Is it possible to perform a multivariate multilevel model with Stata?

[Raffaele Palladino]

Dear Statalist,

I have a serie of outcomes (all continuous) measured in two groups at two different time points (after 2 years). I would like to assess the overall difference in the outcomes between the two groups (under the assumption that all the outcomes are correlated) and I was wondering whether it is possible with Stata to perform a multivariate multilevel linear mixed model. I would prefer this than a MANOVA test as it would consider that my database has two time points for the same population.


My outcome variables are outcome1 outcome2 outcome3 outcome4 outcome5. I also have an id variable (ID), a time variable (time 0/1), and a group variable (group 0/1).

I have not even tried to run the command as I am not entirely sure how I should do that.

Thanks in advance.

[Clyde Schechter]

From -help me-

Mixed-effects multinomial regression

Although there is no memlogit command, multilevel mixed-effects multinomial logistic models can be fit using gsem; see [SEM] example 41g.

So the short answer is no, but there is a workaround. I have not used GSEM to do this myself, and I don't feel qualified to assist you in the process, but the examples in the manuals are usually pretty clear. If you give it a start and get stuck, there are others on the forum who could probably help you get over the bumps.

[Alex Gamma]

Raffaele,
I just recently posted about this after researching a bit. Apparently it is possible if you arrange your data properly. Tabachnik & Fidell (section 15.5.5, p. 821) say that you need to dummy-code the different outcomes and then use this as your level-1 predictor. Baldwin et al (2014), in their supplemental material, give a concrete example of how to do it in Stata. If you don't have the paper, write to me privately.
Alex

References:
1. Baldwin SA, Imel ZE, Braithwaite SR, Atkins DC. Analyzing multiple outcomes in clinical research using multivariate multilevel models. J Consult Clin Psychol. 2014;82(5):920–30.
2. Tabachnik BG, Fidell LS. Using multivariate statistics. Pearson Education, Inc. 5th ed. Boston; 2007.

[Raffaele Palladino]

Thanks Alex and Clyde,
It looks feasible indeed but mostly when you do not have many outcomes (otherwise generating all the interaction variables might be messy).
What if you have many outcomes? I guess the MANOVA would be the only option. Is there a way to fit the MANOVA in a way that recognize the longitudinal nature of my dataset with multiple measures for the same person within multiple outcomes?

[Chris Martin]

In John Nezlek's book on Diary Methods, he has an example of how to dummy code the data to run a multinomial model.
Chris

[Raffaele Palladino]

I have tried to transform my database using the example provided by Baldwin et al but the model gives the "initial values not feasible" error (I guess I tried to fit a model with too many dummy variables in a database which has less than 150 observations (65 at time 0 and 65 at time 1)).
Therefore, I have tried to look at the MANOVA Stata manual to substitute xtmixed with MANOVA but the manual only provides examples for either one/many ways MANOVA or repeated measures MANOVA, while in my case I have a between ('group' variable) and within ('time' variable) effect to account for. Could you help/advise me on how to fit the model?

Thanks in advance,
R.

[Joseph Coveney]

Try something like that below. Start at the "Begin here" comment. First is multivariate multilevel using gsem and then with MANOVA.

.ÿversionÿ14.1

.ÿ
.ÿclearÿ*

.ÿsetÿmoreÿoff

.ÿsetÿseedÿ`=date("2016-02-29",ÿ"YMD")'

.ÿ
.ÿ/*ÿ"Myÿoutcomeÿvariablesÿareÿoutcome1ÿoutcome2ÿoutcome3ÿoutcome4ÿoutcome5.ÿ
>ÿÿÿÿÿIÿalsoÿhaveÿanÿidÿvariableÿ(ID),ÿaÿtimeÿvariableÿ(timeÿ0/1),ÿandÿaÿgroupÿvariableÿ(groupÿ0/1)."ÿ*/
.ÿ
.ÿquietlyÿsetÿobsÿ65ÿ//ÿ".ÿ.ÿ.ÿobservationsÿ(65ÿatÿtimeÿ0ÿandÿ65ÿatÿtimeÿ1)"

.ÿgenerateÿbyteÿgroupÿ=ÿmod(_n,ÿ2)

