repeated measures anova
Locked
 
|
I'm trying to conduct a Repeated Measures ANOVA in SPSS. My syntax is pasted below.
I am trying to obtain results where both time and scale are repeated/within-subject factors and then test for time, scale, and time*scale effects---where all scales at all times listed as variables. My data are in wide form. The wave variables are coded two ways: one set has variables wave1 to wave 5 each coded as 0,1, indicating which subject has data for each wave, the other set has variables w1 to W5 coded as 1,2,3,4,5 depending on which subject has data for each wave. The problem is that I am unable to obtain an F-test for the time variable. And I also cannot generate the interaction. GLM wave1 wave2 wave3 wave4 wave5 W1Char W2Char W3Char W4Char W5Char /WSFACTOR=time 5 Polynomial /MEASURE=wave char /METHOD=SSTYPE(3) /PLOT=PROFILE(time) /EMMEANS=TABLES(time) COMPARE ADJ(BONFERRONI) /PRINT=DESCRIPTIVE ETASQ /CRITERIA=ALPHA(.05) /WSDESIGN=time. I suspect that a) I have a fundamental error in my setup, b) I am overlooking a minor problem with my syntax, or c) I suspect that my wave variables (i.e. time) are coded incorrectly. Or is there some other problem with my setup? I am not interested in recasting the analysis in SAS, HLM, or Structural Equation Modeling. My boss is looking for a quick and simple solution to this in SPSS. Thanks in advance. |
Re: repeated measures anova
|
|
Dear "Thanks in advance", I find your description of your data/design a bit confusing. First of all, RM ANOVA is used to test longitudinal changes in a continuous dependent variable. Do you have a continuous DV in your design? That is, did you measure some quantifiable something (intelligence, body temperature, etc.) at each time point?
it seems to me that this is the first question that needs to be answered here before we can help you further. What do others think? bozena ----- Original Message ----- From: "pjiman1" <[hidden email]> To: [hidden email] Sent: Tuesday, July 19, 2011 11:06:46 AM GMT -08:00 US/Canada Pacific Subject: repeated measures anova I'm trying to conduct a Repeated Measures ANOVA in SPSS. My syntax is pasted below. I am trying to obtain results where both time and scale are repeated/within-subject factors and then test for time, scale, and time*scale effects---where all scales at all times listed as variables. My data are in wide form. The wave variables are coded two ways: one set has variables wave1 to wave 5 each coded as 0,1, indicating which subject has data for each wave, the other set has variables w1 to W5 coded as 1,2,3,4,5 depending on which subject has data for each wave. The problem is that I am unable to obtain an F-test for the time variable. And I also cannot generate the interaction. GLM wave1 wave2 wave3 wave4 wave5 W1Char W2Char W3Char W4Char W5Char /WSFACTOR=time 5 Polynomial /MEASURE=wave char /METHOD=SSTYPE(3) /PLOT=PROFILE(time) /EMMEANS=TABLES(time) COMPARE ADJ(BONFERRONI) /PRINT=DESCRIPTIVE ETASQ /CRITERIA=ALPHA(.05) /WSDESIGN=time. I suspect that a) I have a fundamental error in my setup, b) I am overlooking a minor problem with my syntax, or c) I suspect that my wave variables (i.e. time) are coded incorrectly. Or is there some other problem with my setup? I am not interested in recasting the analysis in SAS, HLM, or Structural Equation Modeling. My boss is looking for a quick and simple solution to this in SPSS. Thanks in advance. -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/repeated-measures-anova-tp4613114p4613114.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 |
|
|
In reply to this post by pjiman1
This message was on another listserver and I replied as follows to it.
Peter, First off, I don't understand what you are trying to do. That said, your syntax indicates that you have a multivariate repeated measures with two measures at each of five time points. You have only one factor: time. There is no interaction. >>The problem is that I am unable to obtain an F-test for the time variable. You get something, i'll bet, so what do you get? >>And I also cannot generate the interaction. If you want an interaction, you have to set up your data differently. I think this is the correct setup. There may be some other things you want to change downstream of the Wsfactor subcommand. GLM wave1 wave2 wave3 wave4 wave5 W1Char W2Char W3Char W4Char W5Char /WSFACTOR=char 2 time 5 Polynomial /METHOD=SSTYPE(3) /PLOT=PROFILE(time) /EMMEANS=TABLES(time) COMPARE ADJ(BONFERRONI) /PRINT=DESCRIPTIVE ETASQ /CRITERIA=ALPHA(.05) /WSDESIGN=time. Peter, since you reposted to this listserv, I'm concluding that my suggestions were not helpful. What did I not understand? Gene Maguin -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of pjiman1 Sent: Tuesday, July 19, 2011 2:07 PM To: [hidden email] Subject: repeated measures anova I'm trying to conduct a Repeated Measures ANOVA in SPSS. My syntax is pasted below. I am trying to obtain results where both time and scale are repeated/within-subject factors and then test for time, scale, and time*scale effects---where all scales at all times listed as variables. My data are in wide form. The wave variables are coded two ways: one set has variables wave1 to wave 5 each coded as 0,1, indicating which subject has data for each wave, the other set has variables w1 to W5 coded as 1,2,3,4,5 depending on which subject has data for each wave. The problem is that I am unable to obtain an F-test for the time variable. And I also cannot generate the interaction. GLM wave1 wave2 wave3 wave4 wave5 W1Char W2Char W3Char W4Char W5Char /WSFACTOR=time 5 Polynomial /MEASURE=wave char /METHOD=SSTYPE(3) /PLOT=PROFILE(time) /EMMEANS=TABLES(time) COMPARE ADJ(BONFERRONI) /PRINT=DESCRIPTIVE ETASQ /CRITERIA=ALPHA(.05) /WSDESIGN=time. I suspect that a) I have a fundamental error in my setup, b) I am overlooking a minor problem with my syntax, or c) I suspect that my wave variables (i.e. time) are coded incorrectly. Or is there some other problem with my setup? I am not interested in recasting the analysis in SAS, HLM, or Structural Equation Modeling. My boss is looking for a quick and simple solution to this in SPSS. Thanks in advance. -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/repeated-measures-anova-tp4613 114p4613114.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 |
|
|
In reply to this post by pjiman1
I clarified the design with mn teds. He wanted a doubly repeated measures anova. Here's the syntax and the output was what he needed.
