[R-sig-ME] plot response by levels of factor

Wed Apr 6 19:37:10 CEST 2011

I think you may need to contact Fabian Scheipl
<fabian.scheipl at stat.uni-muenchen.de>, the author of the amer package
to get advice on this question.  (And, by the way, thanks for making
the data available - it helps a lot in reproducing the problem.)

On Thu, Mar 31, 2011 at 7:33 PM, Matthew Finkbeiner
<matthew.finkbeiner at mq.edu.au> wrote:
> I am new to mixed-effects modeling, so this is probably an embarrassingly
> easy problem to solve.  I have a repeated measures design where, in each
> block of trials, subjects (N=10) performed a reaching response to 10
> different target stimuli that were presented 10 times each (once per level
> of 'FactorA'), for a total of 100 trials per block.  There were 5
> experimental blocks.  The dependent measure is the 'angle' of the reaching
> trajectory, and this was sampled/calculated at 5 different times per
> response.
>
> I would like to plot angles ~ Sample for each level of FactorA (with CIs),
> but have not been able to figure out how to do this.  The amer package
> provides a nice example of how to do this, using the following call to amer
> with the 'dog' data:
> d1<- amer(y ~ -1 + group + tp(time, k=6, by = group) + (1|dog), data = dog)
> plotF(d1, legend = 'topleft', lwd=3)
>
> But when I run a similarly structured analysis with my data, it fails.  For
> example:
> fm1<- amer(angles ~ -1 + FactorA + tp(Sample, k=4, by = FactorA) + (1|Subj)
> + (1|TargItem), data = VelInts)
> Error in `*tmp*`[[ind]] : subscript out of bounds
>
> I suspect that this is because of the repeated-measures design, but don't
> know how to get around that.
>
> I have run the following analysis successfully with lmer:
>
> lmer1 <- lmer(angles ~ FactorA + Sample + (1|TargItem) + (0+key|Subj), data
> = VelInts)
>
> which produces the following:
>
> Fixed effects:
>              Estimate Std. Error t value
> (Intercept)  1.6177246  0.0124393  130.05
> FactorA2    -0.0300888  0.0110805   -2.72
> FactorA3     0.0746474  0.0111035    6.72
> FactorA4     0.0506684  0.0110861    4.57
> FactorA5     0.1578699  0.0111209   14.20
> FactorA6     0.0272700  0.0110862    2.46
> FactorA7     0.0613134  0.0111036    5.52
> FactorA8    -0.0470794  0.0111091   -4.24
> FactorA9    -0.0248243  0.0111382   -2.23
> FactorA10   -0.1258571  0.0110807  -11.36
> Sample      -0.0038139  0.0000701  -54.41
>
>
> But I can't figure out how to go from this to visualizing 'angles ~ Sample'
> for each level of FactorA.
>
> Any help would be much appreciated.
>
> The structure of my data is as follows:
>
>> str(VelInts)
> 'data.frame':   24205 obs. of  31 variables:
>  $ .id         : chr  "101" "101" "101" "101" ...
>  $ Subj        : Factor w/ 10 levels "S1","S2","S3",..: 1 1 1 1 1 1 1 1 1 1
> ...
>  $ block       : num  2 2 2 2 2 2 2 2 2 2 ...
>  $ trial_n     : num  1 1 1 1 1 2 2 2 2 2 ...
>  $ FactorA     : Factor w/ 10 levels "1","2","3","4",..: 6 6 6 6 6 1 1 1 1 1
> ...
>  $ fixedpattern: num  1 1 1 1 1 0 0 0 0 0 ...
>  $ pattern     : num  2 2 2 2 2 0 0 0 0 0 ...
>  $ TargPos     : num  2 2 2 2 2 1 1 1 1 1 ...
>  $ LiftOff     : num  241 241 241 241 241 290 290 290 290 290 ...
>  $ Acc         : num  1 1 1 1 1 1 1 1 1 1 ...
>  $ FixDur      : num  529 529 529 529 529 662 662 662 662 662 ...
>  $ Missing     : num  0 0 0 0 0 0 0 0 0 0 ...
>  $ TargItem    : num  6 6 6 6 6 1 1 1 1 1 ...
>  $ RT          : num  712 712 712 712 712 959 959 959 959 959 ...
>  $ x           : num  0.477 25.26 68.063 121.514 213.708 ...
>  $ y           : num  7.09 30.97 69.54 149.23 270.24 ...
>  $ z           : num  10.8 73.9 138.2 191.8 240.5 ...
>  $ t           : num  270 336 402 469 535 ...
>  $ presT       : num  270 336 402 468 534 ...
>  $ key         : num  101 101 101 101 101 102 102 102 102 102 ...
>  $ pathoffset  : num  0.00583 -0.00899 -0.03706 -0.043 -0.06842 ...
>  $ xvel        : num  92.8 571.6 746.9 869.1 1803.9 ...
>  $ yvel        : num  150 431 817 1476 2092 ...
>  $ tangvel     : num  177 716 1107 1713 2762 ...
>  $ xaccel      : num  5310 2967 3155 5330 1387 ...
>  $ yaccel      : num  7812 1132 10343 6860 2647 ...
>  $ angles      : num  1.503 0.887 0.796 0.887 0.902 ...
>  $ Sample      : int  0 25 50 75 100 0 25 50 75 100 ...
>  $ G           : num  6 6 6 6 6 1 1 1 1 1 ...
>  $ S           : num  51 51 51 51 51 1 1 1 1 1 ...
>  $ SK          : Factor w/ 4841 levels "100103","100113",..: 2228 2228 2228
> 2228 2228 96 96 96 96 96 ...
>
>
> And if it is helpful, the full data frame can be accessed here:
>  http://www.maccs.mq.edu.au/~mfinkbei/VelInts.RData
>
>
> Thank you kindly for suggestions,
>
> Matthew
>
>        [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>