t test with a covariate
Locked
 
|
Dear all,
I am comparing performance on a task from two groups of participants using independent t test. I would also like to add in a covariate (memory span) to control for any initial differences in memory ability. I wonder if anyone has done a t test with a covariate before and can share with me how to do it is Spss. Many thanks! Tim |
Administrator
|
As has been discussed in another recent thread, the ordinary (pooled variance) version of the independent groups t-test is equivalent to one-way ANOVA on the same data: t^2 = F. So use UNIANOVA with the grouping variable as the "fixed factor" and the covariate as the "covariate".
I would suggest looking at a scatter-plot first though, with X = the covariate, Y = the outcome variable, and separate fit-lines for the sub-groups. This will allow you to visually inspect how reasonable it is to force the slopes to be the same for the two groups (which is what the ANCOVA model does). If the scatter-plot suggests that the slopes are different, include the Group x Covariate term in your model. (You can use the F-test on the interaction to confirm what you suspected from the scatter-plot; but I would not necessarily exclude the interaction if it failed to reach the .05 level of significance. The power for interaction terms is generally lower than for main effects.) If you include the Group x Covariate term, the difference between groups will depend on where you look along the X-axis. So you'll probably want to report the group difference at a few points (low, medium & high on X). You can do that with multiple EMMEANS sub-commands, specifying a different X value each time for the WITH option. E.g., suppose the low, medium and high values you want are 40, 50 and 60, and the grouping variable is Grp: /EMMEANS = table(Grp) WITH(X=40) /EMMEANS = table(Grp) WITH(X=50) /EMMEANS = table(Grp) WITH(X=60) If you omit the WITH option, the comparison of the groups will be with X = its mean. HTH.
--
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/). |
|
|
In reply to this post by Tim Wei
Another example of why we should ditch the old terminology and just use general linear models! Forget t tests, ANOVA, ANCOVA, and regression. GLM!
Dr. Paul R. Swank, Children's Learning Institute Professor, Department of Pediatrics, Medical School Adjunct Professor, School of Public Health University of Texas Health Science Center-Houston -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Tim Wei Sent: Friday, October 07, 2011 6:14 AM To: [hidden email] Subject: t test with a covariate Dear all, I am comparing performance on a task from two groups of participants using independent t test. I would also like to add in a covariate (memory span) to control for any initial differences in memory ability. I wonder if anyone has done a t test with a covariate before and can share with me how to do it is Spss. Many thanks! Tim -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/t-test-with-a-covariate-tp4879717p4879717.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 |
|
Administrator
|
That's what Judd & McClelland (1989) do in "Data Analysis: A Model-Comparison Approach". In general, I agree we should take this approach going forward. But as stats instructors, we also have to enable people to understand all the existing literature that does report t-tests, ANOVA, ANCOVA, regression, etc.
--
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/). |
|
|
In reply to this post by Swank, Paul R
Thank you for your response and explanation. This is very helpful.
Tim |
|
|
In reply to this post by Bruce Weaver
Unfortunately, we do it backwards. We teach the t test and ANOVA, etc, and then introduce them to linear models. It is difficult to overcome that previous learning. They would always ask " why not just do a t test?" I would prefer to teach GLM and then go back and indicate that when someone talks about an independent samples t test, they really mean a GLM with a categorical predictor having two levels.
Paul Dr. Paul R. Swank, Children's Learning Institute Professor, Department of Pediatrics, Medical School Adjunct Professor, School of Public Health University of Texas Health Science Center-Houston -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Bruce Weaver Sent: Friday, October 07, 2011 10:49 AM To: [hidden email] Subject: Re: t test with a covariate That's what Judd & McClelland (1989) do in "Data Analysis: A Model-Comparison Approach". In general, I agree we should take this approach going forward. But as stats instructors, we also have to enable people to understand all the existing literature that /does/ report t-tests, ANOVA, ANCOVA, regression, etc. Swank, Paul R wrote: > > Another example of why we should ditch the old terminology and just use > general linear models! Forget t tests, ANOVA, ANCOVA, and regression. GLM! > > Dr. Paul R. Swank, > Children's Learning Institute > Professor, Department of Pediatrics, Medical School > Adjunct Professor, School of Public Health > University of Texas Health Science Center-Houston > > > -----Original Message----- > From: SPSSX(r) Discussion [mailto:SPSSX-L@.UGA] On Behalf Of Tim Wei > Sent: Friday, October 07, 2011 6:14 AM > To: SPSSX-L@.UGA > Subject: t test with a covariate > > Dear all, > > I am comparing performance on a task from two groups of participants using > independent t test. I would also like to add in a covariate (memory span) > to > control for any initial differences in memory ability. I wonder if anyone > has done a t test with a covariate before and can share with me how to do > it > is Spss. Many thanks! > > Tim > > > -- > View this message in context: > http://spssx-discussion.1045642.n5.nabble.com/t-test-with-a-covariate-tp4879717p4879717.html > Sent from the SPSSX Discussion mailing list archive at Nabble.com. > > ===================== > To manage your subscription to SPSSX-L, send a message to > LISTSERV@.UGA (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 > LISTSERV@.UGA (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 > ----- -- 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/t-test-with-a-covariate-tp4879717p4880446.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 |
Administrator
|
Hell,
Just begin with SEM (Structural Equations Modeling) and then we can toss Factor Analysis into the mix and GLM (ANOVA, ANCOVA, t-test, Regression) as just another special case ;-) LOL, David --
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
Don't forget general linear and nonlinear mixed models.
Paul Dr. Paul R. Swank, Children's Learning Institute Professor, Department of Pediatrics, Medical School Adjunct Professor, School of Public Health University of Texas Health Science Center-Houston -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of David Marso Sent: Friday, October 07, 2011 11:46 AM To: [hidden email] Subject: Re: t test with a covariate Hell, Just begin with SEM (Structural Equations Modeling) and then we can toss Factor Analysis into the mix and GLM (ANOVA, ANCOVA, t-test, Regression) as just another special case ;-) LOL, David -- Swank, Paul R wrote: > > Unfortunately, we do it backwards. We teach the t test and ANOVA, etc, and > then introduce them to linear models. It is difficult to overcome that > previous learning. They would always ask " why not just do a t test?" I > would prefer to teach GLM and then go back and indicate that when someone > talks about an independent samples t test, they really mean a GLM with a > categorical predictor having two levels. > > Paul > > Dr. Paul R. Swank, > Children's Learning Institute > Professor, Department of Pediatrics, Medical School > Adjunct Professor, School of Public Health > University of Texas Health Science Center-Houston > > > -----Original Message----- > From: SPSSX(r) Discussion [mailto:SPSSX-L@.UGA] On Behalf Of Bruce Weaver > Sent: Friday, October 07, 2011 10:49 AM > To: SPSSX-L@.UGA > Subject: Re: t test with a covariate > > That's what Judd & McClelland (1989) do in "Data Analysis: A > Model-Comparison > Approach". In general, I agree we should take this approach going > forward. > But as stats instructors, we also have to enable people to understand all > the existing literature that /does/ report t-tests, ANOVA, ANCOVA, > regression, etc. > > > > Swank, Paul R wrote: >> >> Another example of why we should ditch the old terminology and just use >> general linear models! Forget t tests, ANOVA, ANCOVA, and regression. >> GLM! >> >> Dr. Paul R. Swank, >> Children's Learning Institute >> Professor, Department of Pediatrics, Medical School >> Adjunct Professor, School of Public Health >> University of Texas Health Science Center-Houston >> >> >> -----Original Message----- >> From: SPSSX(r) Discussion [mailto:SPSSX-L@.UGA] On Behalf Of Tim Wei >> Sent: Friday, October 07, 2011 6:14 AM >> To: SPSSX-L@.UGA >> Subject: t test with a covariate >> >> Dear all, >> >> I am comparing performance on a task from two groups of participants >> using >> independent t test. I would also like to add in a covariate (memory span) >> to >> control for any initial differences in memory ability. I wonder if anyone >> has done a t test with a covariate before and can share with me how to do >> it >> is Spss. Many thanks! >> >> Tim >> >> >> -- >> View this message in context: >> http://spssx-discussion.1045642.n5.nabble.com/t-test-with-a-covariate-tp4879717p4879717.html >> Sent from the SPSSX Discussion mailing list archive at Nabble.com. >> >> ===================== >> To manage your subscription to SPSSX-L, send a message to >> LISTSERV@.UGA (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 >> LISTSERV@.UGA (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 >> > > > ----- > -- > Bruce Weaver > bweaver@ > 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/t-test-with-a-covariate-tp4879717p4880446.html > Sent from the SPSSX Discussion mailing list archive at Nabble.com. > > ===================== > To manage your subscription to SPSSX-L, send a message to > LISTSERV@.UGA (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 > LISTSERV@.UGA (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 > -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/t-test-with-a-covariate-tp4879717p4880647.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 |
|
Administrator
|
I think maybe we'll save that for the second semester.
--
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |