Plots one- or two-way tables of regression estimates (R.W. Payne).
Options
GRAPHICS = string token |
Type of graph (highresolution, lineprinter); default high |
|---|---|
METHOD = string token |
What to plot (estimates, lines); default esti |
XFREPRESENTATION = string token |
How to label the x-axis (levels, labels); default labels uses the XFACTOR labels, if available |
PSE = string token |
What s.e. to plot to represent variation (average, individual); default aver |
SAVE = regression save structure |
Save structure of the analysis to display; default * shows the most recently fitted regression |
Parameters
XFACTOR = factors |
Factor providing the x-values for each plot |
|---|---|
GROUPS = factors |
Factor identifying the different sets of points from a two-way table of estimates |
XVARIATES = variates |
X-variates for regression coefficients or pointer |
NEWXLEVELS = variates |
Values to be used for XFACTOR instead of its existing levels |
TITLE = texts |
Title for the graph; default defines a title automatically |
YTITLE = texts |
Title for the y-axis; default ‘ ‘ |
XTITLE = texts |
Title for the x-axis; default is to use the identifier of the XFACTOR |
Description
RDESTIMATES plots tables of estimates from a regression analysis. By default the estimates are from the most recent regression, but you use the SAVE option to specify the save structure (from a MODEL statement) from some other analysis.
The XFACTOR parameter indicates the factor against whose levels the estimates are plotted. You can also specify a second factor, using the GROUPS parameter, to plot a two-way table of estimates. A separate set of points is then plotted for every level of GROUPS.
By default, the estimates will be for the model term XFACTOR (if GROUPS is not set) or XFACTOR.GROUPS (if GROUPS is set). You can also specify one, or more, variates for the term, using the XVARIATES parameter. If XVARIATES is set to a single variate, xvar say, the term will be XFACTOR.xvar or XFACTOR.GROUPS.xvar (representing regression coefficients for xvar). Alternatively, it can be set to a pointer containing several variates, for example x1var and x2var. The term will be then be XFACTOR.x1var.x2var or XFACTOR.GROUPS.x1var.x2var (representing regression coefficients for the product of the variates x1var and x2var).
The NEWXLEVELS parameter enables different levels to be supplied for XFACTOR if the existing levels are unsuitable. If XFACTOR has labels, these are used to label the x-axis unless you set option XFREPRESENTATION=levels.
Usually, each estimate is represented by a point (using pens 1, 2, and so on for each level in turn of the GROUPS factor). However, with high-resolution plots, the METHOD option can be set to lines to draw lines between the points. The GRAPHICS option controls whether a high-resolution or a line-printer graph is plotted; by default GRAPHICS=high.
The PSE option specifies how to represent the variability of the estimates, as follows:
average |
plots an error bar showing the average standard error of the estimates; |
|---|---|
individual |
plots a bar around each estimate showing plus and minus its standard error. |
The TITLE, YTITLE and XTITLE parameters allow you to supply titles for the graph, the y-axis and the x-axis respectively.
Options: GRAPHICS, METHOD, XFREPRESENTATION, PSE, SAVE.
Parameters: XFACTOR, GROUPS, XVARIATES, NEWXLEVELS, TITLE, YTITLE, XTITLE.
Method
RDESTIMATES uses the GET directive, if necessary, to obtain the regression save structure, and RKESTIMATES to obtain the tables of estimates.
See also
Procedures: RCHECK, RGRAPH, AGRAPH, DTABLE, VDEFFECTS, VGRAPH.
Commands for: Regression analysis.
Example
CAPTION 'RDESTIMATES example',!t('Probit analysis, see the',\
'Guide to Genstat, Part 2, Section 3.5.'); STYLE=meta,plain
FACTOR [LABELS=!T(Morphine,Amidone,Phenadoxone,Pethidine)] Drug
READ Drug,Dose,Ntest,Nrespond
1 1.50 103 19 1 3.00 120 53 1 6.00 123 83
2 1.50 60 14 2 3.00 110 54 2 6.00 100 81
3 0.75 90 31 3 1.50 80 54 3 3.00 90 80
4 5.00 60 13 4 7.50 85 27 4 10.00 60 32
4 15.00 90 55 4 20.00 60 44 :
CALCULATE Logdose = LOG10(Dose)
MODEL [DISTRIBUTION=binomial; LINK=probit] Nrespond; NBINOMIAL=Ntest
FIT Logdose*Drug
RDESTIMATES Drug
RDESTIMATES Drug; XVARIATE=Logdose