Hi All,

Apologies for multiple postings but I forgot to share my full SPSS output for the linear mixed model that I executed and imputed dataset. Kindly find both of them in the attachment. The SPSS syntax that I use is as follows:

SET SUMMARY=None TOLERANCE=1 CELLSBREAK=10000 TABLERENDER=light ROWSBREAK=100 TLook=None TFit=Both.
DATASET ACTIVATE LMMM1.
MIXED VAS_Actual_by_Period BY Gender TreatmentbyPeriod Period Seq ID WITH VAS_Baseline_by_Period Age
/CRITERIA=CIN(95) MXITER(100) MXSTEP(10) SCORING(1) SINGULAR(0.000000000001) HCONVERGE(0,
ABSOLUTE) LCONVERGE(0, ABSOLUTE) PCONVERGE(0.000001, ABSOLUTE)
/FIXED=Age VAS_Baseline_by_Period Period TreatmentbyPeriod Gender | SSTYPE(3)
/METHOD=REML
/PRINT=CPS DESCRIPTIVES G SOLUTION TESTCOV
/RANDOM=ID(Seq) | COVTYPE(VC)
/SAVE=PRED RESID
/EMMEANS=TABLES(TreatmentbyPeriod) COMPARE ADJ(BONFERRONI)

Is there a way to obtain the pooled estimates for pairwise comparisons (estimated marginal means) for linear mixed model? So far I could not find a workround for this issue in SPSS. I have tried SPSS 23, 25 and 26 but I still could not solve the problems. All pointers are greatly and humbly appreciated. Many thanks.


Kind regards
Jagjeet

------------------------------
Jagjeet Singh Hullon Data Scientist
------------------------------

#SPSSStatistics

Hi. I forwarded your question to one of our statisticians, who just got back to me.

Although we have to look into it in some detail, this seems like a bug. If we cannot provide the pooled results, we have to at least document why. Investigation is called for. Thanks for bringing this to our attention.

------------------------------
Rick Marcantonio
Quality Assurance
IBM
------------------------------

Original Message:
Sent: Wed November 16, 2022 03:16 AM
From: Jagjeet Singh Hullon
Subject: Pooled Estimates for Pairwise Comparisons in Linear Mixed Model

Hi All,

Apologies for multiple postings but I forgot to share my full SPSS output for the linear mixed model that I executed and imputed dataset. Kindly find both of them in the attachment. The SPSS syntax that I use is as follows:

SET SUMMARY=None TOLERANCE=1 CELLSBREAK=10000 TABLERENDER=light ROWSBREAK=100 TLook=None TFit=Both.
DATASET ACTIVATE LMMM1.
MIXED VAS_Actual_by_Period BY Gender TreatmentbyPeriod Period Seq ID WITH VAS_Baseline_by_Period Age
/CRITERIA=CIN(95) MXITER(100) MXSTEP(10) SCORING(1) SINGULAR(0.000000000001) HCONVERGE(0,
ABSOLUTE) LCONVERGE(0, ABSOLUTE) PCONVERGE(0.000001, ABSOLUTE)
/FIXED=Age VAS_Baseline_by_Period Period TreatmentbyPeriod Gender | SSTYPE(3)
/METHOD=REML
/PRINT=CPS DESCRIPTIVES G SOLUTION TESTCOV
/RANDOM=ID(Seq) | COVTYPE(VC)
/SAVE=PRED RESID
/EMMEANS=TABLES(TreatmentbyPeriod) COMPARE ADJ(BONFERRONI)

Is there a way to obtain the pooled estimates for pairwise comparisons (estimated marginal means) for linear mixed model? So far I could not find a workround for this issue in SPSS. I have tried SPSS 23, 25 and 26 but I still could not solve the problems. All pointers are greatly and humbly appreciated. Many thanks.


Kind regards
Jagjeet

------------------------------
Jagjeet Singh Hullon Data Scientist
------------------------------

#SPSSStatistics

Hi Rick,

Many thanks for your reply. Really appreciate it. I hope this issue can be resolved soon. And please do inform me on any update / progress made by the IBM SPSS Statistics team on this issue. Cheers and all the best.

------------------------------
Jagjeet Singh Hullon Data Scientist
------------------------------

Original Message:
Sent: Thu November 17, 2022 12:08 PM
From: Rick Marcantonio
Subject: Pooled Estimates for Pairwise Comparisons in Linear Mixed Model

Hi. I forwarded your question to one of our statisticians, who just got back to me.

Although we have to look into it in some detail, this seems like a bug. If we cannot provide the pooled results, we have to at least document why. Investigation is called for. Thanks for bringing this to our attention.

------------------------------
Rick Marcantonio
Quality Assurance
IBM

Original Message:
Sent: Wed November 16, 2022 03:16 AM
From: Jagjeet Singh Hullon
Subject: Pooled Estimates for Pairwise Comparisons in Linear Mixed Model

Hi All,

Apologies for multiple postings but I forgot to share my full SPSS output for the linear mixed model that I executed and imputed dataset. Kindly find both of them in the attachment. The SPSS syntax that I use is as follows:

SET SUMMARY=None TOLERANCE=1 CELLSBREAK=10000 TABLERENDER=light ROWSBREAK=100 TLook=None TFit=Both.
DATASET ACTIVATE LMMM1.
MIXED VAS_Actual_by_Period BY Gender TreatmentbyPeriod Period Seq ID WITH VAS_Baseline_by_Period Age
/CRITERIA=CIN(95) MXITER(100) MXSTEP(10) SCORING(1) SINGULAR(0.000000000001) HCONVERGE(0,
ABSOLUTE) LCONVERGE(0, ABSOLUTE) PCONVERGE(0.000001, ABSOLUTE)
/FIXED=Age VAS_Baseline_by_Period Period TreatmentbyPeriod Gender | SSTYPE(3)
/METHOD=REML
/PRINT=CPS DESCRIPTIVES G SOLUTION TESTCOV
/RANDOM=ID(Seq) | COVTYPE(VC)
/SAVE=PRED RESID
/EMMEANS=TABLES(TreatmentbyPeriod) COMPARE ADJ(BONFERRONI)

Is there a way to obtain the pooled estimates for pairwise comparisons (estimated marginal means) for linear mixed model? So far I could not find a workround for this issue in SPSS. I have tried SPSS 23, 25 and 26 but I still could not solve the problems. All pointers are greatly and humbly appreciated. Many thanks.


Kind regards
Jagjeet

------------------------------
Jagjeet Singh Hullon Data Scientist
------------------------------

#SPSSStatistics