Pushes and pulls from below: anatomical variation, articulation and sound change – Rmarkdown analysis script

Introduction

This is released under a GPLv2 license.

This script contains the analyses discussed in the paper.
Copyright (C) 2017-2018, Dan Dediu

This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version 2
of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, 
MA  02110-1301, USA.

Data

The whole dataset has 80 participants.

Variables related to the participant’s native language and their English proficiency

We added the following variables related to the participant’s native language (all are dichotomic yes/no) deduced from the segment inventories in PHOIBLE and the phonological descriptions in Wikipedia:

• Retroflexes: does the language have retroflex stops?
• R.trill: does the language use the alveaolr trill?
• R.retroflex: does the language use retroflex r?
• R.flap.tap: does the language use flap or tap?
• R.approximant: does the language use an r approximant?
• R.uvular: does the language use an uvular r?

Also, we coded self-declared English proficiency on a continuous scale 0 (none) to 11 (native) in English.proficiency.

Resolve multicolinearity by manually removing highly intercorrelated variables

Some anatomical variables are quite strongly correlated, and we manually removed one of each pair of variables inter-correlated (as measured by the absolute value of Pearson’s \(r\) ≥ 0.80) by chosing the variable of the pair that has the clearest interpretation while, at the same time, reducing the overall number of kept variables (i.e., some variables appear in more than one such pair).

Hierarchically clustering of the anatomical variables using psychometric principles

An alternative way is to use the iclust method (implemented by R’s psych package). The clusters fit the data well (cluster fit = 0.63, structure fit 0.82 and pattern RMSE 0.06).

A possible interpretation of the clusters is:

• C8: the “V” vs “U” shape of the mouth, with larger values representing a more “U”-shaped mouth;
• C12: the verticality of lower incisors, with larger meaning more vertical lower incisors;
• C37: the alveolar ridge, with higher values representing a bulgier, bigger alveolar ridge;
• C27: the length and height of the anterior hard palate, with higher values refering to a longer and higher anterior hard palate;
• C40: the overall shape of the mouth, with higher values refering to a longer, bigger mouth with a larger alveolar ridge;
• C44: the width of the mouth, with larger values standing for a narrower palate and mouth;
• C43: the length of the mouth, with larger values stading for a longer, heigher mouth;
• C41: the front of the mouth, with larger values representing a shorter upper jaw.

Principal Components Analysis (PCA) on the anatomical variables

We also conducted PCA (Principal Components Analysis) on the anatomical predictors and retained the first principal components that together explain approximately 90% of the variance. For this dataset, the first 13 PCs cumulatively explain 91.1% of the variance.

A possible interpretation of these PCs is:

• PC1 represents the overall dimensions of the mouth, with larger values meaning a smaller mouth;
• PC2 represents the front of the lower mouth, larger being a mouth with less overbite and more sublingual space;
• PC3 represents the width of the mouth, larger being a narrow mouth;
• PC4 represents the length and height of the front palate, larger being a heigher and longer front palate;
• PC5 represents the height of the palate, larger being lower palate;
• PC6 represents the alveolar ridge, larger being a heigher, larger ridge;
• PC7 represents the U vs V shaped mouth, larger being a “U”-shaped mouth;
• PC8 represents the low incisors, larger being less vertical incisors;
• PC9 represents the anterior mouth;
• PC10 represents the width and slope of the anterior palate roof;
• PC11 represents the overjet;
• PC12 represents the shape of the anterior mouth and alveolar ridge;
• PC13 represents the shape of the teeth and the alveaolr ridge.

The DVs

For “r” (as opposed to clicks), we only look at data from the MRI session (as we are interested in the articulation for now). For each participant we have 7 data points (or segments) (some might be missing):

• 1 static (sustained articulation MRI) – denoted as “segment 0”
• 6 dynamic (real-time MRI), of which:
  • 5 in the citation form (/ara/) – segments 1:4 and 6
  • 1 carrier /ara/ in the sentence “I said /ara/ for him” – segment 5

For each of these 7 segments, we have manually classified the following:

• r type (rtype): tip up (shortened as tu), tip up pharyngealized (tup), trill (trl), tap (tap), retroflex (rtr), bunched (bn), or double bunched (bnd);
  
• place of articulation (poa): dental (d), alveolar (a), alveolar rounded (ar), postalveolar (pa), postalveolar rounded (par), palatal (pl), or palatal rounded (plr);
• the shape of the dorsum of the tongue (i.e., the portion behind the blade but in front of the root): convex (cv), flat (f), or concave (cn).

and, just for segment 0 (the static, sustained MRI) we have also conducted PCA on the tongue shapes (see Figure 3 in the main paper).

Figure 1. POA vs tongue.PC1 x tongue.PC2

Figure 2. RTYPE vs tongue.PC1 x tongue.PC2

Figure 3. DORSUM vs tongue.PC1 x tongue.PC2

Figure 4. POA vs tongue.PC3 x tongue.PC4

Figure 5. RTYPE vs tongue.PC3 x tongue.PC4

Figure 6. DORSUM vs tongue.PC3 x tongue.PC4

Results

Linear regressions of each tPC on anatomy and covariates

Here we do linear regressions of the first 6 tPCs using the 3 data reduction sets + covariates (sex, group, phonexp & age) + native language properties (Retroflexes, R.trill, R.retroflex, R.flap.tap, R.approximant & R.uvular) + English proficiency, and we use an alpha-level of 0.01 (a Bonferroni correction is too conservative given that the 3 reduced sets are not independent and we have a very low statistical power in this case with N=79).

