tag:blogger.com,1999:blog-6710487119650146215.post8574974866783965268..comments2017-03-27T01:51:05.632-07:00Comments on R Tutorial Series: R Tutorial Series: Two-Way Repeated Measures ANOVAJohnhttp://www.blogger.com/profile/05331039307550313006noreply@blogger.comBlogger21125tag:blogger.com,1999:blog-6710487119650146215.post-44289697841263740842014-02-19T02:04:00.594-07:002014-02-19T02:04:00.594-07:00how to run post hoc after two way repeated measure...how to run post hoc after two way repeated measures ANOVA?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-85305649732337012932013-12-14T12:30:36.556-07:002013-12-14T12:30:36.556-07:00How do I run a three way repeated measures anova?
...How do I run a three way repeated measures anova?<br />Aynara Chávez Wulsinhttp://www.blogger.com/profile/04050775513146206288noreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-75580679445398440602013-12-02T15:22:54.429-07:002013-12-02T15:22:54.429-07:00I also get these errors, no matter how careful I w...I also get these errors, no matter how careful I write the code according the the instructions above:<br /><br />"Note: model has only an intercept; equivalent type-III tests substituted.<br />Warning message:<br />In summary.Anova.mlm(analysisABC): Singular error SSP matrix:<br />non-sphericity test and corrections not available"Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-39030633418227975702013-11-04T11:25:22.522-07:002013-11-04T11:25:22.522-07:00Hello, I get as well an error message after my ana...Hello, I get as well an error message after my analysis. I am farily new to R, so any suggestions what went wrong here would be very much appreciated. Many thanks, Wiebke<br /><br />Error in eigen(qr.coef(SSPE.qr, x$SSPH), symmetric = FALSE) : <br /> infinite or missing values in 'x'<br />In addition: Warning message:<br />In summary.Anova.mlm(analysis) : Singular error SSP matrix:<br />non-sphericity test and corrections not availableAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-91567582871201259202013-07-19T18:00:13.178-07:002013-07-19T18:00:13.178-07:00You need to convert the equivalent variable to Age...You need to convert the equivalent variable to Age in this case to a factor.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-85376724987264063572013-06-17T13:59:28.838-07:002013-06-17T13:59:28.838-07:00Thanks for the tutorial. Unfortunately, I get this...Thanks for the tutorial. Unfortunately, I get this error message when using my own data:<br /><br />Error in `contrasts<-`(`*tmp*`, value = if (is.ordered(idata[, i])) icontrasts[2] else icontrasts[1]) : <br /> contrasts apply only to factors<br /><br />Any idea what I'm doing wrong?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-69744350846272937902013-05-21T14:40:28.934-07:002013-05-21T14:40:28.934-07:00John, thank you for this tutorials!
Sorry if this...John, thank you for this tutorials!<br /><br />Sorry if this is an idiotic question, but if Jen creates another column for the site location, then how would she distinguish between the fish at different sites within the statistics?<br /><br />I have a similar study in which I have two treatment groups measured by a pretest, posttest, and delayed posttest. I can make the subject ID as column 1, and each of the scores for pretest, posttest, and delayed posttest as columns, as well as a column for treatment type. <br /><br />But (very much a beginner to R) how do I get it to statistically compare the two treatment groups over time (i.e. treatment, time, treatment*time)?<br /><br />Thank you,<br />JohnAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-78625666962323654802013-03-07T09:25:14.775-07:002013-03-07T09:25:14.775-07:00The example (two-way repeated measures ANOVA examp...The example (two-way repeated measures ANOVA example (.txt) ) is great and I could apply it very easily to my analysis, thanks! I am desperate to know how to look at the residuals and check distribution, outliers, etc. from there. Any hint would be very helpful. Thanks, Solene. Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-2858408968973181172012-04-04T20:18:04.486-07:002012-04-04T20:18:04.486-07:00Yes, this can be done with things other than time....Yes, this can be done with things other than time. The important thing is identifying something that has been repeated in the study. From the description, it sounds like they measured the same birds' behaviors multiple times. For example, Bird A dove when confronted with a male intruder, Bird A chased when confronted with a female intruder, and so on. Each confrontation situation presented should be a repeated measures column. For example, BehaviorMale1, BehaviorFemale1, etc. to represent each behavior measured in each confrontation situation.John M. Quickhttp://www.blogger.com/profile/05331039307550313006noreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-45183530964035337842012-04-04T20:00:37.161-07:002012-04-04T20:00:37.161-07:00Hi Jen. Look at the first image in this tutorial. ...Hi Jen. Look at the first image in this tutorial. Instead of subject, you would have a column called fish, which would have fish1, fish2, ... fish56. Your repeated measures variable is growth, which is taken at four time points. Therefore, you will need columns for each possible combination: sizeTime1, sizeTime2, sizeTime3, sizeTime4. Site is not a repeated measures variable, so that is just one more column in your dataset that associates each fish with its location.John M. Quickhttp://www.blogger.com/profile/05331039307550313006noreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-83649921986183652362012-04-04T19:14:06.010-07:002012-04-04T19:14:06.010-07:00Hi John,
