I don't think you really want a learning curve model (aka
latent growth curve model, see HLM software for these
models), but you do want to allow for round-specific fixed
effects, probably. Just -tab round, gen(rd)- and then
include rd* as regressors in each model, like so:
g A=ln(pA/(1-pA))
reg A theoreticalA rd* , score(sA)
est store A
g B=ln(pB/(1-pB))
reg B theoreticalB rd* , score(sB)
est store B
g C=ln(pC/(1-pC))
reg C theoreticalC rd* , score(sC)
est store C
suest A B C, cluster(id)
test [A]theoreticalA=0
test [B]theoreticalB=0, accum
test [C]theoreticalC=0, accum
Maybe you want indiv fixed effects of some kind, too...
cap drop dummy*
tab id, gen(dummy)
foreach v in A B C {
cap drop `v'
cap drop s`v'
g `v'=ln(p`v'/(1-p`v'))
reg `v' theoretical`v' rd* dummy*, score(s`v')
est store `v'
}
suest A B C, cluster(id)
test [A]theoreticalA=0
test [B]theoreticalB=0, accum
test [C]theoreticalC=0, accum
Does any of this help at all? Maybe, maybe not.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
