Skip to content

Calculating fixed effect contrasts with log-transformed data

4 messages · ONKELINX, Thierry, Gus Jespersen, Ben Bolker

Original
Gus Jespersen
# 
Greetings,
I doubt this is a particularly interesting question for you mixed
model gurus, but here goes.  As you can see in the output below, I
have a model with twelve fixed effect parameters.  I am interested in
each of the "Treatment" vs. "Control" comparisons for each "site"(in
each fixed effect parameter name, these are specified by the text
immediately following "sitett"). To produce a 95% CI for such a
comparison I was advised to take two steps:

(1) Subtract the Control parameter estimate from the Treatment
parameter estimate for each site.
(2) Compute the SE for this comparison via:  sqrt( var(treatment) +
var(control) - 2*cov(treatmentt,control)).  To get these values I am
using the vcov matrix for the model.

When I move to log10-transformed data, I am thinking I should
backtransform the fixed effects and SE's before moving ahead  with the
Control-Treatment comparisons.  However, the calculations become more
problematic as ( var(treatment) + var(control) -
2*cov(treatmentt,control)) is consistently negative.  I am uncertain
on how to proceed here.  Any advice would be much appreciated.

Thank you,
Gus

Data: data.file.final
Models:
Mod.NO3.1.2: NO3Nyearone ~ 1 + (1 | pr)
Mod.NO3.1.1: NO3Nyearone ~ 1 + sitett + (1 | pr)
            Df    AIC    BIC  logLik  Chisq Chi Df Pr(>Chisq)
Mod.NO3.1.2  3 1163.5 1172.2 -578.72
Mod.NO3.1.1 14 1155.8 1196.7 -563.90 29.637     11   0.001806 **
---
Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1
Linear mixed model fit by REML
Formula: NO3Nyearone ~ 1 + sitett + (1 | pr)
   Data: data.file.final
  AIC  BIC logLik deviance REMLdev
 1098 1139 -534.8     1128    1070
Random effects:
 Groups   Name        Variance Std.Dev.
 pr       (Intercept)  33.348   5.7747
 Residual             210.115  14.4954
Number of obs: 137, groups: pr, 72

Fixed effects:
                             Estimate Std. Error t value
(Intercept)                    20.118      4.701   4.280
sitettLepAddition Treatment     3.032      6.069   0.500
sitettMossAddition Control      5.677      6.809   0.834
sitettMossAddition Treatment    9.418      6.648   1.417
sitettMossRemoval Control      -9.951      6.510  -1.529
sitettMossRemoval Treatment    -9.601      6.510  -1.475
sitettSaddle Control          -10.985      6.510  -1.687
sitettSaddle Treatment        -12.269      6.648  -1.846
sitettToeAdditions Control      0.932      6.510   0.143
sitettToeAdditions Treatment  -11.678      6.809  -1.715
sitettToeRemoval Control      -12.351      6.510  -1.897
sitettToeRemoval Treatment    -13.168      6.510  -2.023
2 days later
ONKELINX, Thierry
# 
Dear Gus,

Have a look at glht() from the multcomp package. It allows you to define the contrasts that you are interested in.

Best regards,

Thierry

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
+ 32 2 525 02 51
+ 32 54 43 61 85
Thierry.Onkelinx at inbo.be
www.inbo.be

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

-----Oorspronkelijk bericht-----
Van: r-sig-mixed-models-bounces at r-project.org [mailto:r-sig-mixed-models-bounces at r-project.org] Namens Gus Jespersen
Verzonden: vrijdag 13 juli 2012 20:33
Aan: r-sig-mixed-models at r-project.org
Onderwerp: [R-sig-ME] Calculating fixed effect contrasts with log-transformed data

Greetings,
I doubt this is a particularly interesting question for you mixed model gurus, but here goes.  As you can see in the output below, I have a model with twelve fixed effect parameters.  I am interested in each of the "Treatment" vs. "Control" comparisons for each "site"(in each fixed effect parameter name, these are specified by the text immediately following "sitett"). To produce a 95% CI for such a comparison I was advised to take two steps:

(1) Subtract the Control parameter estimate from the Treatment parameter estimate for each site.
(2) Compute the SE for this comparison via:  sqrt( var(treatment) +
var(control) - 2*cov(treatmentt,control)).  To get these values I am using the vcov matrix for the model.

When I move to log10-transformed data, I am thinking I should backtransform the fixed effects and SE's before moving ahead  with the Control-Treatment comparisons.  However, the calculations become more problematic as ( var(treatment) + var(control) -
2*cov(treatmentt,control)) is consistently negative.  I am uncertain on how to proceed here.  Any advice would be much appreciated.

Thank you,
Gus

Data: data.file.final
Models:
Mod.NO3.1.2: NO3Nyearone ~ 1 + (1 | pr)
Mod.NO3.1.1: NO3Nyearone ~ 1 + sitett + (1 | pr)
            Df    AIC    BIC  logLik  Chisq Chi Df Pr(>Chisq)
Mod.NO3.1.2  3 1163.5 1172.2 -578.72
Mod.NO3.1.1 14 1155.8 1196.7 -563.90 29.637     11   0.001806 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 Linear mixed model fit by REML
Formula: NO3Nyearone ~ 1 + sitett + (1 | pr)
   Data: data.file.final
  AIC  BIC logLik deviance REMLdev
 1098 1139 -534.8     1128    1070
Random effects:
 Groups   Name        Variance Std.Dev.
 pr       (Intercept)  33.348   5.7747
 Residual             210.115  14.4954
Number of obs: 137, groups: pr, 72

Fixed effects:
                             Estimate Std. Error t value
(Intercept)                    20.118      4.701   4.280
sitettLepAddition Treatment     3.032      6.069   0.500
sitettMossAddition Control      5.677      6.809   0.834
sitettMossAddition Treatment    9.418      6.648   1.417
sitettMossRemoval Control      -9.951      6.510  -1.529
sitettMossRemoval Treatment    -9.601      6.510  -1.475
sitettSaddle Control          -10.985      6.510  -1.687
sitettSaddle Treatment        -12.269      6.648  -1.846
sitettToeAdditions Control      0.932      6.510   0.143
sitettToeAdditions Treatment  -11.678      6.809  -1.715
sitettToeRemoval Control      -12.351      6.510  -1.897
sitettToeRemoval Treatment    -13.168      6.510  -2.023



--
R. Gus Jespersen
PhD Candidate
College of Forest Resources
University of Washington
Box 352100
Seattle, WA 98195-2100
(206) 543-5777
jesper at u.washington.edu

_______________________________________________
R-sig-mixed-models at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
* * * * * * * * * * * * * D I S C L A I M E R * * * * * * * * * * * * *
Dit bericht en eventuele bijlagen geven enkel de visie van de schrijver weer en binden het INBO onder geen enkel beding, zolang dit bericht niet bevestigd is door een geldig ondertekend document.
The views expressed in this message and any annex are purely those of the writer and may not be regarded as stating an official position of INBO, as long as the message is not confirmed by a duly signed document.
1 day later
Gus Jespersen
# 
Thank you Thierry,
I have looked through the glht function in multcomp, and have two
further questions:

(1) For the lmer output below, I would like to go through and
calculate a 95% CI for the difference between each "Treatment" and
"Control" fixed effect.  Based on my reading of the glht instructions,
this should look like:
Yet I get the following error message:

Error in parse(text = ex[i]) : <text>:1:20: unexpected symbol
1: sitettMossAddition Treatment
                                       ^
Any ideas on what I'm doing wrong here?

(2) As you can see, I am working with a log10 transformed response
variable.  I'd like to stay with this for homog. of variance reasons,
and for reporting the "Treatment -Control" CI previously mentioned,
I'd like to report the backtransformed limits of the CI.  At what
point in this process should the back-transformation happen?  When
attempting this calculation without glht, I am uncertain of where in
the process to back-transform as well.  I had been hoping to simply
use 10^ for each fixed effect and its SE, as well as each element of
the vcov matrix, but I fear I am overlooking some basic math here.

Finally, I have pasted my data below the model output, in case this is
helpful.

Thank you,
Gus

###  Model Code  ###

Mod.NO3.1.1<-lmer(NO3Nyearone ~ 1 + sitett +(1 | pr), data=data.file.final)
Mod.NO3.1.2<-lmer(log10(NO3Nyearone)~ 1  + (1 | pr), data=data.file.final)

anova(Mod.NO3.1.1,Mod.NO3.1.2)
Mod.NO3.1.1

###   Output   ###
Data: data.file.final
Models:
Mod.NO3.1.2: log10(NO3Nyearone) ~ 1 + (1 | pr)
Mod.NO3.1.1: log10(NO3Nyearone) ~ 1 + sitett + (1 | pr)
            Df    AIC    BIC  logLik  Chisq Chi Df Pr(>Chisq)
Mod.NO3.1.2  3 99.830 108.59 -46.915
Mod.NO3.1.1 14 74.329 115.21 -23.165 47.501     11  1.752e-06 ***
---
Signif. codes:  0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1
Linear mixed model fit by REML
Formula: log10(NO3Nyearone) ~ 1 + sitett + (1 | pr)
   Data: data.file.final
   AIC   BIC logLik deviance REMLdev
 110.9 151.8 -41.46    46.33   82.92
Random effects:
 Groups   Name        Variance  Std.Dev.
 pr       (Intercept) 0.0029239 0.054073
 Residual             0.0871127 0.295149
Number of obs: 137, groups: pr, 72

Fixed effects:
                             Estimate Std. Error t value
(Intercept)                                        1.24011    0.09047  13.708
sitettLepAddition Treatment       -0.03859    0.12329  -0.313
sitettMossAddition Control           0.07747    0.13110   0.591
sitettMossAddition Treatment      0.13940    0.12794   1.090
sitettMossRemoval Control        -0.36994    0.12525  -2.954
sitettMossRemoval Treatment   -0.25789    0.12525  -2.059
sitettSaddle Control                     -0.33039    0.12525  -2.638
sitettSaddle Treatment                -0.46379    0.12794  -3.625
sitettToeAdditions Control          -0.17356    0.12525  -1.386
sitettToeAdditions Treatment     -0.33236    0.13110  -2.535
sitettToeRemoval Control          -0.38744    0.12525  -3.093
sitettToeRemoval Treatment     -0.44006    0.12525  -3.513

Correlation of Fixed Effects:
            (Intr) sttLAT sttMAC sttMAT sttMRC sttMRT stttSC stttST
sttTAC sttTAT sttTRC
stttLpAddtT -0.712
stttMssAddC -0.690  0.491
stttMssAddT -0.707  0.503  0.502
stttMssRmvC -0.722  0.514  0.498  0.511
stttMssRmvT -0.722  0.514  0.498  0.511  0.537
stttSddlCnt -0.722  0.514  0.498  0.511  0.522  0.522
stttSddlTrt -0.707  0.503  0.488  0.500  0.511  0.511  0.526
stttTAddtnC -0.722  0.514  0.498  0.511  0.522  0.522  0.522  0.511
stttTAddtnT -0.690  0.491  0.476  0.488  0.498  0.498  0.498  0.488
0.513
stttTRmvlCn -0.722  0.514  0.498  0.511  0.522  0.522  0.522  0.511
0.522  0.498
stttTRmvlTr -0.722  0.514  0.498  0.511  0.522  0.522  0.522  0.511
0.522  0.498  0.537

