Weighting cases and AMOS
Locked
 
|
Dear list,
I have an unequal gender split in my sample and a reviewer of my paper has asked whether I can weight the cases and then run my analyses in AMOS, so that the gender split more accurately reflects the general population. I know how to weight cases in SPSS, but when the weight is turned on this doesn't carry over to AMOS when I run my analyses. Is there any way of setting up the file so my cases are then weighted when I run my structural equation models in AMOS? Thanks Kathryn Get a new e-mail account with Hotmail - Free. Sign-up now. |
|
|
Kathryn Without having more background, it seems an odd request from the
reviewer. Are there reasons to expect gender differences, or do you examine
the effect of gender elsewhere in your study? If so, the answer would be
run a multi-group analysis (Male vs Female) which is easily done in AMOS, and
check for differences by constraining the parameters to equality. If there are no reasons to expect gender differences, the
composition of your sample shouldn’t matter. Having said that, you can weight cases in MPlus, I don’t
know about AMOS unfortunately. Garry Gelade Business Analytic Ltd. From: SPSSX(r) Discussion
[mailto:[hidden email]] On Behalf Of Kathryn Gardner Dear list, Get
a new e-mail account with Hotmail - Free. Sign-up
now. |
|
|
Hi Garry,
Thanks for your response. My paper looked at factorial invariance of a measure across 2 age groups, but the reviewer said the 77% female sample was a problem given known gender differences on the measure. They commented that the results are not generalisable to the general population. The editor said I should address the reviewer's comment, perhaps by comparing separate male and female groups, or possibly by weighting cases. However, I can't do invariance testing acorss groups of males and females as the male sample is too small. I can't collect more data for males either as I don't have the funds to pay for the questionnaire. Kathryn Date: Wed, 21 Apr 2010 23:01:54 +0100 From: [hidden email] Subject: Re: Weighting cases and AMOS To: [hidden email] Kathryn
Without having more background, it seems an odd request from the reviewer. Are there reasons to expect gender differences, or do you examine the effect of gender elsewhere in your study? If so, the answer would be run a multi-group analysis (Male vs Female) which is easily done in AMOS, and check for differences by constraining the parameters to equality.
If there are no reasons to expect gender differences, the composition of your sample shouldn’t matter.
Having said that, you can weight cases in MPlus, I don’t know about AMOS unfortunately.
Garry Gelade Business Analytic Ltd.
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Kathryn Gardner
Dear list, Get a new e-mail account with Hotmail - Free. Sign-up now. Get a new e-mail account with Hotmail – Free. Sign-up now. |
|
|
Hi Kathryn again. I believe AMOS will accept a correlation matrix as input (I’m
not sure of the exact file format, maybe an SPSS matrix data file will work). In that case, weight your cases in SPSS, produce the weighted correlation
matrix, and use that as input to AMOS instead of the cases file. Garry Gelade Business Analytic Ltd From: SPSSX(r) Discussion
[mailto:[hidden email]] On Behalf Of Kathryn Gardner Hi
Garry, Date: Wed, 21 Apr 2010 23:01:54 +0100 Kathryn Without having more background, it seems an odd request from the
reviewer. Are there reasons to expect gender differences, or do you
examine the effect of gender elsewhere in your study? If so, the answer
would be run a multi-group analysis (Male vs Female) which is easily done in
AMOS, and check for differences by constraining the parameters to
equality. If there are no reasons to expect gender differences, the
composition of your sample shouldn’t matter. Having said that, you can weight cases in MPlus, I don’t know
about AMOS unfortunately. Garry Gelade Business Analytic Ltd. From: SPSSX(r) Discussion
[mailto:[hidden email]] On Behalf Of Kathryn Gardner Dear list, Get
a new e-mail account with Hotmail - Free. Sign-up now. Get
a new e-mail account with Hotmail – Free. Sign-up
now. |
|
|
Hi Kathryn,
Please follow the syntax below to create a correlation matrix file. what concerns you will be the ones highlighted in red. The ones not highlighted are for file generations only. The correlation matrix file can be created through your correlation syntax except that you will need to type in a subcommand /matrix=out. I hope this helps you in your AMOS work and is what you need. Regards Dorraj Oet NEW FILE. INPUT PROGRAM. VECTOR y(20F8.0). LOOP v1=1 TO 100. COMPUTE x1=UNIFORM(5). COMPUTE x2=UNIFORM(5). COMPUTE x3=UNIFORM(5). LOOP #cnt=1 TO 20. COMPUTE y(#cnt)=UNIFORM(10). END LOOP. END CASE. END LOOP. END FILE. END INPUT PROGRAM. EXECUTE. CORRELATIONS /VARIABLES=y1 y2 /matrix=out ('C:\test.sav') /stat=desc /PRINT=TWOTAIL NOSIG /MISSING=PAIRWISE. get file 'C:\test.sav'. Date: Wed, 21 Apr 2010 23:40:33 +0100 From: [hidden email] Subject: Re: Weighting cases and AMOS To: [hidden email] Hi Kathryn again.
I believe AMOS will accept a correlation matrix as input (I’m not sure of the exact file format, maybe an SPSS matrix data file will work). In that case, weight your cases in SPSS, produce the weighted correlation matrix, and use that as input to AMOS instead of the cases file.
Garry Gelade Business Analytic Ltd
From: SPSSX(r) Discussion
[mailto:[hidden email]] On Behalf Of Kathryn Gardner
Hi
Garry, Date: Wed, 21 Apr 2010 23:01:54 +0100 Kathryn
Without having more background, it seems an odd request from the reviewer. Are there reasons to expect gender differences, or do you examine the effect of gender elsewhere in your study? If so, the answer would be run a multi-group analysis (Male vs Female) which is easily done in AMOS, and check for differences by constraining the parameters to equality.
If there are no reasons to expect gender differences, the composition of your sample shouldn’t matter.
Having said that, you can weight cases in MPlus, I don’t know about AMOS unfortunately.
Garry Gelade Business Analytic Ltd.
From: SPSSX(r) Discussion
[mailto:[hidden email]] On Behalf Of Kathryn Gardner
Dear list, Get a new e-mail account with Hotmail - Free. Sign-up now.
Get a new e-mail account with Hotmail �C Free. Sign-up now. Your E-mail and More On-the-Go. Get Windows Live Hotmail Free. Sign up now. |
|
|
|
| Free forum by Nabble | Edit this page |