Multivariate Analysis in MIXED
Locked
 
|
For those interested, I've provided steps to running a multivariate analysis in the MIXED procedure in PASW 17.
In order to run a multivariate analysis employing the MIXED procedure, one would need to add an indicator variable as a link to the response variables to the data set. Also, in order to allow for separate intercepts for each response variable, the grand intercept must be excluded, and to allow for separate slopes, the fixed effects covariates should be added only by interacting them with the indicator variable. Note that the response is assumed to be a multivariate normal. Here is an example: --------------------------- ID Indic Y X1 1 1 150 22 1 2 70 33 2 1 180 24 2 2 72 48 3 1 163 2 3 2 62 23 . . . N --------------------------- where ID = identification number, repeating for each unit Indic = indicator of the dependent variables (i.e. 1=height, 2=weight) Y = value on that specific dependent variable (i.e. in inches for height, in pounds for weight) X1 = fixed effects covariate --------------------------- The code to run such a model would be as follows: MIXED Y BY Indic WITH X1 /FIXED=Indic Indic*X1 | NOINT SSTYPE(3) /METHOD=REML /PRINT=DESCRIPTIVES SOLUTION /RANDOM=INTERCEPT | SUBJECT(ID) COVTYPE(VC). ---------------------------- Here are a couple of articles that explain running a multivariate analysis in linear mixed modeling: http://www2.sas.com/proceedings/sugi23/Stats/p229.pdf http://arxiv.org/ftp/arxiv/papers/0705/0705.0568.pdf ---------------------------- ***Here's a more eloquent explanation in the SAS google group that I read earlier on when learning the how to run such an analysis in SAS: http://listserv.uga.edu/cgi-bin/wa?A2=ind0402D&L=sas-l&P=R20058 ---------------------------- I welcome thoughts on this topic. Best, Ryan |
|
Very well done, thanks
WMB Statistical Services ============ mailto: [hidden email] http:\\home.earthlink.net\~info.statman Skype: zstatman ============ -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of rblack Sent: Wednesday, January 06, 2010 9:20 PM To: [hidden email] Subject: Multivariate Analysis in MIXED For those interested, I've provided steps to running a multivariate analysis in the MIXED procedure in PASW 17. In order to run a multivariate analysis employing the MIXED procedure, one would need to add an indicator variable as a link to the response variables to the data set. Also, in order to allow for separate intercepts for each response variable, the grand intercept must be excluded, and to allow for separate slopes, the fixed effects covariates should be added only by interacting them with the indicator variable. Note that the response is assumed to be a multivariate normal. Here is an example: --------------------------- ID Indic Y X1 1 1 150 22 1 2 70 33 2 1 180 24 2 2 72 48 3 1 163 2 3 2 62 23 . . . N --------------------------- where ID = identification number, repeating for each unit Indic = indicator of the dependent variables (i.e. 1=height, 2=weight) Y = value on that specific dependent variable (i.e. in inches for height, in pounds for weight) X1 = fixed effects covariate --------------------------- The code to run such a model would be as follows: MIXED Y BY Indic WITH X1 /FIXED=Indic Indic*X1 | NOINT SSTYPE(3) /METHOD=REML /PRINT=DESCRIPTIVES SOLUTION /RANDOM=INTERCEPT | SUBJECT(ID) COVTYPE(VC). ---------------------------- Here are a couple of articles that explain running a multivariate analysis in linear mixed modeling: http://www2.sas.com/proceedings/sugi23/Stats/p229.pdf http://arxiv.org/ftp/arxiv/papers/0705/0705.0568.pdf ---------------------------- ***Here's a more eloquent explanation in the SAS google group that I read earlier on when learning the how to run such an analysis in SAS: http://listserv.uga.edu/cgi-bin/wa?A2=ind0402D&L=sas-l&P=R20058 ---------------------------- I welcome thoughts on this topic. Best, Ryan -- View this message in context: http://old.nabble.com/Multivariate-Analysis-in-MIXED-tp27027781p27027781.htm l Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Will
Statistical Services ============ info.statman@earthlink.net http://home.earthlink.net/~z_statman/ ============ |
Re: Multivariate Analysis in MIXED
Administrator
|
In reply to this post by Ryan
I just noticed that textbook author Ronald Heck has a set of slides showing this same approach to multivariate analysis via MIXED. You can view them here:
http://www2.hawaii.edu/~ltabata/mlm/ppt5-multivarmodels.html
--
Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." PLEASE NOTE THE FOLLOWING: 1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. 2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/). |
|
|
Bruce,
That's a pretty old post. I actually wrote back to the list suggesting a better approach (another old post, but more recent). Whether it is consistent with that online presentation I cannot investigate at this time. More recent post here: https://listserv.uga.edu/cgi-bin/wa?A2=ind1102&L=SPSSX-L&P=R51830 Ryan Sent from my iPhone > On Oct 27, 2015, at 3:37 PM, Bruce Weaver <[hidden email]> wrote: > > I just noticed that textbook author Ronald Heck has a set of slides showing > this same approach to multivariate analysis via MIXED. You can view them > here: > > http://www2.hawaii.edu/~ltabata/mlm/ppt5-multivarmodels.html > > > > > Ryan Black wrote >> For those interested, I've provided steps to running a multivariate >> analysis in the MIXED procedure in PASW 17. >> >> In order to run a multivariate analysis employing the MIXED procedure, one >> would need to add an indicator variable as a link to the response >> variables to the data set. Also, in order to allow for separate intercepts >> for each response variable, the grand intercept must be excluded, and to >> allow for separate slopes, the fixed effects covariates should be added >> only by interacting them with the indicator variable. Note that the >> response is assumed to be a multivariate normal. >> >> Here is an example: >> >> --------------------------- >> ID Indic Y X1 >> 1 1 150 22 >> 1 2 70 33 >> 2 1 180 24 >> 2 2 72 48 >> 3 1 163 2 >> 3 2 62 23 >> . >> . >> . >> N >> --------------------------- >> >> where >> >> ID = identification number, repeating for each unit >> Indic = indicator of the dependent variables (i.e. 1=height, 2=weight) >> Y = value on that specific dependent variable (i.e. in inches for height, >> in pounds for weight) >> X1 = fixed effects covariate >> --------------------------- >> >> The code to run such a model would be as follows: >> >> MIXED Y BY Indic WITH X1 >> /FIXED=Indic Indic*X1 | NOINT SSTYPE(3) >> /METHOD=REML >> /PRINT=DESCRIPTIVES SOLUTION >> /RANDOM=INTERCEPT | SUBJECT(ID) COVTYPE(VC). >> >> ---------------------------- >> >> Here are a couple of articles that explain running a multivariate analysis >> in linear mixed modeling: >> >> >> http://www2.sas.com/proceedings/sugi23/Stats/p229.pdf >> >> http://arxiv.org/ftp/arxiv/papers/0705/0705.0568.pdf >> >> ---------------------------- >> >> ***Here's a more eloquent explanation in the SAS google group that I read >> earlier on when learning the how to run such an analysis in SAS: >> >> http://listserv.uga.edu/cgi-bin/wa?A2=ind0402D&L=sas-l&P=R20058 >> >> >> ---------------------------- >> >> I welcome thoughts on this topic. >> >> Best, >> >> Ryan > > > > > > ----- > -- > Bruce Weaver > [hidden email] > http://sites.google.com/a/lakeheadu.ca/bweaver/ > > "When all else fails, RTFM." > > NOTE: My Hotmail account is not monitored regularly. > To send me an e-mail, please use the address shown above. > > -- > View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Multivariate-Analysis-in-MIXED-tp1092023p5730884.html > Sent from the SPSSX Discussion mailing list archive at Nabble.com. > > ===================== > To manage your subscription to SPSSX-L, send a message to > [hidden email] (not to SPSSX-L), with no body text except the > command. To leave the list, send the command > SIGNOFF SPSSX-L > For a list of commands to manage subscriptions, send the command > INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Multivariate Analysis in MIXED
|
Administrator
|
Thanks Ryan. I'd forgotten about your follow-up post, and didn't spot it when searching the archives. For those who prefer using the Nabble archive, here's the link to Ryan's more recent post:
http://spssx-discussion.1045642.n5.nabble.com/Multivariate-Mixed-Model-Follow-Up-td3392871.html
--
Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." PLEASE NOTE THE FOLLOWING: 1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. 2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/). |
Re: Multivariate Analysis in MIXED
|
Administrator
|
Due to some changes at Nabble a few years ago, the link I gave below no longer works. Here is the current version:
http://spssx-discussion.165.s1.nabble.com/Multivariate-Mixed-Model-Follow-Up-td3392871.html
--
Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." PLEASE NOTE THE FOLLOWING: 1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. 2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/). |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |