Provides a summary of results from an ANOVA analysis (R.W. Payne).
Options
PRINT = string tokens |
What to print (description, means, significant); default desc, mean, sign |
|---|---|
PSE = string tokens |
Standard errors to be printed with the means (sed, sedsummary, lsd, lsdsummary, dfmeans); default sed, dfme |
LSDLEVEL = scalar |
Significance level (%) for least significant differences; default 5 |
SAVE = ANOVA save structure |
Save structure for the analysis; default uses the save structure from the most recent ANOVA |
No parameters
Description
ARESULTSUMMARY investigates an ANOVA analysis, to provide the information that would be useful for a report. By default, all the information is printed, but you can control this with the PRINT option, whose settings are:
description |
prints the name of the y-variate, any covariates and the block and treatment models, |
means |
prints relevant tables of means, and |
significant |
lists the significant treatment terms. |
The relevant tables of means are those that contain significant treatment effects. Also, each table contains all the significant effects involving any of its factors. In the example for the procedure, terms A, D, S and A.S are significant. Two tables of means are therefore presented, one classified by A and S, and the other by D. However, if the significant terms were A.S and D.S. there would be only one table, classified by factors A, D and S.
The PSE option controls the information provided with the tables of means:
sed |
standard errors for differences between means, |
sedsummary |
summary of the standard errors for differences, |
dfmeans |
degrees of freedom for the standard errors of differences, |
lsd |
least significant differences between the means, and |
lsdsummary |
summary of the least significant differences. |
The default is to print the standard errors of differences and their degrees of freedom. Note: if all the differences between means have the same standard error of difference, a summary is printed for the settings sed and lsd, instead of the full symmetric matrices of values.
The LSDLEVEL option specifies the significance level (%) to use in the calculation of least significant differences (default 5%).
Options: PRINT, PSE, LSDLEVEL, SAVE.
Parameters: none.
See also
Directives: ADISPLAY, ANOVA.
Procedure: AFMEANS, A2RESULTSUMMARY.
Commands for: Analysis of variance.
Example
CAPTION 'ARESULTSUMMARY example',\
!t('Randomized-block design with 3 treatment factors:',\
'yields of mangold roots(t/acre) from Snedecor & Cochran (1980),',\
'Statistical Methods (7th edition), page 314.'); \
STYLE=meta,plain
FACTOR [NVALUES=32; LEVELS=2] A,S,D
& [LEVELS=4] Block
GENERATE Block,D,S,A
READ Yields
19.2 20.6 18.9 25.3 20.8 26.8 22.2 27.7
15.5 16.9 20.2 27.6 18.5 17.8 18.6 28.6
17.0 19.5 16.7 29.1 20.1 18.6 22.3 28.7
11.7 21.9 20.7 25.4 19.2 19.0 21.1 28.5 :
TREATMENTS D*S*A
BLOCKS Block
ANOVA [PRINT=aovtable; FPROB=yes] Yields
ARESULTSUMMARY