[R-sig-ME] Meaning of Corr of random-effects with a cross-level interaction
th|erry@onke||nx @end|ng |rom |nbo@be
Mon Sep 28 09:14:01 CEST 2020
The plot shows a set of dots on a vertical line indicating an upper bound
on ses. The few points outside the bound need investigating in a real
dataset. They might be wrong measurements.
If you have a response variable with a boundary, you need to use a
distribution that copes with that. Using a Gaussian distribution with
please of data close to a boundary, might lead to predictions and
confidence intervals outside of the boundary. I've seen people give a talk
on mortality with confidence intervals like (80%; 120%)...
I've no idea WHY you are getting the false convergence, just WHEN it starts
to kick in.
If you are cross posting you question, please do mention that.
ir. Thierry Onkelinx
Statisticus / Statistician
Vlaamse Overheid / Government of Flanders
INSTITUUT VOOR NATUUR- EN BOSONDERZOEK / RESEARCH INSTITUTE FOR NATURE AND
Team Biometrie & Kwaliteitszorg / Team Biometrics & Quality Assurance
thierry.onkelinx using inbo.be
Havenlaan 88 bus 73, 1000 Brussel
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
Op vr 25 sep. 2020 om 19:35 schreef Simon Harmel <sim.harmel using gmail.com>:
> Thank you Thierry! Would you please clarify one of your sentences: "both
> math and ses have bounds. Ses even seems to have some data above its
> upper bound."
> Specifically, would please clarify what you mean by "ses has some data
> above its upper bound"?(you mean the couple of outlying ses values in red
> as shown in your plot?)
> Of course, real world data always have some lower and upper bound based on
> the instrument (e.g., a math test) used to collect the data. But my
> question is what are the relative required lower and upper bounds on
> NUMERIC OUTCOME & NUMERIC PREDICTORS so we don't face convergence issues
> of the type I have shown in my question?
> Thank you,
> On Fri, Sep 25, 2020 at 3:03 AM Thierry Onkelinx <thierry.onkelinx using inbo.be>
>> Dear Simon,
>> A perfect correlation between random effect parameters indicates a
>> problem. Note the failed convergence warning.
>> Standardising ses makes things even worse: it yields a singular fit error.
>> Removing the random slope of ses or the sector interaction solves the
>> problem. i.e. the model runs and yields sensible output.
>> Looking at the data, it seems like both math and ses have bounds. Ses
>> even seems to have some data above its upper bound.
>> The model assumes no bounds in the response variable. Maybe this is the
>> cause of the problem.
>> ggplot(hsb, aes(x = ses, y = math, colour = factor(sector))) +
>> Best regards,
>> ir. Thierry Onkelinx
>> Statisticus / Statistician
>> Vlaamse Overheid / Government of Flanders
>> INSTITUUT VOOR NATUUR- EN BOSONDERZOEK / RESEARCH INSTITUTE FOR NATURE
>> AND FOREST
>> Team Biometrie & Kwaliteitszorg / Team Biometrics & Quality Assurance
>> thierry.onkelinx using inbo.be
>> Havenlaan 88 bus 73, 1000 Brussel
>> 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
>> Op do 24 sep. 2020 om 18:39 schreef Simon Harmel <sim.harmel using gmail.com>:
>>> Dear All,
>>> I had a quick question. I have a cross-level interaction in my model
>>> (ses*sector). My cluster-level predictor "sector" is a binary variable
>>> (0=Public, 1=Private). My level-1 predictor is numeric.
>>> QUESTION: The `Corr = 1` is indicating the correlation between
>>> intercepts and slopes across BOTH public & private sectors (like their
>>> average) OR something else?
>>> hsb <- read.csv('
>>> summary(lmer(math ~ ses*sector + (ses|sch.id), data = hsb))
>>> Random effects:
>>> Groups Name Variance Std.Dev. Corr
>>> sch.id (Intercept) 3.82107 1.9548
>>> ses 0.07587 0.2754 1.00
>>> Residual 36.78760 6.0653
>>> [[alternative HTML version deleted]]
>>> R-sig-mixed-models using r-project.org mailing list
[[alternative HTML version deleted]]
More information about the R-sig-mixed-models