL cumulative upper probability for a negative binomial distribution.

EDNEGATIVEBINOMIAL equivalent deviate for a negative binomial distribution.

PRNEGATIVEBINOMIAL probability for a negative binomial distribution.

3. What’s changed

Most of the changes are compatible with Release 23, 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 24 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

AKEEP can now save results from the unadjusted analysis in an analysis of covariance, or from the analysis of one of the covariates.

At the start of each job, Genstat now loads the graphics environment that is defined in the computer’s Registry. The Registry setting can be set by the GRAPHICSENVIRONMENT option of SET (as well as by the Graphics Environment menu). You can save the Registry setting in the ‘graphicsenvironment’ element of the pointer saved by the ENVIRONMENT option of GET. The currently loaded graphics environment can be saved by the GRAPHICSENVIRONMENT option of GET. (This may differ from the Registry environment, if DLOAD has been used in the interim.)

3.2 Procedures

ABLUPS can form means from the unadjusted analysis in an analysis of covariance.

ACHECK can now save the results of the tests.

APOLYNOMIAL can now calculate standard errors and correlations for the coefficients.

AFMEANS can form means from the unadjusted analysis or from the analysis of a covariate in an analysis of covariance.

ASPREADSHEET and A2RDA can now save residuals, replication and BLUPs for block terms.

A2RDA can save information from the unadjusted analysis or from the analysis of a covariate in an analysis of covariance. It can also save standard errors of residuals and information about large residuals.

BOXPLOT can now colour the boxes.

DDESIGN can now plot the factor labels vertically instead of horizontally, and you can provide colours for the plots e.g. to superimpose a heat-map onto the plan to represent yields or covariates. It can also display the axes (and x- and y-coordinates).

DERRORBAR can now set the thickness of the bar.

DMOSAIC now allows you to supply titles for the lower and upper x-axis and y-axis.

DOTHISTOGRAM now allows you to supply a title for the axis representing different groups or variates.

DREFERENCELINE and DTEXT can now resize the window, to allow for situations where the graph has no spare space.

DXDENSITY can now include means, medians, percentiles and confidence limits for the means or medians, to produce a “pirate” plot.

GRANDOM can now generate random numbers from the negative binomial, Neyman type A and Poisson log-Normal distributions.

SPSHEWHART can now print and save the variates plotted in the charts.

GLTOBITPOISSON, HGTOBITPOISSON, RTOBITPOISSON and TOBIT can now analyse data with both left and right censoring, and where the bound varies across the data set.

3.3 Functions

None

3.4 Incompatibilities