[R] Help Interpreting Linear Mixed Model

[Thierry Onkelinx]

Hello Josh,

One is never too old to study ;-)

Your question seems quite broad. You might be better off to read some books
on mixed models (e.g. Pinheiro & Bates (2000) or Zuur et al (2009)) or try
to find a local statistician. Email is not a suitable medium to teach
statistics.

Note that r-sig-mixed-models is a more suitable list for _specific_
questions on mixed models.

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

2015-04-27 9:54 GMT+02:00 Joshua Dixon <joshuamichaeldixon op gmail.com>:

> Hello Thierry,
>
> No, this isn't homework. Not that young unfortunately.
>
> Josh
>
> On 27 Apr 2015, at 08:06, Thierry Onkelinx <thierry.onkelinx op inbo.be>
> wrote:
>
> Dear Josh,
>
> Is this homework? Because the list has a no homework policy.
>
> 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
>
> 2015-04-27 2:26 GMT+02:00 Joshua Dixon <joshuamichaeldixon op gmail.com>:
>
>> Hello!
>>
>> Very new to R (10 days), and I've run the linear mixed model, below.
>> Attempting to interpret what it means...  What do I need to look for?
>> Residuals, correlations of fixed effects?!
>>
>> How would I look at very specific interactions, such as PREMIER_LEAGUE
>> (Level) 18 (AgeGr) GK (Position) mean difference to CHAMPIONSHIP 18  GK?
>>
>> For reference my data set looks like this:
>>
>> Id Level AgeGr   Position Height Weight BMI YoYo
>> 7451 CHAMPIONSHIP 14 M NA 63 NA 80
>> 148 PREMIER_LEAGUE 16 D NA 64 NA 80
>> 10393 CONFERENCE 10 D NA 36 NA 160
>> 10200 CHAMPIONSHIP 10 F NA 46 NA 160
>> 1961 LEAGUE_TWO 13 GK NA 67 NA 160
>> 10428 CHAMPIONSHIP 10 GK NA 40 NA 160
>> 10541 LEAGUE_ONE 10 F NA 25 NA 160
>> 10012 CHAMPIONSHIP 10 GK NA 30 NA 160
>> 9895 CHAMPIONSHIP 10 D NA 36 NA 160
>>
>>
>> Many thanks in advance for time and help.  Really appreciate it.
>>
>> Josh
>>
>>
>> > summary(lmer(YoYo~AgeGr+Position+(1|Id)))
>> Linear mixed model fit by REML ['lmerMod']
>> Formula: YoYo ~ AgeGr + Position + (1 | Id)
>>
>> REML criterion at convergence: 125712.2
>>
>> Scaled residuals:
>>     Min      1Q  Median      3Q     Max
>> -3.4407 -0.5288 -0.0874  0.4531  4.8242
>>
>> Random effects:
>>  Groups   Name        Variance Std.Dev.
>>  Id       (Intercept) 15300    123.7
>>  Residual             16530    128.6
>> Number of obs: 9609, groups:  Id, 6071
>>
>> Fixed effects:
>>              Estimate Std. Error t value
>> (Intercept) -521.6985    16.8392  -30.98
>> AgeGr         62.6786     0.9783   64.07
>> PositionD    139.4682     7.8568   17.75
>> PositionM    141.2227     7.7072   18.32
>> PositionF    135.1241     8.1911   16.50
>>
>> Correlation of Fixed Effects:
>>           (Intr) AgeGr  PostnD PostnM
>> AgeGr     -0.910
>> PositionD -0.359 -0.009
>> PositionM -0.375  0.001  0.810
>> PositionF -0.349 -0.003  0.756  0.782
>> > model=lmer(YoYo~AgeGr+Position+(1|Id))
>> > summary(glht(model,linfct=mcp(Position="Tukey")))
>>
>>  Simultaneous Tests for General Linear Hypotheses
>>
>> Multiple Comparisons of Means: Tukey Contrasts
>>
>>
>> Fit: lmer(formula = YoYo ~ AgeGr + Position + (1 | Id))
>>
>> Linear Hypotheses:
>>             Estimate Std. Error z value Pr(>|z|)
>> D - GK == 0  139.468      7.857  17.751   <1e-04 ***
>> M - GK == 0  141.223      7.707  18.323   <1e-04 ***
>> F - GK == 0  135.124      8.191  16.496   <1e-04 ***
>> M - D == 0     1.754      4.799   0.366    0.983
>> F - D == 0    -4.344      5.616  -0.774    0.862
>> F - M == 0    -6.099      5.267  -1.158    0.645
>> ---
>> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
>> (Adjusted p values reported -- single-step method)
>>
>>         [[alternative HTML version deleted]]
>>
>> ______________________________________________
>> R-help op r-project.org mailing list -- To UNSUBSCRIBE and more, see
>> 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.
>
>
>

	[[alternative HTML version deleted]]