.ÿgenerateÿdoubleÿidÿ=ÿ_n

.ÿgenerateÿdoubleÿuÿ=ÿrnormal()

.ÿquietlyÿexpandÿ2

.ÿbysortÿid:ÿgenerateÿbyteÿtimeÿ=ÿ_nÿ-ÿ1

.ÿforvaluesÿiÿ=ÿ1/5ÿ{
ÿÿ2.ÿÿÿÿÿÿÿÿÿgenerateÿdoubleÿoutcome`i'ÿ=ÿuÿ+ÿrnormal()
ÿÿ3.ÿ}

.ÿ
.ÿ*
.ÿ*ÿBeginÿhere
.ÿ*
.ÿgsemÿ(outcome1ÿoutcome2ÿoutcome3ÿoutcome4ÿoutcome5ÿ<-ÿi.group##i.timeÿM1[id]),ÿÿ///
>ÿÿÿÿÿÿÿÿÿcovariance(e.outcome1*e.outcome2ÿe.outcome1*e.outcome3ÿe.outcome1*e.outcome4ÿe.outcome1*e.outcome5ÿ///
>ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿe.outcome2*e.outcome3ÿe.outcome2*e.outcome4ÿe.outcome2*e.outcome5ÿ///
>ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿe.outcome3*e.outcome4ÿe.outcome3*e.outcome5ÿ///
>ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿe.outcome4*e.outcome5)ÿnocnsreportÿnodvheaderÿnolog

GeneralizedÿstructuralÿequationÿmodelÿÿÿÿÿÿÿÿÿÿÿNumberÿofÿobsÿÿÿÿÿ=ÿÿÿÿÿÿÿÿ130
Logÿlikelihoodÿ=ÿ-970.71494