#####   Data  ######

PlotID	Site	pr	tt	sitett	NO3Nyearone
3156	LepAddition	LeprariaAdditionone	Control	LepAddition Control	35
3155	LepAddition	LeprariaAdditionone	Treatment	LepAddition Treatment	105.6
3161	LepAddition	LeprariaAdditionfour	Treatment	LepAddition Treatment	44.8
3297	LepAddition	LeprariaAdditionseven	Control	LepAddition Control	40.6
3158	LepAddition	LeprariaAdditiontwo	Control	LepAddition Control	26.4
3162	LepAddition	LeprariaAdditionfour	Control	LepAddition Control	19.6
3293	LepAddition	LeprariaAdditioneight	Control	LepAddition Control	28.2
3157	LepAddition	LeprariaAdditiontwo	Treatment	LepAddition Treatment	21.8
2740	LepAddition	LeprariaAdditionthree	Control	LepAddition Control	21.8
2745	LepAddition	LeprariaAdditionnine	Treatment	LepAddition Treatment	20.2
2755	LepAddition	LeprariaAdditionsix	Treatment	LepAddition Treatment	7.8
3284	LepAddition	LeprariaAdditiontwelve	Treatment	LepAddition Treatment	16.4
2749	LepAddition	LeprariaAdditionfive	Control	LepAddition Control	14.6
2744	LepAddition	LeprariaAdditionnine	Control	LepAddition Control	13
2759	LepAddition	LeprariaAdditioneleven	Treatment	LepAddition Treatment	7.4
2758	LepAddition	LeprariaAdditioneleven	Control	LepAddition Control	7.4
3292	LepAddition	LeprariaAdditioneight	Treatment	LepAddition Treatment	12.8
3296	LepAddition	LeprariaAdditionseven	Treatment	LepAddition Treatment	9.6
2748	LepAddition	LeprariaAdditionfive	Treatment	LepAddition Treatment	12.6
3294	LepAddition	LeprariaAdditionthirteen	Treatment	LepAddition Treatment	7.8
3285	LepAddition	LeprariaAdditiontwelve	Control	LepAddition Control	8.4
2754	LepAddition	LeprariaAdditionsix	Control	LepAddition Control	8.4
2741	LepAddition	LeprariaAdditionthree	Treatment	LepAddition Treatment	11
3289	MossAddition	MossAddthree	Treatment	MossAddition Treatment	63.8
2751	MossAddition	MossAddsix	Control	MossAddition Control	46.8
3286	MossAddition	MossAddthree	Control	MossAddition Control	61.8
3288	MossAddition	MossAddeleven	Treatment	MossAddition Treatment	59.4
2747	MossAddition	MossAddseven	Treatment	MossAddition Treatment	54.3
2752	MossAddition	MossAddeight	Treatment	MossAddition Treatment	20.6
3280	MossAddition	MossAddtwo	Control	MossAddition Control	31.8
3163	MossAddition	MossAddtwelve	Control	MossAddition Control	37.8
2750	MossAddition	MossAddsix	Treatment	MossAddition Treatment	30.8
2743	MossAddition	MossAddfive	Treatment	MossAddition Treatment	25
3291	MossAddition	MossAddeleven	Control	MossAddition Control	20.2
3299	MossAddition	MossAddone	Control	MossAddition Control	25
3281	MossAddition	MossAddtwo	Treatment	MossAddition Treatment	16.3
2757	MossAddition	MossAddnine	Control	MossAddition Control	14.8
2753	MossAddition	MossAddeight	Control	MossAddition Control	21
3164	MossAddition	MossAddtwelve	Treatment	MossAddition Treatment	18
3160	MossAddition	MossAddfour	Treatment	MossAddition Treatment	17.6
2742	MossAddition	MossAddfive	Control	MossAddition Control	18.8
2756	MossAddition	MossAddnine	Treatment	MossAddition Treatment	11.6
3287	MossAddition	MossAddten	Control	MossAddition Control	10.8
3290	MossAddition	MossAddten	Treatment	MossAddition Treatment	7.6
2746	MossAddition	MossAddseven	Control	MossAddition Control	4.4
3242	MossRemoval	MossRemone	Control	MossRemoval Control	40.6
3245	MossRemoval	MossRemtwo	Treatment	MossRemoval Treatment	24.6
3253	MossRemoval	MossRemtwelve	Control	MossRemoval Control	7.6
3250	MossRemoval	MossRemnine	Treatment	MossRemoval Treatment	11
3240	MossRemoval	MossRemseven	Treatment	MossRemoval Treatment	6.6
3257	MossRemoval	MossRemfive	Control	MossRemoval Control	7.4
3244	MossRemoval	MossRemtwo	Control	MossRemoval Control	9.2
