repeated measures MANCOVA question

[Sally-15]

Hi all

I am running a repeated measures MANCOVA (4 variables [qja, ija, rja, and
jaga], 2 levels for between-subjects factor of group, 2 levels for within-
subjects
factor of time, 1 covariate- developmental quotient).

I have successfully run the MANOVA, but now need to adjust the syntax to
covary out developmental quotient (since the groups ended up slightly
different on this).

I then have another MANCOVA to run with 3 times points and more variables
(see MANCOVA 2 below). THis is the one I am concerned about- not sure if
hte within S's has Repeated or Polynomial written next to it(See ??? below)


Can someone take a look at this syntax below and tell me what you think?

Regards,
Sally

GLM
qja3 qja4 ija3new ija4 rja3 rja4 jaga3 jaga4 BY cgroup WITH comp
/WSFACTOR = agei 2 Polynomial
/MEASURE = qja ija rja jaga
/METHOD = SSTYPE(3)
/EMMEANS = TABLES(cgroup) WITH(comp=MEAN)
/EMMEANS = TABLES(agei) WITH(comp=MEAN)
/EMMEANS = TABLES(cgroup*agei) WITH(comp=MEAN)
/PRINT = DESCRIPTIVE ETASQ OPOWER HOMOGENEITY
/CRITERIA = ALPHA(.05)
/WSDESIGN = agei
/DESIGN = comp cgroup .


MANCOVA 2

GLM
smtri9b smtri9c smtri9d fpnt13b fpnt13c fpnt13d swgaz18b swgaz18c swgaz18d
sref19b sref19c sref19d match20b match20c match20d BY cgroup WITH comp
/WSFACTOR = agei 3 Repeated OR Polynomial??????????
/MEASURE = smtri9 fpnt13 swgaz18 sref19 match20
/METHOD = SSTYPE(3)
/EMMEANS = TABLES(cgroup) WITH(comp=MEAN)
/EMMEANS = TABLES(agei) WITH(comp=MEAN)
/EMMEANS = TABLES(cgroup*agei) WITH(comp=MEAN)
/PRINT = DESCRIPTIVE ETASQ OPOWER HOMOGENEITY
/CRITERIA = ALPHA(.05)
/WSDESIGN = agei
/DESIGN = cgroup .
* Given a significant interaction effect,
conduct simple effects analyses;
GROUP effects within each AGE, and
AGE effects within each GROUP.
* Question 1: AD Vs CNT.
* Over the five ages, age1 vs age2 .
* age 2 vs age 3.

[statisticsdoc]

Stephen Brand
www.statisticsdoc.com

Sally,

Whether you use Repeated or Polynomial when you have three time points
depends on your research questions.  Repeated will simply contrast the
second time point with the first, and the third with the second.  If you are
interested in testing hypotheses about the shape of trends over time, then
Polynomial would be helpful.  Polynomial will test the significance of the
linear and the quadratic trend over time (i.e., is there some curvilinear
trend over time).  In developmental research, polynomial trends are often of
interest.

HTH,

Stephen Brand


For personalized and professional consultation in statistics and research
design, visit
www.statisticsdoc.com


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]]On Behalf Of
Sally
Sent: Sunday, January 14, 2007 7:16 PM
To: [hidden email]
Subject: repeated measures MANCOVA question


Hi all

I am running a repeated measures MANCOVA (4 variables [qja, ija, rja, and
jaga], 2 levels for between-subjects factor of group, 2 levels for within-
subjects
factor of time, 1 covariate- developmental quotient).

I have successfully run the MANOVA, but now need to adjust the syntax to
covary out developmental quotient (since the groups ended up slightly
different on this).

I then have another MANCOVA to run with 3 times points and more variables
(see MANCOVA 2 below). THis is the one I am concerned about- not sure if
hte within S's has Repeated or Polynomial written next to it(See ??? below)


Can someone take a look at this syntax below and tell me what you think?

Regards,
Sally

GLM
qja3 qja4 ija3new ija4 rja3 rja4 jaga3 jaga4 BY cgroup WITH comp
/WSFACTOR = agei 2 Polynomial
/MEASURE = qja ija rja jaga
/METHOD = SSTYPE(3)
/EMMEANS = TABLES(cgroup) WITH(comp=MEAN)
/EMMEANS = TABLES(agei) WITH(comp=MEAN)
/EMMEANS = TABLES(cgroup*agei) WITH(comp=MEAN)
/PRINT = DESCRIPTIVE ETASQ OPOWER HOMOGENEITY
/CRITERIA = ALPHA(.05)
/WSDESIGN = agei
/DESIGN = comp cgroup .


MANCOVA 2

GLM
smtri9b smtri9c smtri9d fpnt13b fpnt13c fpnt13d swgaz18b swgaz18c swgaz18d
sref19b sref19c sref19d match20b match20c match20d BY cgroup WITH comp
/WSFACTOR = agei 3 Repeated OR Polynomial??????????
/MEASURE = smtri9 fpnt13 swgaz18 sref19 match20
/METHOD = SSTYPE(3)
/EMMEANS = TABLES(cgroup) WITH(comp=MEAN)
/EMMEANS = TABLES(agei) WITH(comp=MEAN)
/EMMEANS = TABLES(cgroup*agei) WITH(comp=MEAN)
/PRINT = DESCRIPTIVE ETASQ OPOWER HOMOGENEITY
/CRITERIA = ALPHA(.05)
/WSDESIGN = agei
/DESIGN = cgroup .
* Given a significant interaction effect,
conduct simple effects analyses;
GROUP effects within each AGE, and
AGE effects within each GROUP.
* Question 1: AD Vs CNT.
* Over the five ages, age1 vs age2 .
* age 2 vs age 3.