-------------------------------------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿÿCoef.ÿÿÿStd.ÿErr.ÿÿÿÿÿÿzÿÿÿÿP>|z|ÿÿÿÿÿ[95%ÿConf.ÿInterval]
--------------------------+----------------------------------------------------------------
outcome1ÿ<-ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1.groupÿ|ÿÿ-.4589077ÿÿÿ.3554206ÿÿÿÿ-1.29ÿÿÿ0.197ÿÿÿÿ-1.155519ÿÿÿÿ.2377039
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1.timeÿ|ÿÿ-.3777953ÿÿÿ.2621226ÿÿÿÿ-1.44ÿÿÿ0.150ÿÿÿÿ-.8915462ÿÿÿÿ.1359556
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿgroup#timeÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1ÿ1ÿÿ|ÿÿÿÿ.374571ÿÿÿ.3678784ÿÿÿÿÿ1.02ÿÿÿ0.309ÿÿÿÿ-.3464573ÿÿÿÿ1.095599
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿM1[id]ÿ|ÿÿÿÿÿÿÿÿÿÿ1ÿÿ(constrained)
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ_consÿ|ÿÿÿÿ.383106ÿÿÿ.2532462ÿÿÿÿÿ1.51ÿÿÿ0.130ÿÿÿÿ-.1132474ÿÿÿÿ.8794595
--------------------------+----------------------------------------------------------------
outcome2ÿ<-ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1.groupÿ|ÿÿÿ.2134708ÿÿÿ.3172271ÿÿÿÿÿ0.67ÿÿÿ0.501ÿÿÿÿ-.4082828ÿÿÿÿ.8352244
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1.timeÿ|ÿÿÿ.2747099ÿÿÿÿ.239912ÿÿÿÿÿ1.15ÿÿÿ0.252ÿÿÿÿ-.1955091ÿÿÿÿ.7449288
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿgroup#timeÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1ÿ1ÿÿ|ÿÿÿÿÿ-.3638ÿÿÿ.3367067ÿÿÿÿ-1.08ÿÿÿ0.280ÿÿÿÿ-1.023733ÿÿÿÿ.2961329
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿM1[id]ÿ|ÿÿÿ.8655812ÿÿÿ.1474601ÿÿÿÿÿ5.87ÿÿÿ0.000ÿÿÿÿÿ.5765647ÿÿÿÿ1.154598
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ_consÿ|ÿÿ-.0859169ÿÿÿ.2260323ÿÿÿÿ-0.38ÿÿÿ0.704ÿÿÿÿ-.5289321ÿÿÿÿ.3570983
--------------------------+----------------------------------------------------------------
outcome3ÿ<-ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1.groupÿ|ÿÿ-.0379962ÿÿÿ.3496783ÿÿÿÿ-0.11ÿÿÿ0.913ÿÿÿÿ-.7233531ÿÿÿÿ.6473606
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1.timeÿ|ÿÿÿ.2269692ÿÿÿ.2603942ÿÿÿÿÿ0.87ÿÿÿ0.383ÿÿÿÿÿ-.283394ÿÿÿÿ.7373324
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿgroup#timeÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1ÿ1ÿÿ|ÿÿÿ.1561582ÿÿÿ.3654525ÿÿÿÿÿ0.43ÿÿÿ0.669ÿÿÿÿ-.5601156ÿÿÿÿ.8724319
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿM1[id]ÿ|ÿÿÿ.9726955ÿÿÿ.1340964ÿÿÿÿÿ7.25ÿÿÿ0.000ÿÿÿÿÿ.7098714ÿÿÿÿÿ1.23552
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ_consÿ|ÿÿÿ.0604537ÿÿÿ.2491546ÿÿÿÿÿ0.24ÿÿÿ0.808ÿÿÿÿ-.4278804ÿÿÿÿ.5487878
--------------------------+----------------------------------------------------------------
outcome4ÿ<-ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1.groupÿ|ÿÿ-.2738214ÿÿÿ.3116547ÿÿÿÿ-0.88ÿÿÿ0.380ÿÿÿÿ-.8846535ÿÿÿÿ.3370106
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1.timeÿ|ÿÿ-.2281447ÿÿÿ.2389922ÿÿÿÿ-0.95ÿÿÿ0.340ÿÿÿÿ-.6965608ÿÿÿÿ.2402713
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿgroup#timeÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1ÿ1ÿÿ|ÿÿÿ.1296324ÿÿÿ.3354157ÿÿÿÿÿ0.39ÿÿÿ0.699ÿÿÿÿ-.5277702ÿÿÿÿ.7870351
