I apologize for sending this message again. The last time I sent it, the subject line was not correct. I have corrected the subject line. I am trying to run a repeated measures analysis of data in which each subject (identified by SS) has 3 observations at three different times (0, 3, and 6). There are two groups of subjects (identified by group). I want to know if the response differs in the two groups. I have tried to used lme. Lme tell me if there is a time effect, but does not tell me if there is a group effect. Once I get this to work I will want to know if there is a significant group*time effect. Can someone tell me how to get an estimate for group. Once I get that, I believe getting an estimate for group*time should be straight forward. The code I have tired to use follows. Thank you, John> # This is my data > data1SS group time value baseline 1 1 Cont 0 9.000000 9.000000 2 2 Cont 0 3.000000 3.000000 3 3 Cont 0 8.000000 8.000000 4 4 Inte 0 5.690702 5.690702 5 5 Inte 0 7.409493 7.409493 6 6 Inte 0 7.428018 7.428018 7 1 Cont 3 13.713148 9.000000 8 2 Cont 3 9.841107 3.000000 9 3 Cont 3 12.843236 8.000000 10 4 Inte 3 9.300899 5.690702 11 5 Inte 3 10.936389 7.409493 12 6 Inte 3 12.358499 7.428018 13 1 Cont 6 18.952390 9.000000 14 2 Cont 6 15.091527 3.000000 15 3 Cont 6 17.578812 8.000000 16 4 Inte 6 12.325499 5.690702 17 5 Inte 6 15.486513 7.409493 18 6 Inte 6 18.284965 7.428018> # Create a grouped data object. SS identifies each subject > # group indentifies group, intervention or control. > GD<- groupedData(value~time|SS/group,data=data1,FUN=mean) > # Fit the model. > fit1 <- lme(GD) > cat("The results give information about time, but does not say if the gruops are different\n")The results give information about time, but does not say if the gruops are different> summary(fit1)Linear mixed-effects model fit by REML Data: GD AIC BIC logLik 74.59447 81.54777 -28.29724 Random effects: Formula: ~time | SS Structure: General positive-definite StdDev Corr (Intercept) 1.3875111 (Intr) time 0.2208046 -0.243 Formula: ~time | group %in% SS Structure: General positive-definite StdDev Corr (Intercept) 1.3875115 (Intr) time 0.2208051 -0.243 Residual 0.3800788 Fixed effects: value ~ time Value Std.Error DF t-value p-value (Intercept) 6.747442 0.8135067 11 8.294268 0 time 1.588653 0.1326242 11 11.978601 0 Correlation: (Intr) time -0.268 Standardized Within-Group Residuals: Min Q1 Med Q3 Max -1.11412947 -0.44986535 0.08034174 0.34615610 1.29943887 Number of Observations: 18 Number of Groups: SS group %in% SS 6 6>John David Sorkin M.D., Ph.D. Professor of Medicine Chief, Biostatistics and Informatics University of Maryland School of Medicine Division of Gerontology and Geriatric Medicine Baltimore VA Medical Center 10 North Greene Street GRECC (BT/18/GR) Baltimore, MD 21201-1524 (Phone) 410-605-7119 (Fax) 410-605-7913 (Please call phone number above prior to faxing) Confidentiality Statement: This email message, including any attachments, is for the sole use of the intended recipient(s) and may contain confidential and privileged information. Any unauthorized use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message.
I never used the groupedData structure, precisely because I found it confusing, but I think: 1. group is *not* a (random) grouping variable; it's a fixed effect covariate. 2. so I believe your groupedData call should be: GD<- groupedData(value~time|SS,data=data1,outer = group) Of course, as you did not give us your data in a convenient form, I can't check. Please let us know if this is wrong, however, as I don't want to mislead others down the primrose path. Cheers, Bert Bert Gunter "The trouble with having an open mind is that people keep coming along and sticking things into it." -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip ) On Wed, Aug 24, 2016 at 3:46 PM, John Sorkin <jsorkin at grecc.umaryland.edu> wrote:> I apologize for sending this message again. The last time I sent it, the subject line was not correct. I have corrected the subject line. > > I am trying to run a repeated measures analysis of data in which each subject (identified by SS) has 3 observations at three different times (0, 3, and 6). There are two groups of subjects (identified by group). I want to know if the response differs in the two groups. I have tried to used lme. Lme tell me if there is a time effect, but does not tell me if there is a group effect. Once I get this to work I will want to know if there is a significant group*time effect. Can someone tell me how to get an estimate for group. Once I get that, I believe getting an estimate for group*time should be straight forward. The code I have tired to use follows. > Thank you, > John > >> # This is my data >> data1 > SS group time value baseline > 1 1 Cont 0 9.000000 9.000000 > 2 2 Cont 0 3.000000 3.000000 > 3 3 Cont 0 8.000000 8.000000 > 4 4 Inte 0 5.690702 5.690702 > 5 5 Inte 0 7.409493 7.409493 > 6 6 Inte 0 7.428018 7.428018 > 7 1 Cont 3 13.713148 9.000000 > 8 2 Cont 3 9.841107 3.000000 > 9 3 Cont 3 12.843236 8.000000 > 10 4 Inte 3 9.300899 5.690702 > 11 5 Inte 3 10.936389 7.409493 > 12 6 Inte 3 12.358499 7.428018 > 13 1 Cont 6 18.952390 9.000000 > 14 2 Cont 6 15.091527 3.000000 > 15 3 Cont 6 17.578812 8.000000 > 16 4 Inte 6 12.325499 5.690702 > 17 5 Inte 6 15.486513 7.409493 > 18 6 Inte 6 18.284965 7.428018 >> # Create a grouped data object. SS identifies each subject >> # group indentifies group, intervention or control. >> GD<- groupedData(value~time|SS/group,data=data1,FUN=mean) >> # Fit the model. >> fit1 <- lme(GD) >> cat("The results give information about time, but does not say if the gruops are different\n") > The results give information about time, but does not say if the gruops are different >> summary(fit1) > Linear mixed-effects model fit by REML > Data: GD > AIC BIC logLik > 74.59447 81.54777 -28.29724 > > Random effects: > Formula: ~time | SS > Structure: General positive-definite > StdDev Corr > (Intercept) 1.3875111 (Intr) > time 0.2208046 -0.243 > > Formula: ~time | group %in% SS > Structure: General positive-definite > StdDev Corr > (Intercept) 1.3875115 (Intr) > time 0.2208051 -0.243 > Residual 0.3800788 > > Fixed effects: value ~ time > Value Std.Error DF t-value p-value > (Intercept) 6.747442 0.8135067 11 8.294268 0 > time 1.588653 0.1326242 11 11.978601 0 > Correlation: > (Intr) > time -0.268 > > Standardized Within-Group Residuals: > Min Q1 Med Q3 Max > -1.11412947 -0.44986535 0.08034174 0.34615610 1.29943887 > > Number of Observations: 18 > Number of Groups: > SS group %in% SS > 6 6 > > > >> > John David Sorkin M.D., Ph.D. > Professor of Medicine > Chief, Biostatistics and Informatics > University of Maryland School of Medicine Division of Gerontology and Geriatric Medicine > Baltimore VA Medical Center > 10 North Greene Street > GRECC (BT/18/GR) > Baltimore, MD 21201-1524 > (Phone) 410-605-7119 > (Fax) 410-605-7913 (Please call phone number above prior to faxing) > > Confidentiality Statement: > This email message, including any attachments, is for ...{{dropped:12}}
Dear John, lme() not longer requires a GroupedData object. You can directly use a data.frame which is easier to specify different models. You want something like lme(value ~ time * group, random = ~ time|SS, data = data1) PS Note that the R-Sig-mixedmodels is more suited for this kind of question. Best regards, ir. Thierry Onkelinx Instituut voor natuur- en bosonderzoek / Research Institute for Nature and Forest team Biometrie & Kwaliteitszorg / team Biometrics & Quality Assurance Kliniekstraat 25 1070 Anderlecht Belgium To call in the statistician after the experiment is done may be no more than asking him to perform a post-mortem examination: he may be able to say what the experiment died of. ~ Sir Ronald Aylmer Fisher The plural of anecdote is not data. ~ Roger Brinner The combination of some data and an aching desire for an answer does not ensure that a reasonable answer can be extracted from a given body of data. ~ John Tukey 2016-08-25 0:46 GMT+02:00 John Sorkin <jsorkin at grecc.umaryland.edu>:> I apologize for sending this message again. The last time I sent it, the > subject line was not correct. I have corrected the subject line. > > I am trying to run a repeated measures analysis of data in which each > subject (identified by SS) has 3 observations at three different times (0, > 3, and 6). There are two groups of subjects (identified by group). I want > to know if the response differs in the two groups. I have tried to used > lme. Lme tell me if there is a time effect, but does not tell me if there > is a group effect. Once I get this to work I will want to know if there is > a significant group*time effect. Can someone tell me how to get an estimate > for group. Once I get that, I believe getting an estimate for group*time > should be straight forward. The code I have tired to use follows. > Thank you, > John > > > # This is my data > > data1 > SS group time value baseline > 1 1 Cont 0 9.000000 9.000000 > 2 2 Cont 0 3.000000 3.000000 > 3 3 Cont 0 8.000000 8.000000 > 4 4 Inte 0 5.690702 5.690702 > 5 5 Inte 0 7.409493 7.409493 > 6 6 Inte 0 7.428018 7.428018 > 7 1 Cont 3 13.713148 9.000000 > 8 2 Cont 3 9.841107 3.000000 > 9 3 Cont 3 12.843236 8.000000 > 10 4 Inte 3 9.300899 5.690702 > 11 5 Inte 3 10.936389 7.409493 > 12 6 Inte 3 12.358499 7.428018 > 13 1 Cont 6 18.952390 9.000000 > 14 2 Cont 6 15.091527 3.000000 > 15 3 Cont 6 17.578812 8.000000 > 16 4 Inte 6 12.325499 5.690702 > 17 5 Inte 6 15.486513 7.409493 > 18 6 Inte 6 18.284965 7.428018 > > # Create a grouped data object. SS identifies each subject > > # group indentifies group, intervention or control. > > GD<- groupedData(value~time|SS/group,data=data1,FUN=mean) > > # Fit the model. > > fit1 <- lme(GD) > > cat("The results give information about time, but does not say if the > gruops are different\n") > The results give information about time, but does not say if the gruops > are different > > summary(fit1) > Linear mixed-effects model fit by REML > Data: GD > AIC BIC logLik > 74.59447 81.54777 -28.29724 > > Random effects: > Formula: ~time | SS > Structure: General positive-definite > StdDev Corr > (Intercept) 1.3875111 (Intr) > time 0.2208046 -0.243 > > Formula: ~time | group %in% SS > Structure: General positive-definite > StdDev Corr > (Intercept) 1.3875115 (Intr) > time 0.2208051 -0.243 > Residual 0.3800788 > > Fixed effects: value ~ time > Value Std.Error DF t-value p-value > (Intercept) 6.747442 0.8135067 11 8.294268 0 > time 1.588653 0.1326242 11 11.978601 0 > Correlation: > (Intr) > time -0.268 > > Standardized Within-Group Residuals: > Min Q1 Med Q3 Max > -1.11412947 -0.44986535 0.08034174 0.34615610 1.29943887 > > Number of Observations: 18 > Number of Groups: > SS group %in% SS > 6 6 > > > > > > John David Sorkin M.D., Ph.D. > Professor of Medicine > Chief, Biostatistics and Informatics > University of Maryland School of Medicine Division of Gerontology and > Geriatric Medicine > Baltimore VA Medical Center > 10 North Greene Street > GRECC (BT/18/GR) > Baltimore, MD 21201-1524 > (Phone) 410-605-7119 > (Fax) 410-605-7913 (Please call phone number above prior to faxing) > > Confidentiality Statement: > This email message, including any attachments, is for ...{{dropped:16}}