***tongue.PC1.AN***


Call:
lm(formula = tongue.PC1 ~ 1, data = d)

Residuals:
    Min      1Q  Median      3Q     Max 
-2.3347 -0.6818 -0.0734  0.6429  2.6104 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)
(Intercept) -1.25e-17   1.13e-01       0        1

Residual standard error: 1 on 78 degrees of freedom

***tongue.PC2.AN***


Call:
lm(formula = tongue.PC2 ~ overjet + m2Height + C.P2w, data = d)

Residuals:
    Min      1Q  Median      3Q     Max 
-1.9006 -0.5569  0.0049  0.4755  2.2163 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)   
(Intercept) -2.29e-16   9.91e-02    0.00   1.0000   
overjet      3.31e-01   1.02e-01    3.26   0.0017 **
m2Height     3.18e-01   9.99e-02    3.18   0.0022 **
C.P2w       -2.98e-01   1.02e-01   -2.93   0.0046 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.88 on 75 degrees of freedom
Multiple R-squared:  0.253, Adjusted R-squared:  0.223 
F-statistic: 8.48 on 3 and 75 DF,  p-value: 6.35e-05

***tongue.PC3.AN***


Call:
lm(formula = tongue.PC3 ~ R.trill, data = d)

Residuals:
    Min      1Q  Median      3Q     Max 
-2.1498 -0.6879  0.0698  0.7231  2.0673 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)   
(Intercept)   0.0852     0.1105    0.77    0.443   
R.trillyes   -1.3457     0.4391   -3.07    0.003 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.95 on 77 degrees of freedom
Multiple R-squared:  0.109, Adjusted R-squared:  0.0972 
F-statistic: 9.39 on 1 and 77 DF,  p-value: 0.003

***tongue.PC4.AN***


Call:
lm(formula = tongue.PC4 ~ sex + lowIA + C.P2l + aRBulge, data = d)

Residuals:
   Min     1Q Median     3Q    Max 
-1.930 -0.479  0.108  0.396  1.841 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept)   0.3223     0.1087    2.97  0.00406 ** 
sexF         -0.8487     0.1783   -4.76  9.4e-06 ***
lowIA        -0.2714     0.0873   -3.11  0.00267 ** 
C.P2l         0.3481     0.0882    3.95  0.00018 ***
aRBulge       0.3631     0.0880    4.13  9.5e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.76 on 74 degrees of freedom
Multiple R-squared:  0.459, Adjusted R-squared:  0.429 
F-statistic: 15.7 on 4 and 74 DF,  p-value: 2.47e-09

***tongue.PC5.AN***


Call:
lm(formula = tongue.PC5 ~ 1, data = d)

Residuals:
   Min     1Q Median     3Q    Max 
-2.702 -0.730  0.108  0.584  3.186 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.25e-17   1.13e-01       0        1

Residual standard error: 1 on 78 degrees of freedom

***tongue.PC6.AN***


Call:
lm(formula = tongue.PC6 ~ 1, data = d)

Residuals:
    Min      1Q  Median      3Q     Max 
-2.3561 -0.6723  0.0352  0.6542  2.6121 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)
(Intercept) -3.12e-17   1.13e-01       0        1

Residual standard error: 1 on 78 degrees of freedom

***tongue.PC1.CL***


Call:
lm(formula = tongue.PC1 ~ 1, data = d)

Residuals:
    Min      1Q  Median      3Q     Max 
-2.3347 -0.6818 -0.0734  0.6429  2.6104 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)
(Intercept) -1.25e-17   1.13e-01       0        1

Residual standard error: 1 on 78 degrees of freedom

***tongue.PC2.CL***


Call:
lm(formula = tongue.PC2 ~ age, data = d)

Residuals:
   Min     1Q Median     3Q    Max 
-2.037 -0.693 -0.134  0.660  1.994 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)   
(Intercept) -4.42e-17   1.06e-01    0.00   1.0000   
age          3.60e-01   1.06e-01    3.39   0.0011 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.94 on 77 degrees of freedom
Multiple R-squared:  0.13,  Adjusted R-squared:  0.119 
F-statistic: 11.5 on 1 and 77 DF,  p-value: 0.00111

***tongue.PC3.CL***


Call:
lm(formula = tongue.PC3 ~ R.trill, data = d)

Residuals:
    Min      1Q  Median      3Q     Max 
-2.1498 -0.6879  0.0698  0.7231  2.0673 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)   
(Intercept)   0.0852     0.1105    0.77    0.443   
R.trillyes   -1.3457     0.4391   -3.07    0.003 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.95 on 77 degrees of freedom
Multiple R-squared:  0.109, Adjusted R-squared:  0.0972 
F-statistic: 9.39 on 1 and 77 DF,  p-value: 0.003

***tongue.PC4.CL***


Call:
lm(formula = tongue.PC4 ~ sex + C44, data = d)

Residuals:
    Min      1Q  Median      3Q     Max 
-1.8497 -0.4725 -0.0213  0.3733  2.2412 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept)    0.276      0.125    2.20  0.03082 *  
sexF          -0.726      0.211   -3.45  0.00093 ***
C44           -0.298      0.103   -2.89  0.00498 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.86 on 76 degrees of freedom
Multiple R-squared:  0.285, Adjusted R-squared:  0.266 
F-statistic: 15.1 on 2 and 76 DF,  p-value: 2.94e-06

***tongue.PC5.CL***


Call:
lm(formula = tongue.PC5 ~ 1, data = d)

Residuals:
   Min     1Q Median     3Q    Max 
-2.702 -0.730  0.108  0.584  3.186 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.25e-17   1.13e-01       0        1

Residual standard error: 1 on 78 degrees of freedom

***tongue.PC6.CL***


Call:
lm(formula = tongue.PC6 ~ 1, data = d)

Residuals:
    Min      1Q  Median      3Q     Max 
-2.3561 -0.6723  0.0352  0.6542  2.6121 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)
(Intercept) -3.12e-17   1.13e-01       0        1

Residual standard error: 1 on 78 degrees of freedom

***tongue.PC1.PC***


Call:
lm(formula = tongue.PC1 ~ 1, data = d)

Residuals:
    Min      1Q  Median      3Q     Max 
-2.3347 -0.6818 -0.0734  0.6429  2.6104 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)
(Intercept) -1.25e-17   1.13e-01       0        1