my question is now: is there a syntax command to obtain the output tests of time effect for each scale? If I can't obtain that output, then I'll conduct 6 individual repeated ANOVAs for each of the 6 variables and report the time effect for each one. pjiman1 GLM S1_peer S2_peer S3_peer S4_peer S5_peer S1_honesty S2_honesty S3_honesty S4_honesty S5_honesty S1_selfdev S2_selfdev S3_selfdev S4_selfdev S5_selfdev S1_Control S2_Control S3_Control S4_Control S5_Control S1_School S2_School S3_School S4_School S5_School S1_Home S2_Home S3_Home S4_Home S5_Home /WSFACTOR=char 6 Polynomial wave 5 Polynomial /METHOD=SSTYPE(3) /PLOT=PROFILE(wave*char) /EMMEANS=TABLES(char) /EMMEANS=TABLES(wave) /EMMEANS=TABLES(char*wave) compare(char) /PRINT=DESCRIPTIVE ETASQ /CRITERIA=ALPHA(.05) /WSDESIGN=char wave char*wave. |
Re: repeated measures anova
|
|
Can you clarify if the study collected the same measure under six different conditions (peer, honesty, selfdev, etc) or if these six are different measures, for example, subscales of the char scale. This will define what syntax should be used.
thanks. bozena zdaniuk ----- Original Message ----- From: "pjiman1" <[hidden email]> To: [hidden email] Sent: Tuesday, July 26, 2011 7:58:21 AM GMT -08:00 US/Canada Pacific Subject: Re: repeated measures anova I clarified the design with mn teds. He wanted a doubly repeated measures anova. Here's the syntax and the output was what he needed. my question is now: is there a syntax command to obtain the output tests of time effect for each scale? If I can't obtain that output, then I'll conduct 6 individual repeated ANOVAs for each of the 6 variables and report the time effect for each one. pjiman1 GLM S1_peer S2_peer S3_peer S4_peer S5_peer S1_honesty S2_honesty S3_honesty S4_honesty S5_honesty S1_selfdev S2_selfdev S3_selfdev S4_selfdev S5_selfdev S1_Control S2_Control S3_Control S4_Control S5_Control S1_School S2_School S3_School S4_School S5_School S1_Home S2_Home S3_Home S4_Home S5_Home /WSFACTOR=char 6 Polynomial wave 5 Polynomial /METHOD=SSTYPE(3) /PLOT=PROFILE(wave*char) /EMMEANS=TABLES(char) /EMMEANS=TABLES(wave) /EMMEANS=TABLES(char*wave) compare(char) /PRINT=DESCRIPTIVE ETASQ /CRITERIA=ALPHA(.05) /WSDESIGN=char wave char*wave. -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/repeated-measures-anova-tp4613114p4635111.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 |
|
|
they are six different subscales of the character scale.
out of curiosity, if the study did involve collecting the same measure under six different conditions, what would be the syntax? thanks. pjiman1 |
Re: repeated measures anova
|
|
In reply to this post by pjiman1
It means that you have six different measures (subscales) measured at 5 time points. This is what SPSS calls " A Doubly Multivariate Analysis of Variance" or, in your case, something like "sextuply" instead of "Doubly". It also means that you have only one within subject factor, the time (or what you call "wave"). So, basically it is as if you measured weight, BP, temp, and three other things, all of them at five different time points. Since you have only one within subject factor (and no btw subject factors) there is no interaction to be tested. In this case your syntax should be:
(if you had measured the same thing under six different "char" conditions then your original syntax would make sense.) bozena GLM S1_peer S2_peer S3_peer S4_peer S5_peer S1_honesty S2_honesty S3_honesty S4_honesty S5_honesty S1_selfdev S2_selfdev S3_selfdev S4_selfdev S5_selfdev S1_Control S2_Control S3_Control S4_Control S5_Control S1_School S2_School S3_School S4_School S5_School S1_Home S2_Home S3_Home S4_Home S5_Home /WSFACTOR=wave 5 Repeated /MEASURE= peer honesty selfdev control school home /METHOD=SSTYPE(3) /PLOT=PROFILE(wave) /EMMEANS=TABLES(wave) /PRINT=DESCRIPTIVE ETASQ /CRITERIA=ALPHA(.05) /WSDESIGN=wave. ----- Original Message ----- From: "pjiman1" <[hidden email]> To: [hidden email] Sent: Wednesday, July 27, 2011 7:34:32 PM GMT -08:00 US/Canada Pacific Subject: Re: repeated measures anova they are six different subscales of the character scale. out of curiosity, if the study did involve collecting the same measure under six different conditions, what would be the syntax? thanks. pjiman1 -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/repeated-measures-anova-tp4613114p4641033.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 |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |