Issues with Mixed Models

39
Issues with Mixed Models

description

Issues with Mixed Models. Model doesn’t converge…. OR. Convergence. Likelihood Landscape. Likelihood Landscape. Maximum Likelihood Estimation. Mean. Maximum Likelihood Estimation. Mean. Maximum Likelihood Estimation. Mean. Maximum Likelihood Estimation. Mean. - PowerPoint PPT Presentation

Transcript of Issues with Mixed Models

Page 1: Issues with Mixed Models

Issues withMixed Models

Page 2: Issues with Mixed Models

Model doesn’t converge…

OR

Page 3: Issues with Mixed Models

Convergence

Page 4: Issues with Mixed Models

Likelihood Landscape

Page 5: Issues with Mixed Models

Likelihood Landscape

Page 6: Issues with Mixed Models

Maximum Likelihood Estimation

Mea

n

Page 7: Issues with Mixed Models

Maximum Likelihood Estimation

Mea

n

Page 8: Issues with Mixed Models

Maximum Likelihood Estimation

Mea

n

Page 9: Issues with Mixed Models

Maximum Likelihood Estimation

Mea

n

Page 10: Issues with Mixed Models

Maximum Likelihood Estimation

Likelihood = the probability of seeing the data we actually collected given a particular model

Maximum Likelihood Estimates = those values that make the observed data most likely to have happened

Page 11: Issues with Mixed Models

Sources of Convergence Problems

• You estimate more parameters than data (or, in general, too many parameters

• Severe collinearity (e.g., two predictors are exactly correlated)

• Missing cells in your design

• Predictors of vastly different metrics

Page 12: Issues with Mixed Models

Failure to converge

GENDERATTITUDE male

femalepolite 16

0informal 1632

… and then trying to test the ATTITUDE*GENDER interaction

Page 13: Issues with Mixed Models

How can this happen?

“Death by Design”

(coined byRoger Mundry)

Page 14: Issues with Mixed Models

designanalysis

Page 15: Issues with Mixed Models

Solutions to Convergence Problems

• Drop a random slope(not preferred, should be reported)

• Drop subjects/items for which there is not enough data (not preferred, should be reported)

• Rescale variables so that they lie range between 0 and 1; or make them on similar metrics overall

• Center continuous predictors

• Nonlinear transformations of skewed predictors

Page 16: Issues with Mixed Models

Solutions to Convergence Problems

• Change order of variable names in model formula

• Have a balanced and complete design

Page 17: Issues with Mixed Models

p-values

Page 18: Issues with Mixed Models

The p-value conundrum

What are the degrees of freedom?

How to get p-values out ofmixed models is not entirely straightforward…

DouglasBates

Page 19: Issues with Mixed Models

“There are a number of ways to compute p-values from LMEMs,

none of which is uncontroversially the best.”

Barr et al. (2013)

Page 20: Issues with Mixed Models

Ways to get p-values

• t-test/F-test with normal approximation• Likelihood Ratio Test• Boostrapping• Permutation• Markov Chain Monte Carlo (MCMC)

Page 21: Issues with Mixed Models

Getting p-vals with normal approximation

xmdlcoefs=data.frame(summary(xmdl)@coefs)coefs$p = 2*(1-pnorm(abs(coefs$t.value)))coefs

Page 22: Issues with Mixed Models

Function for getting p-vals with normal approximation

create.sig.table = function(x){

coefs=data.frame(summary(x)@coefs)coefs$p = 2*(1-pnorm(abs(coefs$t.value)))coefs$sig = character(nrow(coefs))coefs[which(coefs$p < 0.05),]$sig = "*"coefs[which(coefs$p < 0.01),]$sig = "**"coefs[which(coefs$p < 0.001),]$sig = "***"return(coefs)

}

Page 23: Issues with Mixed Models

Likelihood Ratio Test

First model needs to be nested in second

Page 24: Issues with Mixed Models

Likelihood RatioThe likelihood ratio expresses how many times more likely the data are under one model than the other

Mea

n

Mea

n

Page 25: Issues with Mixed Models

Likelihood Ratio Test

Page 26: Issues with Mixed Models

Likelihood Ratio Test

Page 27: Issues with Mixed Models

Important whendoing likelihood ratio tests

lmer(…,REML=FALSE)

http://anythingbutrbitrary.blogspot.com/2012/06/random-regression-coefficients-using.html

Page 28: Issues with Mixed Models

Final issue:Random slopes

Page 29: Issues with Mixed Models

DANGEROUS!!!

Random intercept onlymodels are known to bevery anti-conservativein many circumstances

(cf. Barr et al., 2013,Schielzeth & Forstmeier, 2008)

Page 30: Issues with Mixed Models

Schielzeth & Forstmeier (2008)

Random intercept only

Page 31: Issues with Mixed Models

Type I error simulation

10 subjects

10 data points each

5 of those in condition A,5 in B

LRT intercept ML 0.052LRT slope ML 0.035LRT intercept REML 0.052LRT slope REML 0.035

z-test intercept ML 0.053z-test slope ML 0.039z-test intercept REML 0.054z-test slope REML 0.042

Page 32: Issues with Mixed Models

Add to this explicit subject slopesfor A/B

10 subjects

10 data points each

5 of those in condition A,5 in B

LRT intercept ML 0.24LRT slope ML 0.15LRT intercept REML 0.24LRT slope REML 0.069

z-test intercept ML 0.24z-test slope ML 0.079z-test intercept REML 0.25z-test slope REML 0.091

Page 33: Issues with Mixed Models

Add to this explicit subject slopesfor A/B

10 subjects

10 data points each

5 of those in condition A,5 in B

LRT intercept ML 0.24LRT slope ML 0.15LRT intercept REML 0.24LRT slope REML 0.069

z-test intercept ML 0.24z-test slope ML 0.079z-test intercept REML 0.25z-test slope REML 0.091

Page 34: Issues with Mixed Models

Add to this explicit subject slopesfor A/B + take item slopes

10 subjects

10 data points each

5 of those in condition A,5 in B

LRT intercept ML 0.18LRT slope ML 0.085LRT intercept REML 0.18LRT slope REML 0.052

z-test intercept ML 0.21z-test slope ML 0.064z-test intercept REML 0.23z-test slope REML 0.08

Page 35: Issues with Mixed Models

LRT intercept ML 0.18LRT slope ML 0.085LRT intercept REML 0.18LRT slope REML 0.052

z-test intercept ML 0.21z-test slope ML 0.064z-test intercept REML 0.23z-test slope REML 0.08

Add to this explicit subject slopesfor A/B + take item slopes

10 subjects

10 data points each

5 of those in condition A,5 in B

Page 36: Issues with Mixed Models

“Keep it maximal”

Page 37: Issues with Mixed Models

“Keep it maximal”

random effectsjustified by the design

vs.random effects

justified by the dataBarr et al.

(2013)

Page 38: Issues with Mixed Models

“Keep it maximal”

“for whatever fixed effects are of critical interest, the

corresponding random effects should be in that analysis”

Barr et al. (2013)

Page 39: Issues with Mixed Models

That’s it(for now)