Residual standard error: 1 on 78 degrees of freedom

***tongue.PC2.PC***


Call:
lm(formula = tongue.PC2 ~ age + PC1, data = d)

Residuals:
   Min     1Q Median     3Q    Max 
-2.062 -0.581 -0.172  0.561  2.180 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept) -5.12e-17   1.01e-01    0.00  1.00000    
age          3.53e-01   1.02e-01    3.47  0.00087 ***
PC1         -2.84e-01   1.02e-01   -2.78  0.00684 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.9 on 76 degrees of freedom
Multiple R-squared:  0.21,  Adjusted R-squared:  0.189 
F-statistic: 10.1 on 2 and 76 DF,  p-value: 0.000128

***tongue.PC3.PC***


Call:
lm(formula = tongue.PC3 ~ PC2 + R.trill, data = d)

Residuals:
    Min      1Q  Median      3Q     Max 
-1.8260 -0.6848 -0.0122  0.7270  2.0158 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept)   0.0934     0.1054    0.89  0.37849    
PC2           0.3033     0.1032    2.94  0.00435 ** 
R.trillyes   -1.4751     0.4211   -3.50  0.00077 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.91 on 76 degrees of freedom
Multiple R-squared:   0.2,  Adjusted R-squared:  0.179 
F-statistic: 9.48 on 2 and 76 DF,  p-value: 0.00021

***tongue.PC4.PC***


Call:
lm(formula = tongue.PC4 ~ PC1 + PC3 + PC8 + PC13, data = d)

Residuals:
    Min      1Q  Median      3Q     Max 
-2.2242 -0.4760 -0.0713  0.2643  2.2323 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept) -7.02e-19   9.31e-02    0.00  1.00000    
PC1         -2.64e-01   9.46e-02   -2.79  0.00664 ** 
PC3         -2.49e-01   9.39e-02   -2.65  0.00972 ** 
PC8         -3.41e-01   9.39e-02   -3.63  0.00051 ***
PC13        -2.87e-01   9.42e-02   -3.05  0.00321 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.83 on 74 degrees of freedom
Multiple R-squared:  0.351, Adjusted R-squared:  0.315 
F-statistic: 9.99 on 4 and 74 DF,  p-value: 1.62e-06

***tongue.PC5.PC***


Call:
lm(formula = tongue.PC5 ~ 1, data = d)

Residuals:
   Min     1Q Median     3Q    Max 
-2.702 -0.730  0.108  0.584  3.186 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)
(Intercept) 1.25e-17   1.13e-01       0        1

Residual standard error: 1 on 78 degrees of freedom

***tongue.PC6.PC***


Call:
lm(formula = tongue.PC6 ~ PC5 + PC10, data = d)

Residuals:
    Min      1Q  Median      3Q     Max 
-1.6913 -0.6092 -0.0459  0.4857  1.9041 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)   
(Intercept) -2.18e-17   1.04e-01    0.00   1.0000   
PC5         -3.00e-01   1.04e-01   -2.88   0.0052 **
PC10         3.07e-01   1.04e-01    2.94   0.0043 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.92 on 76 degrees of freedom
Multiple R-squared:  0.174, Adjusted R-squared:  0.152 
F-statistic:    8 on 2 and 76 DF,  p-value: 0.000703

Summary of linear regressions

Summary of the linear regression of tongue PCs on anatomy and covariates.

tPC	DV.set	adj.R2	IV	beta	p
tongue.PC1	AN	0.000	(Intercept)	0.000	1.000
tongue.PC2	AN	0.223	(Intercept)	0.000	1.000
tongue.PC2	AN	0.223	overjet	0.331	0.002
tongue.PC2	AN	0.223	m2Height	0.318	0.002
tongue.PC2	AN	0.223	C.P2w	-0.298	0.005
tongue.PC3	AN	0.097	(Intercept)	0.085	0.443
tongue.PC3	AN	0.097	R.trillyes	-1.346	0.003
tongue.PC4	AN	0.429	(Intercept)	0.322	0.004
tongue.PC4	AN	0.429	sexF	-0.849	0.000
tongue.PC4	AN	0.429	lowIA	-0.271	0.003
tongue.PC4	AN	0.429	C.P2l	0.348	0.000
tongue.PC4	AN	0.429	aRBulge	0.363	0.000
tongue.PC5	AN	0.000	(Intercept)	0.000	1.000
tongue.PC6	AN	0.000	(Intercept)	0.000	1.000
tongue.PC1	CL	0.000	(Intercept)	0.000	1.000
tongue.PC2	CL	0.119	(Intercept)	0.000	1.000
tongue.PC2	CL	0.119	age	0.360	0.001
tongue.PC3	CL	0.097	(Intercept)	0.085	0.443
tongue.PC3	CL	0.097	R.trillyes	-1.346	0.003
tongue.PC4	CL	0.266	(Intercept)	0.276	0.031
tongue.PC4	CL	0.266	sexF	-0.726	0.001
tongue.PC4	CL	0.266	C44	-0.298	0.005
tongue.PC5	CL	0.000	(Intercept)	0.000	1.000
tongue.PC6	CL	0.000	(Intercept)	0.000	1.000
tongue.PC1	PC	0.000	(Intercept)	0.000	1.000
tongue.PC2	PC	0.189	(Intercept)	0.000	1.000
tongue.PC2	PC	0.189	age	0.353	0.001
tongue.PC2	PC	0.189	PC1	-0.284	0.007
tongue.PC3	PC	0.179	(Intercept)	0.093	0.378
tongue.PC3	PC	0.179	PC2	0.303	0.004
tongue.PC3	PC	0.179	R.trillyes	-1.475	0.001
tongue.PC4	PC	0.315	(Intercept)	0.000	1.000
tongue.PC4	PC	0.315	PC1	-0.264	0.007
tongue.PC4	PC	0.315	PC3	-0.249	0.010
tongue.PC4	PC	0.315	PC8	-0.341	0.001
tongue.PC4	PC	0.315	PC13	-0.287	0.003
tongue.PC5	PC	0.000	(Intercept)	0.000	1.000
tongue.PC6	PC	0.152	(Intercept)	0.000	1.000
tongue.PC6	PC	0.152	PC5	-0.300	0.005
tongue.PC6	PC	0.152	PC10	0.307	0.004

Same linear regressions in an even more condensed form

Significant predictors for each DV and case using lm. The sign in parantheses is the sign of the estimated beta.

DV	AN	CL	PC
tongue.PC1
tongue.PC2	overjet (+), m2Height (+), C.P2w (-)	age (+)	age (+), PC1 (-)
tongue.PC3	R.trillyes (-)	R.trillyes (-)	PC2 (+), R.trillyes (-)
tongue.PC4	sexF (-), lowIA (-), C.P2l (+), aRBulge (+)	sexF (-), C44 (-)	PC1 (-), PC3 (-), PC8 (-), PC13 (-)
tongue.PC5
tongue.PC6			PC5 (-), PC10 (+)

Consolidating the value sets for rtype, poa and dorsal

Given that we have a relatively small sample size and 3 variables encoding tongue shape (excluding the tongue.PCs), we must consolidate the set of their discrete values. Based on a priori ideas, as well as on their plots in the tongue.PCs planes, we decided on the following consolidation schemes:

• rtype: {{tu, tup}, {bn, bnd}, {rtr}, {tap, trl}}
• poa: {{d, a, ar}, {pa, par, pl, plr}}
• lipround (lip rounding): {{d, a, pa, pl} , {ar, par, plr}}

• dorsum - as original

• Tip up / true retroflex unrounded: rtype = {tip up, tip up pharyngealized, retroflex} & poa = {dental, alveolar, postalveolar, palatal} & dorsum = {flat D, concave D, convex D}
• Tip up / true retroflex rounded: rtype = {tip up, tip up pharyngealized, retroflex} & poa = {alveolar rounded, postalveolar rounded, palatal rounded} & dorsum = {flat D, concave D, convex D}
• Bunched unrounded: rtype = {bunched, double bunched} & poa = {dental, alveolar, postalveolar, palatal} & dorsum = {flat D, concave D, convex D}

• Bunched rounded: rtype = { bunched, double bunched} & poa = { alveolar rounded, postalveolar rounded, palatal rounded} & dorsum = {flat D, concave D, convex D}

• Tip up / true retrofle: rounded | unrounded

• Bunched: bnchu | rounded | unrounded

Logistic regressions

For the real-time MRI data, we performed multi-level logistic regressions (with participant as random effect) using both glmer and MCMCglmm approaches, as well as “flat” (i.e., collapsing each participant to its most frequent value) using glm. For the sustained articulation (i.e., static) MRI data, we only performend “flat” logistic regression (using glm). In all cases, we used the three sets of anatomical variables (AN, CL and PC), the covariates (sex, group, phonexp and age), the properties of the native language (Retroflexes, R.trill, R.retroflex, R.flap.tap, R.approximant & R.uvular), and English proficiency as DVs; to these, for the “static” case, we also added the tongue shape PCs.

In general, the participants are extremely consistent in their behaviour, as shown by the histograms of their proportion of FALSE responses across trials and their very high intra-calss correlation coefficients. Nevertheless, we still performed multi-level regressions with participant as random effects, in parallel with the “flat” regression of the majority value per participant (i.e., a participant would be coded as TRUE iff (s)he had a majority of TRUE responses across the real-time MRI trials).

Plots

The distribution of answers for the Real-Time MRI data.

The distribution of answers for the static MRI data.

The distribution of answers for the Real-time MRI /for/ data.

Summary of logistic regressions

Below is the summary of all logistic regressions (the full logistic regression results are in Appendix I).

Summary of non-trivial logistic regression models (i.e., models where there is more than the intercept). IC = the appropriate information criterion: AIC applies to glm(er) and DIC to mcmc, and they refer to the whole model, not just individual IVs. GR = Gelman-Rubin criterion (or scale reduction factor.) for mcmc: the closer to 1, the better the convergence of parallel chains (< ≈1.1 shows good convergence).