3252	MossRemoval	MossRemtwelve	Treatment	MossRemoval Treatment	6.8
3254	MossRemoval	MossRemsix	Treatment	MossRemoval Treatment	17.4
3241	MossRemoval	MossRemseven	Control	MossRemoval Control	9
3255	MossRemoval	MossRemsix	Control	MossRemoval Control	6
2687	MossRemoval	MossRemten	Treatment	MossRemoval Treatment	11.8
3246	MossRemoval	MossRemthree	Control	MossRemoval Control	11
3243	MossRemoval	MossRemone	Treatment	MossRemoval Treatment	10.2
3259	MossRemoval	MossRemfour	Treatment	MossRemoval Treatment	9
3247	MossRemoval	MossRemthree	Treatment	MossRemoval Treatment	6.8
3251	MossRemoval	MossRemnine	Control	MossRemoval Control	8.4
2684	MossRemoval	MossRemeleven	Treatment	MossRemoval Treatment	7.2
3248	MossRemoval	MossRemeight	Treatment	MossRemoval Treatment	7
3249	MossRemoval	MossRemeight	Control	MossRemoval Control	10
3258	MossRemoval	MossRemfour	Control	MossRemoval Control	8.8
3256	MossRemoval	MossRemfive	Treatment	MossRemoval Treatment	7.8
2686	MossRemoval	MossRemten	Control	MossRemoval Control	1
2685	MossRemoval	MossRemeleven	Control	MossRemoval Control	3
2999	Saddle	LeprariaRemovalnine	Treatment	Saddle Treatment	21.8
2998	Saddle	LeprariaRemovalnine	Control	Saddle Control	12
2956	Saddle	LeprariaRemovaltwelve	Control	Saddle Control	11.6
2485	Saddle	LeprariaRemovalthree	Control	Saddle Control	13.4
2489	Saddle	LeprariaRemovalten	Control	Saddle Control	3
2497	Saddle	LeprariaRemovalsix	Control	Saddle Control	13.2
2487	Saddle	LeprariaRemovalten	Treatment	Saddle Treatment	3.4
2958	Saddle	LeprariaRemovaleleven	Treatment	Saddle Treatment	7.4
2491	Saddle	LeprariaRemovalone	Treatment	Saddle Treatment	12.2
2483	Saddle	LeprariaRemovalfour	Control	Saddle Control	16.4
2484	Saddle	LeprariaRemovalthree	Treatment	Saddle Treatment	14.4
2498	Saddle	LeprariaRemovalseven	Control	Saddle Control	9.8
2493	Saddle	LeprariaRemovaltwo	Control	Saddle Control	8.6
2481	Saddle	LeprariaRemovaleight	Treatment	Saddle Treatment	10.8
2480	Saddle	LeprariaRemovaleight	Control	Saddle Control	5.4
2959	Saddle	LeprariaRemovaleleven	Control	Saddle Control	6.6
2496	Saddle	LeprariaRemovalsix	Treatment	Saddle Treatment	3.6
2494	Saddle	LeprariaRemovalfive	Control	Saddle Control	4.2
2490	Saddle	LeprariaRemovalone	Control	Saddle Control	5.4
2482	Saddle	LeprariaRemovalfour	Treatment	Saddle Treatment	3.2
2495	Saddle	LeprariaRemovalfive	Treatment	Saddle Treatment	3.4
2499	Saddle	LeprariaRemovalseven	Treatment	Saddle Treatment	3
2492	Saddle	LeprariaRemovaltwo	Treatment	Saddle Treatment	2.8
2577	ToeAdditions	FlavocetrariaAdditiontwelve	Control	ToeAdditions Control	121.4
2581	ToeAdditions	FlavocetrariaAdditionone	Control	ToeAdditions Control	17.2
2576	ToeAdditions	FlavocetrariaAdditionfour	Control	ToeAdditions Control	20.8
2568	ToeAdditions	FlavocetrariaAdditiontwo	Control	ToeAdditions Control	17.6
2580	ToeAdditions	FlavocetrariaAdditionthree	Control	ToeAdditions Control	18.4
2572	ToeAdditions	FlavocetrariaAdditioneight	Treatment	ToeAdditions
Treatment	14.6
2573	ToeAdditions	FlavocetrariaAdditionfour	Treatment	ToeAdditions Treatment	12
2571	ToeAdditions	FlavocetrariaAdditionfive	Control	ToeAdditions Control	10.8
2574	ToeAdditions	FlavocetrariaAdditionseven	Treatment	ToeAdditions
Treatment	13.6
2588	ToeAdditions	FlavocetrariaAdditiontwo	Treatment	ToeAdditions Treatment	10.2
2575	ToeAdditions	FlavocetrariaAdditionsix	Control	ToeAdditions Control	11.8
2579	ToeAdditions	FlavocetrariaAdditionnine	Treatment	ToeAdditions
Treatment	13.8
2584	ToeAdditions	FlavocetrariaAdditionseven	Control	ToeAdditions Control	8.6
2582	ToeAdditions	FlavocetrariaAdditionten	Control	ToeAdditions Control	10
