*  do-file to demonstrate commands in the STB insert for aovsum
*  John R. Gleason
*  Syracuse University
*  430 Huntington Hall
*  Syracuse, NY  13244-2340
*  loesljrg@accucom.net
*  04Jan2000


di in red "OK to clear?"
more

clear
** Example 1:

aovsum, n(12 11 9 9) m(18.1 28.2 48.3 70.2) sd(8.54 14.3 12.4 14.0)

* Repeat quietly, to get the synthetic data:
more
quietly aovsum, n(12 11 9 9) m(18.1 28.2 48.3 70.2) sd(8.54 14.3 12.4 14.0) keep

list freq_ y_ cond_

** Create a copy of the group variable:
more
gen byte Cond_ = cond_

** Test for linear trend and deviation from linear trend:
more
anova y_ cond_ Cond_ [fw=freq_], cont(cond_) seq

** Repeat the trend tests, using the actual raw data:
more
use ldose, clear

describe

** Create a copy of the group variable, again:
more
gen byte Rate = rate

** And run -anova-, again:
more
anova ldose rate Rate, cont(rate) seq

more
** Example 2:

** Create synthetic data for 8 groups:
more

clear
quietly aovsum, n(13 14 20 26 20 20 16 17)   /*
   */    m(13.62 15.79 16.55 7.00 31.75 10.45 20.06 19.18)  /*
   */    sd(12.07 17.49 14.75 6.11 22.07 10.68 14.22 17.90) /*
   */    keep names(days group)


** Create factors for race and grade in school:
more
gen byte race = mod(group-1, 2)
gen byte grade = int((group-1)/2)

** Run the usual 2-way ANOVA with interaction:
more
anova days race grade race*grade [fw=freq_]

** Repeat, this time using the actual raw data:
more
use absences, replace
describe

** And the ANOVA:
more
anova days race grade race*grade

** Done