DV	Set	Case	Method	IV	Estimate	p	GR	IC
bunched.vs.tipuprtr	AN	RT.flat	glm(er)	(Intercept)	-0.66	0.0317	NA	73
bunched.vs.tipuprtr	AN	RT.flat	glm(er)	Retroflexesyes	-2.67	0.0120	NA	73
bunched.vs.tipuprtr	AN	static	glm(er)	(Intercept)	-0.60	0.0234	NA	93
bunched.vs.tipuprtr	AN	static	glm(er)	jLowPC2	0.79	0.0058	NA	93
bunched.vs.tipuprtr	AN	static	mcmc	(Intercept)	-0.84	0.0098	1	62
bunched.vs.tipuprtr	AN	static	mcmc	tongue.PC1	1.96	0.0001	1	62
bunched.vs.tipuprtr	AN	static	mcmc	tongue.PC6	-0.85	0.0079	1	62
bunched.vs.tipuprtr	CL	RT.flat	glm(er)	(Intercept)	-0.84	0.0051	NA	77
bunched.vs.tipuprtr	CL	RT.flat	glm(er)	R.flap.tapyes	-2.25	0.0344	NA	77
bunched.vs.tipuprtr	CL	RT.flat	mcmc	(Intercept)	-1.43	0.0001	1	76
bunched.vs.tipuprtr	CL	RT.flat	mcmc	C44	0.76	0.0077	1	76
bunched.vs.tipuprtr	CL	static	glm(er)	(Intercept)	-0.82	0.0062	NA	84
bunched.vs.tipuprtr	CL	static	glm(er)	tongue.PC4	-1.38	0.0003	NA	84
bunched.vs.tipuprtr	CL	static	mcmc	(Intercept)	-0.71	0.0307	1	54
bunched.vs.tipuprtr	CL	static	mcmc	tongue.PC1	1.83	0.0001	1	54
bunched.vs.tipuprtr	CL	static	mcmc	tongue.PC4	-1.28	0.0001	1	54
bunched.vs.tipuprtr	CL	RT.for	mcmc	(Intercept)	0.23	0.4066	1	82
bunched.vs.tipuprtr	CL	RT.for	mcmc	C44	0.71	0.0090	1	82
bunched.vs.tipuprtr	PC	RT.flat	glm(er)	(Intercept)	-0.84	0.0051	NA	77
bunched.vs.tipuprtr	PC	RT.flat	glm(er)	R.flap.tapyes	-2.25	0.0344	NA	77
bunched.vs.tipuprtr	PC	RT.flat	mcmc	(Intercept)	-0.91	0.0013	1	76
bunched.vs.tipuprtr	PC	RT.flat	mcmc	R.flap.tapyes	-1.82	0.0098	1	76
bunched.vs.tipuprtr	PC	static	mcmc	(Intercept)	-0.84	0.0100	1	62
bunched.vs.tipuprtr	PC	static	mcmc	tongue.PC1	1.97	0.0001	1	62
bunched.vs.tipuprtr	PC	static	mcmc	tongue.PC6	-0.85	0.0092	1	62
bunched.vs.tipuprtr	PC	RT.for	mcmc	(Intercept)	0.34	0.2304	1	81
bunched.vs.tipuprtr	PC	RT.for	mcmc	PC11	0.75	0.0059	1	81
bunched.vs.others	AN	RT.flat	mcmc	(Intercept)	-0.91	0.0014	1	76
bunched.vs.others	AN	RT.flat	mcmc	R.flap.tapyes	-1.93	0.0064	1	76
bunched.vs.others	AN	static	glm(er)	(Intercept)	-0.70	0.0068	NA	97
bunched.vs.others	AN	static	glm(er)	jLowPC2	0.85	0.0037	NA	97
bunched.vs.others	AN	static	mcmc	(Intercept)	-1.77	0.0001	1	66
bunched.vs.others	AN	static	mcmc	tongue.PC1	1.76	0.0001	1	66
bunched.vs.others	AN	static	mcmc	tongue.PC6	-1.10	0.0008	1	66
bunched.vs.others	AN	static	mcmc	sexF	1.55	0.0086	1	66
bunched.vs.others	AN	RT.for	glm(er)	(Intercept)	0.17	0.5215	NA	84
bunched.vs.others	AN	RT.for	glm(er)	jTotPC2	0.88	0.0080	NA	84
bunched.vs.others	CL	RT.flat	glm(er)	(Intercept)	-0.84	0.0051	NA	77
bunched.vs.others	CL	RT.flat	glm(er)	R.flap.tapyes	-2.38	0.0251	NA	77
bunched.vs.others	CL	RT.flat	mcmc	(Intercept)	-1.47	0.0001	1	77
bunched.vs.others	CL	RT.flat	mcmc	C44	0.80	0.0058	1	77
bunched.vs.others	CL	static	glm(er)	(Intercept)	-0.70	0.0069	NA	96
bunched.vs.others	CL	static	glm(er)	C44	0.90	0.0033	NA	96
bunched.vs.others	CL	static	mcmc	(Intercept)	-0.83	0.0098	1	56
bunched.vs.others	CL	static	mcmc	tongue.PC1	1.74	0.0001	1	56
bunched.vs.others	CL	static	mcmc	tongue.PC4	-1.44	0.0001	1	56
bunched.vs.others	CL	RT.for	mcmc	(Intercept)	0.16	0.5641	1	85
bunched.vs.others	CL	RT.for	mcmc	C44	0.73	0.0079	1	85
bunched.vs.others	PC	RT.flat	glm(er)	(Intercept)	-0.84	0.0051	NA	77
bunched.vs.others	PC	RT.flat	glm(er)	R.flap.tapyes	-2.38	0.0251	NA	77
bunched.vs.others	PC	RT.flat	mcmc	(Intercept)	-0.91	0.0027	1	76
bunched.vs.others	PC	RT.flat	mcmc	R.flap.tapyes	-1.94	0.0041	1	76
bunched.vs.others	PC	static	glm(er)	(Intercept)	-1.35	0.0006	NA	99
bunched.vs.others	PC	static	glm(er)	R.uvularyes	1.35	0.0076	NA	99
bunched.vs.others	PC	static	mcmc	(Intercept)	-1.06	0.0006	1	72
bunched.vs.others	PC	static	mcmc	tongue.PC1	1.72	0.0001	1	72
bunched.vs.others	PC	static	mcmc	tongue.PC6	-0.91	0.0032	1	72
bunched.vs.others	PC	RT.for	mcmc	(Intercept)	0.27	0.3109	1	84
bunched.vs.others	PC	RT.for	mcmc	PC11	0.75	0.0054	1	84
tipuprtr.vs.others	AN	static	mcmc	(Intercept)	0.46	0.1126	1	72
tipuprtr.vs.others	AN	static	mcmc	tongue.PC1	-1.84	0.0001	1	72
tipuprtr.vs.others	CL	static	mcmc	(Intercept)	0.46	0.1057	1	72
tipuprtr.vs.others	CL	static	mcmc	tongue.PC1	-1.84	0.0001	1	72
tipuprtr.vs.others	PC	static	mcmc	(Intercept)	0.45	0.1093	1	72
tipuprtr.vs.others	PC	static	mcmc	tongue.PC1	-1.84	0.0001	1	72
lipround.vs.nonlipround	AN	RT.flat	mcmc	(Intercept)	0.56	0.0530	1	100
lipround.vs.nonlipround	AN	RT.flat	mcmc	R.flap.tapyes	-1.72	0.0007	1	100
lipround.vs.nonlipround	AN	static	mcmc	(Intercept)	-0.18	0.5200	1	99
lipround.vs.nonlipround	AN	static	mcmc	R.flap.tapyes	-1.43	0.0082	1	99
lipround.vs.nonlipround	CL	RT.flat	glm(er)	(Intercept)	0.58	0.0422	NA	101
lipround.vs.nonlipround	CL	RT.flat	glm(er)	R.flap.tapyes	-1.79	0.0011	NA	101
lipround.vs.nonlipround	CL	RT.flat	mcmc	(Intercept)	0.56	0.0541	1	100
lipround.vs.nonlipround	CL	RT.flat	mcmc	R.flap.tapyes	-1.72	0.0008	1	100
lipround.vs.nonlipround	CL	static	glm(er)	(Intercept)	-0.15	0.5866	NA	101
lipround.vs.nonlipround	CL	static	glm(er)	R.flap.tapyes	-1.51	0.0133	NA	101
lipround.vs.nonlipround	CL	static	mcmc	(Intercept)	-0.18	0.5152	1	99
lipround.vs.nonlipround	CL	static	mcmc	R.flap.tapyes	-1.43	0.0093	1	99
lipround.vs.nonlipround	CL	RT.multilevel	glm(er)	(Intercept)	10.40	0.0000	NA	182
lipround.vs.nonlipround	CL	RT.multilevel	glm(er)	R.flap.tapyes	-21.99	0.0000	NA	182
lipround.vs.nonlipround	PC	RT.flat	glm(er)	(Intercept)	0.58	0.0422	NA	101
lipround.vs.nonlipround	PC	RT.flat	glm(er)	R.flap.tapyes	-1.79	0.0011	NA	101
lipround.vs.nonlipround	PC	RT.flat	mcmc	(Intercept)	0.56	0.0511	1	100
lipround.vs.nonlipround	PC	RT.flat	mcmc	R.flap.tapyes	-1.72	0.0007	1	100
lipround.vs.nonlipround	PC	static	glm(er)	(Intercept)	-0.15	0.5866	NA	101
lipround.vs.nonlipround	PC	static	glm(er)	R.flap.tapyes	-1.51	0.0133	NA	101
lipround.vs.nonlipround	PC	static	mcmc	(Intercept)	-0.61	0.0136	1	99
lipround.vs.nonlipround	PC	static	mcmc	PC10	-0.67	0.0079	1	99
lipround.vs.nonlipround	PC	RT.multilevel	glm(er)	(Intercept)	-0.86	0.6709	NA	220
lipround.vs.nonlipround	PC	RT.multilevel	glm(er)	PC1	7.47	0.0012	NA	220
lipround.vs.nonlipround	PC	RT.multilevel	glm(er)	PC10	-7.97	0.0008	NA	220

Same logistic regressions in an even more condensed form

Significant predictors for each DV and case using glm(er). The sign in parantheses is the sign of the estimated beta.

dv	RT.multilevel	RT.flat	static	RT.for
bunched.vs.tipuprtr		Retroflexesyes (-), R.flap.tapyes (-)	jLowPC2 (+), tongue.PC4 (-)
bunched.vs.others		R.flap.tapyes (-)	jLowPC2 (+), C44 (+), R.uvularyes (+)	jTotPC2 (+)
tipuprtr.vs.others
lipround.vs.nonlipround	R.flap.tapyes (-), PC1 (+), PC10 (-)	R.flap.tapyes (-)	R.flap.tapyes (-)

Significant predictors for each DV and case using mcmc. The sign in parantheses is the sign of the estimated beta.

dv	RT.multilevel	RT.flat	static	RT.for
bunched.vs.tipuprtr		C44 (+), R.flap.tapyes (-)	tongue.PC1 (+), tongue.PC6 (-), tongue.PC4 (-)	C44 (+), PC11 (+)
bunched.vs.others		R.flap.tapyes (-), C44 (+)	tongue.PC1 (+), tongue.PC6 (-), sexF (+), tongue.PC4 (-)	C44 (+), PC11 (+)
tipuprtr.vs.others			tongue.PC1 (-)
lipround.vs.nonlipround		R.flap.tapyes (-)	R.flap.tapyes (-), PC10 (-)

Clustering of tongue shapes

Tongue shape is available only for Static MRI.

   clust
sex  1  2  3  4  5  6
  M 15  2 10  7  9  6
  F  3  4 14  2  4  3

Clusters by tPC1 x tPC2.

Clusters by tPC3 x tPC4.

Clusters by tPC5 x tPC6.

Figure 7. Tongue shape by cluster.

Tongue shapes by tongue PCs and discrete classifications

Tongue shape by tPC1 x tPC2.

Tongue shape by tPC3 x tPC4.

Tongue shape by tPC5 x tPC6.

Tongue shape within participants

We show, for each variable of interest:

• a matrix plot of each participant (x-axis) and segment in order (y-axis, from bottom to top) and for each segment the value it has (or NA if missing) – this allows a quick look at the consistency between and within participants;
• a second plot, where for a given participant we computed the number of changes in value between consecutive segments (NAs are ignored), resulting in an averged (across participants) observed number of changes (the red dotted line) and the histogram (with mean and sd) of 10,000 number of changes for permuted values; p is the proportion of permuted # of changes that are smaller than the observed one.

Plotting the raw rtype within participants

Figure 1. Matrix plot of rtype by participant. Participants are sorted in decreasing order by number of bunched, then retroflex, and then tip-up responses. Missing data is represented by white cells. Rounding is represented by small black circles. Participant IDs are followed by the participant’s sex (in parantheses). The primary participant’s language subfamily is represented by the bottom symbols. The top circles represent the participant’s English proficiency (circle size) and phonetic expertise (circle colour).

Per variable of interest (with permutations test)

Figure 9. Matrix plot of rtype2 by participant.

Figure 10. Observed vs permuted (10000 perms) number of changes for rtype2.

Figure 11. Matrix plot of poa2 by participant.

Figure 12. Observed vs permuted (10000 perms) number of changes for poa2.

Figure 13. Matrix plot of dorsum by participant.

Figure 14. Observed vs permuted (10000 perms) number of changes for dorsum

Figure 15. Matrix plot of lipround by participant.

Figure 16. Observed vs permuted (10000 perms) number of changes for lipround

Figure 17. Matrix plot of tip-up/retroflex vs bunched by participant.