Thanks for the great tutorial, it explai...Hi John,<br /><br />Thanks for the great tutorial, it explained a lot. I have a query about the organisation of the data. I want to compare differences in fish growth between 4 sites, I have 4 repeated measurements of fish size for 14 fish per site. If I organise my data like your example R only thinks I have 14 fish in total and the df in the MANOVA test are wrong. Could you please explain how I should organise my data and idata??<br /><br />Any help would be much appreciated. Cheers, JenJennoreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-65933234846553959372012-04-03T22:03:36.442-07:002012-04-03T22:03:36.442-07:00Hello
could this set up be used to test two withi...Hello<br /><br />could this set up be used to test two within-subject factors like behaviour (chase or dive-bomb)and intruder (female or male) instead of a time series?<br /><br />I am trying to replicate this study:<br /><br />"Contingency analyses were not appropriate because<br />multiple behaviors were recorded for each resident. To determine whether resident birds treated male and female intruders differently,we performed repeated-measures ANOVA.<br />The two within-subject fixed factors were Behavior<br />(chase, dive) and Intruder Sex (male, female).<br />Each subject was regarded as a unit of observation<br />and provided the random variate for analysis.<br />We examined the number of times a subject performed<br />each behavior in the presence of male and female<br />intruders."<br /><br />Please help.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-24524299872288800182011-10-05T07:58:07.303-07:002011-10-05T07:58:07.303-07:00Hi. In the case of an interaction, you need to exa...Hi. In the case of an interaction, you need to examine your data further, which requires separating the dataset along one of the interacting variables. In this case, if Age and Interest are interacting, the dataset must be divided along one of them. This can be accomplished by looking at Interest within all three levels of Age (10, 15, 20) or by looking at Age within both levels of Interest (school, work). Since multiple tests are required, you control for the Type I error rate by reducing it. See the Two-Way ANOVA with Interactions and Simple Main Effects tutorial for details on this process.John M. Quickhttp://www.blogger.com/profile/05331039307550313006noreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-60569147515334099342011-10-05T05:42:56.304-07:002011-10-05T05:42:56.304-07:00Hi what I want to ask is that you suggest braking ...Hi what I want to ask is that you suggest braking down 2 way repeated measures anova to one way repeated measures anova for further analysis is that un-parsimonious??Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-17157253383524518032011-10-01T08:30:41.415-07:002011-10-01T08:30:41.415-07:00Hi. I do not have a specific example for this scen...Hi. I do not have a specific example for this scenario. However, I do have tutorials for two-way repeated measures and two-way ANOVA with unequal sample sizes. You may be able to adapt the methods from both tutorials to your situation. I also recommend validating your methodology with a statistician.John M. Quickhttp://www.blogger.com/profile/05331039307550313006noreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-59852859566678254302011-09-30T10:52:22.011-07:002011-09-30T10:52:22.011-07:00Dear, do you have some script for unbalanced desig...Dear, do you have some script for unbalanced design for two-way ANOVA with repeated measurements?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-39401220327198011492011-02-22T09:38:36.223-07:002011-02-22T09:38:36.223-07:00Using multivariate techniques is certainly appropr...Using multivariate techniques is certainly appropriate. I'm not disagreeing with that at all. What I'm saying is that then it's not an ANOVA, it's a MANOVA. The univariate test doesn't just remove the multivariate results but it also provides different results.<br /><br />As an aside, MANOVA is not the panacea it was first believed to be with respect to circumventing the problems with repeated measures designs.<br /><br />At least this comment might be able to have this come up in people's searches.jchttp://www.blogger.com/profile/00227235335343168838noreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-82846091245998420652011-02-22T06:56:40.734-07:002011-02-22T06:56:40.734-07:00Using multivariate = FALSE removes the multivariat...Using multivariate = FALSE removes the multivariate tests from the output (e.g. Pillai, Wilks, etc.). From every indication that I have seen, it is acceptable practice to use multivariate techniques to execute repeated measures ANOVA, thus evading the sphericity/compound symmetry assumptions.John M. Quickhttp://www.blogger.com/profile/05331039307550313006noreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-36499198414951415802011-02-22T06:37:30.501-07:002011-02-22T06:37:30.501-07:00That's a MANOVA, not an ANOVA. To see the ANO...That's a MANOVA, not an ANOVA. To see the ANOVA you would need to enter summary(analysis, multivariate=FALSE)jchttp://www.blogger.com/profile/00227235335343168838noreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-26456133599841311842011-02-21T13:11:49.753-07:002011-02-21T13:11:49.753-07:00Thanks for the comments. The error is common, but ...Thanks for the comments. The error is common, but does not appear to affect the results in any of the instances that I have used this method.John M. Quickhttp://www.blogger.com/profile/05331039307550313006noreply@blogger.comtag:blogger.com,1999:blog-6710487119650146215.post-23467235999246782002011-02-21T11:06:36.315-07:002011-02-21T11:06:36.315-07:00Two things... didn't work... I got
"Note...Two things... didn't work... I got<br /><br />"Note: model has only an intercept; equivalent type-III tests substituted."<br /><br />when generating the Anova.<br /><br />Also, the cbind command is just really a very verbose way of saying you want to make a matrix out of all of the dependent variable colunns of the data. ... as.matrix(data[,2:7])<br /><br />(note... that's not why I got the error because I actually did it the way things are described... both ways came out the same)Anonymousnoreply@blogger.com