[R] Repeated measures lme or anova
Martin Henry H. Stevens
HStevens at muohio.edu
Sun Jul 6 14:46:30 CEST 2008
Hi John,
1. I do not know why you remove the intercept in the lme model, but
keep it in the aov model.
2. The distributional assumptions are shot --- you can't run any sort
of normal model with these data. You might consider some sort of
binomial (metabolite detected vs. not detected).
Hank
On Jul 4, 2008, at 9:24 AM, John Coulthard wrote:
>
> Hi
>
> As I can't find an example of my data structure I'd like some
> advice on which is the most appropriate test for significant
> effects. If I should be using either lme or anova, is the relevant
> example below the best/correct way to do the test?
>
> The Data...
> 2 groups of patients (5 in GroupA, 7 in GroupB)
> 3 short acting drugs, (I'm not concerned with residual effects from
> the previous test effecting the current test)
> Each patient is given all 8 possible combinations of the 3 drugs (8
> measurements from each subject = repeated measures)
> The dependent effects are metabolite levels in the blood, a
> continuous variable but often skewed towards 0.1 which is the
> minimum detectable level.
> I'm look for individual and combined effects of Group and Drug.
> There are ~400 metabolites which I intend to test independently. (I
> know that will leave me with a multiple testing issue)
>
> What I've worked out for lme is
>> summary(test.lme <- lme(Value ~ Group*Drug1*Drug2*Drug3 - 1,
>> test,random = ~1|Patient))
>> anova(test.lme)
>
>
> and for anova is
>> summary(aov(Value~(Group*Drug1*Drug2*Drug3)+Error(Patient/
>> (Drug1*Drug2*Drug3)),test))
>
> The full structure of the 'test' data table for one metabolite is
> below.
>
> Thanks for your time and any thoughts you may have.
> Ann
>
>> test
> Patient Group Drug1 Drug2 Drug3 Value
> 1 1 A 0 0 0 446.70
> 2 1 A 0 0 1 0.10
> 3 1 A 0 1 1 0.10
> 4 1 A 0 1 0 328.20
> 5 1 A 1 0 0 0.10
> 6 1 A 1 0 1 0.10
> 7 1 A 1 1 1 0.10
> 8 1 A 1 1 0 0.10
> 9 2 B 0 0 0 0.10
> 10 2 B 0 0 1 69.93
> 11 2 B 0 1 0 878.30
> 12 2 B 0 1 1 0.10
> 13 2 B 1 0 1 0.10
> 14 2 B 1 0 0 0.10
> 15 2 B 1 1 1 0.10
> 16 2 B 1 1 0 0.10
> 17 3 A 0 0 0 0.10
> 18 3 A 0 0 1 0.10
> 19 3 A 0 1 0 0.10
> 20 3 A 0 1 1 0.10
> 21 3 A 1 0 1 0.10
> 22 3 A 1 0 0 0.10
> 23 3 A 1 1 0 0.10
> 24 3 A 1 1 1 0.10
> 25 4 B 0 0 1 0.10
> 26 4 B 0 0 0 688.50
> 27 4 B 0 1 0 541.00
> 28 4 B 0 1 1 0.10
> 29 4 B 1 0 0 0.10
> 30 4 B 1 0 1 0.10
> 31 4 B 1 1 1 0.10
> 32 4 B 1 1 0 0.10
> 33 5 B 0 0 0 0.10
> 34 5 B 0 0 1 0.10
> 35 5 B 0 1 0 541.60
> 36 5 B 0 1 1 0.10
> 37 5 B 1 0 0 0.10
> 38 5 B 1 0 1 0.10
> 39 5 B 1 1 0 0.10
> 40 5 B 1 1 1 0.10
> 41 6 B 0 0 1 58.29
> 42 6 B 0 0 0 353.60
> 43 6 B 0 1 0 523.30
> 44 6 B 0 1 1 0.10
> 45 6 B 1 0 0 0.10
> 46 6 B 1 0 1 0.10
> 47 6 B 1 1 0 0.10
> 48 6 B 1 1 1 0.10
> 49 7 A 0 0 0 351.03
> 50 7 A 0 0 1 0.10
> 51 7 A 0 1 1 0.10
> 52 7 A 0 1 0 0.10
> 53 7 A 1 0 0 0.10
> 54 7 A 1 0 1 0.10
> 55 7 A 1 1 1 0.10
> 56 7 A 1 1 0 0.10
> 57 8 A 0 0 0 299.80
> 58 8 A 0 0 1 0.10
> 59 8 A 0 1 0 0.10
> 60 8 A 0 1 1 0.10
> 61 8 A 1 0 1 0.10
> 62 8 A 1 0 0 0.10
> 63 8 A 1 1 0 0.10
> 64 8 A 1 1 1 0.10
> 65 9 B 0 0 0 355.50
> 66 9 B 0 0 1 0.10
> 67 9 B 0 1 0 737.90
> 68 9 B 0 1 1 322.60
> 69 9 B 1 0 0 0.10
> 70 9 B 1 0 1 0.10
> 71 9 B 1 1 0 0.10
> 72 9 B 1 1 1 0.10
> 73 10 A 0 0 1 462.90
> 74 10 A 0 0 0 657.50
> 75 10 A 0 1 0 82.13
> 76 10 A 0 1 1 0.10
> 77 10 A 1 0 1 0.10
> 78 10 A 1 0 0 0.10
> 79 10 A 1 1 1 0.10
> 80 10 A 1 1 0 0.10
> 81 11 B 0 0 0 386.70
> 82 11 B 0 0 1 279.40
> 83 11 B 0 1 1 407.30
> 84 11 B 0 1 0 485.20
> 85 11 B 1 0 1 0.10
> 86 11 B 1 0 0 0.10
> 87 11 B 1 1 0 0.10
> 88 11 B 1 1 1 0.10
> 89 12 B 0 0 1 0.10
> 90 12 B 0 0 0 705.10
> 91 12 B 0 1 0 706.20
> 92 12 B 0 1 1 450.20
> 93 12 B 1 0 1 0.10
> 94 12 B 1 0 0 0.10
> 95 12 B 1 1 0 0.10
> 96 12 B 1 1 1 0.10
>>
>
> _________________________________________________________________
> Discover the new Windows Vista
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-
> guide.html
> and provide commented, minimal, self-contained, reproducible code.
Dr. Hank Stevens, Associate Professor
338 Pearson Hall
Botany Department
Miami University
Oxford, OH 45056
Office: (513) 529-4206
Lab: (513) 529-4262
FAX: (513) 529-4243
http://www.cas.muohio.edu/~stevenmh/
http://www.users.muohio.edu/harkesae/
http://www.cas.muohio.edu/ecology
http://www.muohio.edu/botany/
"E Pluribus Unum"
"I love deadlines. I love the whooshing noise they make as they go by."
(Douglas Adams)
If you send an attachment, please try to send it in a format anyone
can read, such as PDF, text, Open Document Format, HTML, or RTF.
Why? See: http://www.gnu.org/philosophy/no-word-attachments.html
More information about the R-help
mailing list