Figure 18. Observed vs permuted (10000 perms) number of changes for tip-up/retroflex vs bunched

Appendix I: Full logistic regression results

Bunched vs. tip-up/retroflex

Coding: TRUE FALSE “bunched” “tip-up/retroflex”

ICC = 0.992

Real-time (multilevel)

AN

glmer

Generalized linear mixed model fit by maximum likelihood (Laplace
  Approximation) [glmerMod]
 Family: binomial  ( logit )
Formula: dv ~ (1 | ID)
   Data: data

     AIC      BIC   logLik deviance df.resid 
     150      158      -73      146      438 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-2.1949 -0.0047 -0.0047 -0.0047  2.2676 

Random effects:
 Groups Name        Variance Std.Dev.
 ID     (Intercept) 390      19.8    
Number of obs: 440, groups:  ID, 77

Fixed effects:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)   -10.66       1.36   -7.81  5.7e-15 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 83 

 G-structure:  ~ID

   post.mean l-95% CI u-95% CI eff.samp
ID      98.5     31.6      192     95.6

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - cWidth - pHeight - pm2Width - 
    pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - antPArea - ttAvailA - jTotPC2 - overbite - m2Width - R.retroflex - slMHeight - antPAreaR - C.M2w - pm2Len - upIA - pWid - R.trill - aRBulge - C.P2w - C.P2l - aRA - overjet - pRoofA - pTransA - phonexp - sex - m2Height - slMDepth - R.approximant - age - English.proficiency - lowIA - jLowPC2 - jLowPC3 - R.flap.tap 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)     -4.87    -7.05    -2.49     1335 <6e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

CL

glmer

Generalized linear mixed model fit by maximum likelihood (Laplace
  Approximation) [glmerMod]
 Family: binomial  ( logit )
Formula: dv ~ (1 | ID)
   Data: data

     AIC      BIC   logLik deviance df.resid 
     150      158      -73      146      438 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-2.1949 -0.0047 -0.0047 -0.0047  2.2676 

Random effects:
 Groups Name        Variance Std.Dev.
 ID     (Intercept) 390      19.8    
Number of obs: 440, groups:  ID, 77

Fixed effects:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)   -10.66       1.36   -7.81  5.7e-15 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 81 

 G-structure:  ~ID

   post.mean l-95% CI u-95% CI eff.samp
ID       118     29.1      255     50.9

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - C8 - C41 - C43 - C37 - C27 - R.trill - sex - C40 - phonexp - R.approximant - age - English.proficiency - C12 - R.flap.tap - C44 

            post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)     -4.93    -7.36    -2.60     1273 0.00033 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

PC

glmer

Generalized linear mixed model fit by maximum likelihood (Laplace
  Approximation) [glmerMod]
 Family: binomial  ( logit )
Formula: dv ~ (1 | ID)
   Data: data

     AIC      BIC   logLik deviance df.resid 
     150      158      -73      146      438 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-2.1949 -0.0047 -0.0047 -0.0047  2.2676 

Random effects:
 Groups Name        Variance Std.Dev.
 ID     (Intercept) 390      19.8    
Number of obs: 440, groups:  ID, 77

Fixed effects:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)   -10.66       1.36   -7.81  5.7e-15 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 82 

 G-structure:  ~ID

   post.mean l-95% CI u-95% CI eff.samp
ID       117     31.8      240     62.7

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - sex - PC6 - PC2 - R.trill - PC7 - PC9 - PC1 - phonexp - PC10 - PC3 - PC13 - R.approximant - PC4 - PC5 - age - English.proficiency - PC11 - PC12 - PC8 - R.flap.tap 

            post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)     -4.95    -7.36    -2.57     1136 0.00011 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1       1.01

Real-time (per-participant, ‘flat’)

AN

glmer

Call:
glm(formula = dv ~ Retroflexes, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-0.912  -0.912  -0.265  -0.265   2.595  

Coefficients:
               Estimate Std. Error z value Pr(>|z|)  
(Intercept)      -0.661      0.308   -2.15    0.032 *
Retroflexesyes   -2.671      1.063   -2.51    0.012 *
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 80.793  on 75  degrees of freedom
Residual deviance: 68.984  on 74  degrees of freedom
  (4 observations deleted due to missingness)
AIC: 72.98

Number of Fisher Scoring iterations: 6

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 82 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - cWidth - pHeight - pm2Width - 
    pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - antPArea - ttAvailA - jTotPC2 - overbite - m2Width - R.retroflex - slMHeight - antPAreaR - C.M2w - pm2Len - upIA - pWid - R.trill - pTransA - C.P2l - pRoofA - R.approximant - m2Height - sex - C.P2w - slMDepth - phonexp - overjet - aRBulge - aRA - jLowPC2 - jLowPC3 - age - R.flap.tap - lowIA - English.proficiency 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)     -1.47    -2.11    -0.87    15669 <6e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

CL

glmer

Call:
glm(formula = dv ~ R.flap.tap, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-0.848  -0.848  -0.298  -0.298   2.504  

Coefficients:
              Estimate Std. Error z value Pr(>|z|)   
(Intercept)     -0.838      0.299   -2.80   0.0051 **
R.flap.tapyes   -2.253      1.065   -2.11   0.0344 * 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 80.793  on 75  degrees of freedom
Residual deviance: 73.147  on 74  degrees of freedom
  (4 observations deleted due to missingness)
AIC: 77.15

Number of Fisher Scoring iterations: 5

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 76 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - C43 - sex - R.trill - C41 - R.approximant - C27 - C40 - C37 - C8 - phonexp - R.flap.tap - age - C12 - English.proficiency 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)    -1.654   -2.337   -0.955    13235 <6e-05 ***
C44             0.883    0.203    1.593    13070 0.0077 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
C44                  1          1

Multivariate psrf

1

PC

glmer

Call:
glm(formula = dv ~ R.flap.tap, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-0.848  -0.848  -0.298  -0.298   2.504  