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿM1[id]ÿ|ÿÿÿ.8347956ÿÿÿ.1295942ÿÿÿÿÿ6.44ÿÿÿ0.000ÿÿÿÿÿ.5807955ÿÿÿÿ1.088796
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ_consÿ|ÿÿÿ.1932194ÿÿÿ.2220619ÿÿÿÿÿ0.87ÿÿÿ0.384ÿÿÿÿ-.2420138ÿÿÿÿ.6284527
--------------------------+----------------------------------------------------------------
outcome5ÿ<-ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1.groupÿ|ÿÿÿ.1245081ÿÿÿ.3029192ÿÿÿÿÿ0.41ÿÿÿ0.681ÿÿÿÿ-.4692027ÿÿÿÿ.7182189
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1.timeÿ|ÿÿ-.1742712ÿÿÿÿ.223598ÿÿÿÿ-0.78ÿÿÿ0.436ÿÿÿÿ-.6125153ÿÿÿÿ.2639729
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿgroup#timeÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ1ÿ1ÿÿ|ÿÿ-.1441375ÿÿÿ.3138106ÿÿÿÿ-0.46ÿÿÿ0.646ÿÿÿÿÿ-.759195ÿÿÿÿ.4709201
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿM1[id]ÿ|ÿÿÿÿ.851424ÿÿÿ.1235063ÿÿÿÿÿ6.89ÿÿÿ0.000ÿÿÿÿÿ.6093561ÿÿÿÿ1.093492
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ_consÿ|ÿÿÿ.1179378ÿÿÿ.2158376ÿÿÿÿÿ0.55ÿÿÿ0.585ÿÿÿÿ-.3050961ÿÿÿÿ.5409718
--------------------------+----------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿvar(M1[id])|ÿÿÿ.9529439ÿÿÿ.2557673ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ.5631283ÿÿÿÿ1.612602
--------------------------+----------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿvar(e.outcome1)|ÿÿÿ1.099332ÿÿÿ.1543833ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ.8348177ÿÿÿÿÿ1.44766
ÿÿÿÿÿÿÿÿÿÿÿvar(e.outcome2)|ÿÿÿ.9209244ÿÿÿ.1293791ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ.6992623ÿÿÿÿ1.212852
ÿÿÿÿÿÿÿÿÿÿÿvar(e.outcome3)|ÿÿÿ1.084882ÿÿÿÿ.155216ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ.8195955ÿÿÿÿ1.436036
ÿÿÿÿÿÿÿÿÿÿÿvar(e.outcome4)|ÿÿÿ.9138761ÿÿÿ.1277603ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ.6948466ÿÿÿÿ1.201948
ÿÿÿÿÿÿÿÿÿÿÿvar(e.outcome5)|ÿÿÿ.7999372ÿÿÿ.1199806ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ.5961921ÿÿÿÿ1.073311
--------------------------+----------------------------------------------------------------
cov(e.outcome2,e.outcome1)|ÿÿ-.3358143ÿÿÿ.1019101ÿÿÿÿ-3.30ÿÿÿ0.001ÿÿÿÿ-.5355545ÿÿÿ-.1360742
cov(e.outcome3,e.outcome1)|ÿÿÿ.1302253ÿÿÿ.1110744ÿÿÿÿÿ1.17ÿÿÿ0.241ÿÿÿÿ-.0874765ÿÿÿÿ.3479271
cov(e.outcome4,e.outcome1)|ÿÿ-.0083711ÿÿÿ.0959868ÿÿÿÿ-0.09ÿÿÿ0.931ÿÿÿÿ-.1965019ÿÿÿÿ.1797596
cov(e.outcome5,e.outcome1)|ÿÿÿ.0588204ÿÿÿ.0945999ÿÿÿÿÿ0.62ÿÿÿ0.534ÿÿÿÿ-.1265921ÿÿÿÿ.2442328
cov(e.outcome3,e.outcome2)|ÿÿ-.1121743ÿÿÿ.0970596ÿÿÿÿ-1.16ÿÿÿ0.248ÿÿÿÿ-.3024075ÿÿÿÿÿ.078059
cov(e.outcome4,e.outcome2)|ÿÿÿ.1298179ÿÿÿ.0943196ÿÿÿÿÿ1.38ÿÿÿ0.169ÿÿÿÿ-.0550452ÿÿÿÿ.3146809
cov(e.outcome5,e.outcome2)|ÿÿ-.1085612ÿÿÿ.0839275ÿÿÿÿ-1.29ÿÿÿ0.196ÿÿÿÿÿ-.273056ÿÿÿÿ.0559337
cov(e.outcome4,e.outcome3)|ÿÿÿ-.034367ÿÿÿ.0983404ÿÿÿÿ-0.35ÿÿÿ0.727ÿÿÿÿ-.2271106ÿÿÿÿ.1583766
cov(e.outcome5,e.outcome3)|ÿÿ-.1678687ÿÿÿ.0945451ÿÿÿÿ-1.78ÿÿÿ0.076ÿÿÿÿ-.3531737ÿÿÿÿ.0174363
cov(e.outcome5,e.outcome4)|ÿÿÿ.1119298ÿÿÿ.0877869ÿÿÿÿÿ1.28ÿÿÿ0.202ÿÿÿÿ-.0601294ÿÿÿÿÿ.283989
-------------------------------------------------------------------------------------------