2569	ToeAdditions	FlavocetrariaAdditionsix	Treatment	ToeAdditions Treatment	8.6
2583	ToeAdditions	FlavocetrariaAdditionone	Treatment	ToeAdditions Treatment	7.2
2578	ToeAdditions	FlavocetrariaAdditioneleven	Control	ToeAdditions Control	10.2
2570	ToeAdditions	FlavocetrariaAdditionfive	Treatment	ToeAdditions Treatment	9.8
2585	ToeAdditions	FlavocetrariaAdditionthree	Treatment	ToeAdditions
Treatment	6.6
2589	ToeAdditions	FlavocetrariaAdditionten	Treatment	ToeAdditions Treatment	4
2683	ToeAdditions	FlavocetrariaAdditionnine	Control	ToeAdditions Control	5
2586	ToeAdditions	FlavocetrariaAdditiontwelve	Treatment	ToeAdditions
Treatment	6.2
2682	ToeAdditions	FlavocetrariaAdditioneight	Control	ToeAdditions Control	0.8
2681	ToeAdditions	FlavocetrariaAdditioneleven	Treatment	ToeAdditions Treatment	5
2986	ToeRemoval	FlavocetrariaRemovalfive	Treatment	ToeRemoval Treatment	16.8
2991	ToeRemoval	FlavocetrariaRemovalfour	Control	ToeRemoval Control	16.6
2983	ToeRemoval	FlavocetrariaRemovaltwo	Control	ToeRemoval Control	9.6
2994	ToeRemoval	FlavocetrariaRemovalnine	Control	ToeRemoval Control	7.8
2989	ToeRemoval	FlavocetrariaRemovalthree	Treatment	ToeRemoval Treatment	7.2
2796	ToeRemoval	FlavocetrariaRemovaleight	Control	ToeRemoval Control	9.2
2982	ToeRemoval	FlavocetrariaRemovaltwo	Treatment	ToeRemoval Treatment	7.2
2985	ToeRemoval	FlavocetrariaRemovalone	Treatment	ToeRemoval Treatment	6.8
2984	ToeRemoval	FlavocetrariaRemovalone	Control	ToeRemoval Control	2.6
2799	ToeRemoval	FlavocetrariaRemovalten	Control	ToeRemoval Control	8
2797	ToeRemoval	FlavocetrariaRemovaleight	Treatment	ToeRemoval Treatment	9.4
2992	ToeRemoval	FlavocetrariaRemovalsix	Control	ToeRemoval Control	8.4
2981	ToeRemoval	FlavocetrariaRemovaltwelve	Control	ToeRemoval Control	7.8
2996	ToeRemoval	FlavocetrariaRemovalseven	Control	ToeRemoval Control	6.6
2980	ToeRemoval	FlavocetrariaRemovaltwelve	Treatment	ToeRemoval Treatment	7.8
2987	ToeRemoval	FlavocetrariaRemovalfive	Control	ToeRemoval Control	6.4
2988	ToeRemoval	FlavocetrariaRemovalthree	Control	ToeRemoval Control	5.4
2990	ToeRemoval	FlavocetrariaRemovalfour	Treatment	ToeRemoval Treatment	6
2993	ToeRemoval	FlavocetrariaRemovalsix	Treatment	ToeRemoval Treatment	4.2
2689	ToeRemoval	FlavocetrariaRemovaleleven	Control	ToeRemoval Control	4.8
2997	ToeRemoval	FlavocetrariaRemovalseven	Treatment	ToeRemoval Treatment	4.6
2798	ToeRemoval	FlavocetrariaRemovalten	Treatment	ToeRemoval Treatment	6
2688	ToeRemoval	FlavocetrariaRemovaleleven	Treatment	ToeRemoval Treatment	4.4
2995	ToeRemoval	FlavocetrariaRemovalnine	Treatment	ToeRemoval Treatment	3





On Mon, Jul 16, 2012 at 1:53 AM, ONKELINX, Thierry
<Thierry.ONKELINX at inbo.be> wrote:

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Ben Bolker
    

    

      
      #
    

  


  

      
Gus Jespersen <jesper <at> u.washington.edu> writes:

            

      
      
      
    

        
[snip]

            

      
      
      
    

        
It is very likely that the glht function is having trouble
with the spaces in your level names.  I would strongly suggest
that you reformulate them as legal R variable names: something like

     levels(mydata$myfactor) <- make.names(levels(mydata$myfactor))
should work.

            

      
      
      
    

        
Yes, you are.  You need to back-transform the confidence intervals,
not the elements and their standard errors.  The basic math you
are thinking about is

  10^(est+1.96*stderr) !== 10^est+10^(1.96*stderr)

If you back-transform first, the RHS is what you will be doing;
you want the LHS (this is for the upper CI, the obvious parallel
applies for the lower CI)