.ÿ
.ÿsetÿemptycellsÿdrop

.ÿmanovaÿoutcome1ÿoutcome2ÿoutcome3ÿoutcome4ÿoutcome5ÿ=ÿgroupÿ/ÿid|groupÿtimeÿgroup#time

ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿNumberÿofÿobsÿ=ÿÿÿÿÿÿÿÿ130

ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿWÿ=ÿWilks'ÿlambdaÿÿÿÿÿÿLÿ=ÿLawley-Hotellingÿtrace
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿPÿ=ÿPillai'sÿtraceÿÿÿÿÿRÿ=ÿRoy'sÿlargestÿroot

ÿÿÿÿÿÿÿÿÿÿÿÿÿÿSourceÿ|ÿStatisticÿÿÿÿÿÿÿÿdfÿÿÿÿF(df1,ÿÿÿÿÿdf2)ÿ=ÿÿÿFÿÿÿProb>F
ÿÿÿÿÿÿÿÿÿÿ-----------+-------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿModelÿ|Wÿÿÿ0.0040ÿÿÿÿÿÿÿÿ66ÿÿÿÿ330.0ÿÿÿÿ299.9ÿÿÿÿÿ1.83ÿ0.0000ÿa
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|Pÿÿÿ3.0110ÿÿÿÿÿÿÿÿÿÿÿÿÿÿ330.0ÿÿÿÿ315.0ÿÿÿÿÿ1.45ÿ0.0005ÿa
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|Lÿÿ15.7463ÿÿÿÿÿÿÿÿÿÿÿÿÿÿ330.0ÿÿÿÿ287.0ÿÿÿÿÿ2.74ÿ0.0000ÿa
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|Rÿÿ11.1444ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ66.0ÿÿÿÿÿ63.0ÿÿÿÿ10.64ÿ0.0000ÿu
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|-------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿResidualÿ|ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ63
ÿÿÿÿÿÿÿÿÿÿ-----------+-------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿgroupÿ|Wÿÿÿ0.9333ÿÿÿÿÿÿÿÿÿ1ÿÿÿÿÿÿ5.0ÿÿÿÿÿ59.0ÿÿÿÿÿ0.84ÿ0.5241ÿe
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|Pÿÿÿ0.0667ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ5.0ÿÿÿÿÿ59.0ÿÿÿÿÿ0.84ÿ0.5241ÿe
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|Lÿÿÿ0.0715ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ5.0ÿÿÿÿÿ59.0ÿÿÿÿÿ0.84ÿ0.5241ÿe
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|Rÿÿÿ0.0715ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ5.0ÿÿÿÿÿ59.0ÿÿÿÿÿ0.84ÿ0.5241ÿe
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|-------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿid|groupÿ|ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ63
ÿÿÿÿÿÿÿÿÿÿ-----------+-------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿtimeÿ|Wÿÿÿ0.9127ÿÿÿÿÿÿÿÿÿ1ÿÿÿÿÿÿ5.0ÿÿÿÿÿ59.0ÿÿÿÿÿ1.13ÿ0.3552ÿe
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|Pÿÿÿ0.0873ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ5.0ÿÿÿÿÿ59.0ÿÿÿÿÿ1.13ÿ0.3552ÿe
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|Lÿÿÿ0.0957ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ5.0ÿÿÿÿÿ59.0ÿÿÿÿÿ1.13ÿ0.3552ÿe
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|Rÿÿÿ0.0957ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ5.0ÿÿÿÿÿ59.0ÿÿÿÿÿ1.13ÿ0.3552ÿe
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|-------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿgroup#timeÿ|Wÿÿÿ0.9685ÿÿÿÿÿÿÿÿÿ1ÿÿÿÿÿÿ5.0ÿÿÿÿÿ59.0ÿÿÿÿÿ0.38ÿ0.8579ÿe
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|Pÿÿÿ0.0315ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ5.0ÿÿÿÿÿ59.0ÿÿÿÿÿ0.38ÿ0.8579ÿe
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|Lÿÿÿ0.0325ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ5.0ÿÿÿÿÿ59.0ÿÿÿÿÿ0.38ÿ0.8579ÿe
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|Rÿÿÿ0.0325ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ5.0ÿÿÿÿÿ59.0ÿÿÿÿÿ0.38ÿ0.8579ÿe
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|-------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿResidualÿ|ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ63
ÿÿÿÿÿÿÿÿÿÿ-----------+-------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿTotalÿ|ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ129
ÿÿÿÿÿÿÿÿÿÿ-------------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿeÿ=ÿexact,ÿaÿ=ÿapproximate,ÿuÿ=ÿupperÿboundÿonÿF

.ÿ
.ÿexit

endÿofÿdo-file


.

Did you change the contents of your post after Clyde and Chris responded? The flow of this thread doesn't make sense.

[RonD McDowell]

This is a helpful thread. I wondered what happens if the outcomes were non-normal eg Binomial. If all of these outcomes were binary, each with a fixed within cluster variance (of pi-squared/3), can within cluster covariances be estimated, or are they too a fixed value?