Coefficients:
              Estimate Std. Error z value Pr(>|z|)   
(Intercept)     -0.838      0.299   -2.80   0.0051 **
R.flap.tapyes   -2.253      1.065   -2.11   0.0344 * 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 80.793  on 75  degrees of freedom
Residual deviance: 73.147  on 74  degrees of freedom
  (4 observations deleted due to missingness)
AIC: 77.15

Number of Fisher Scoring iterations: 5

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 76 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - sex - PC9 - PC6 - R.trill - PC2 - PC7 - R.approximant - PC10 - phonexp - PC5 - PC4 - PC13 - PC8 - PC1 - PC3 - PC12 - age - PC11 - English.proficiency 

              post.mean l-95% CI u-95% CI eff.samp  pMCMC   
(Intercept)      -1.050   -1.765   -0.388    16181 0.0013 **
R.flap.tapyes    -2.110   -3.886   -0.307     9372 0.0098 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

              Point est. Upper C.I.
(Intercept)            1          1
R.flap.tapyes          1          1

Multivariate psrf

1

Static (per participant)

AN

glmer

Call:
glm(formula = dv ~ jLowPC2, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-1.452  -0.957  -0.649   1.084   2.025  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)   
(Intercept)   -0.597      0.263   -2.27   0.0234 * 
jLowPC2        0.789      0.286    2.76   0.0058 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 98.163  on 73  degrees of freedom
Residual deviance: 89.048  on 72  degrees of freedom
  (5 observations deleted due to missingness)
AIC: 93.05

Number of Fisher Scoring iterations: 3

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 62 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency + tongue.PC1 + tongue.PC2 + 
    tongue.PC3 + tongue.PC4 + tongue.PC5 + tongue.PC6 - cWidth - pHeight - m2Width - pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - jTotPC2 - antPArea - ttAvailA - overbite - pWid - R.retroflex - slMHeight - C.P2w - antPAreaR - pm2Len - tongue.PC4 - upIA - pm2Width - aRA - phonexp - C.M2w - English.proficiency - pTransA - aRBulge - R.trill - m2Height - R.flap.tap - pRoofA - tongue.PC5 - R.approximant - slMDepth - tongue.PC3 - 
    C.P2l - overjet - lowIA - jLowPC3 - age - tongue.PC2 - jLowPC2 - sex 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)    -0.980   -1.763   -0.247     9142 0.0098 ** 
tongue.PC1      2.279    1.373    3.235     6739 <6e-05 ***
tongue.PC6     -0.989   -1.752   -0.194     8450 0.0079 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
tongue.PC1           1          1
tongue.PC6           1          1

Multivariate psrf

1

CL

glmer

Call:
glm(formula = dv ~ tongue.PC4, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-1.404  -0.865  -0.495   0.882   2.919  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)   -0.818      0.299   -2.74  0.00619 ** 
tongue.PC4    -1.384      0.386   -3.58  0.00034 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 98.163  on 73  degrees of freedom
Residual deviance: 79.791  on 72  degrees of freedom
  (5 observations deleted due to missingness)
AIC: 83.79

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 54 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency + tongue.PC1 + tongue.PC2 + tongue.PC3 + tongue.PC4 + tongue.PC5 + tongue.PC6 - Retroflexes - R.uvular - group - R.retroflex - sex - R.trill - C12 - C27 - C8 - phonexp - R.flap.tap - C41 - R.approximant - tongue.PC3 - English.proficiency - C43 - tongue.PC5 - C40 - C37 - tongue.PC2 - age - C44 - tongue.PC6 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)   -0.8210  -1.5925  -0.0864    12569  0.031 *  
tongue.PC1     2.1236   1.2423   3.0909     7823 <6e-05 ***
tongue.PC4    -1.4805  -2.3028  -0.6593     9720 <6e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
tongue.PC1           1          1
tongue.PC4           1          1

Multivariate psrf

1

PC

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-0.975  -0.975  -0.975   1.394   1.394  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)  
(Intercept)   -0.496      0.240   -2.07    0.038 *
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 98.163  on 73  degrees of freedom
Residual deviance: 98.163  on 73  degrees of freedom
  (5 observations deleted due to missingness)
AIC: 100.2

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 62 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency + tongue.PC1 + tongue.PC2 + tongue.PC3 + tongue.PC4 + tongue.PC5 + tongue.PC6 - Retroflexes - R.uvular - group - R.retroflex - sex - tongue.PC4 - R.flap.tap - phonexp - PC7 - PC9 - PC5 - R.trill - R.approximant - tongue.PC5 - English.proficiency - PC2 - tongue.PC3 - PC8 - PC11 - PC10 - 
    PC6 - PC3 - age - PC4 - PC12 - tongue.PC2 - PC1 - PC13 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)    -0.977   -1.761   -0.209     9761 0.0100 ** 
tongue.PC1      2.282    1.350    3.219     7787 <6e-05 ***
tongue.PC6     -0.986   -1.754   -0.198     8768 0.0092 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
tongue.PC1           1          1
tongue.PC6           1          1

Multivariate psrf

1

Real-time /for/ context (per participant)

AN

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.30   -1.30    1.06    1.06    1.06  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)    0.288      0.255    1.13     0.26

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 86.046  on 62  degrees of freedom
Residual deviance: 86.046  on 62  degrees of freedom
  (17 observations deleted due to missingness)
AIC: 88.05

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 87 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - cWidth - pHeight - pm2Width - 
    pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - antPArea - ttAvailA - jTotPC2 - overbite - m2Width - R.retroflex - slMHeight - antPAreaR - C.M2w - pm2Len - upIA - pWid - pTransA - pRoofA - jLowPC2 - aRBulge - R.flap.tap - C.P2w - age - R.approximant - jLowPC3 - m2Height - C.P2l - R.trill - English.proficiency - sex - phonexp - slMDepth - overjet - lowIA - aRA 

            post.mean l-95% CI u-95% CI eff.samp pMCMC
(Intercept)     0.339   -0.266    0.916    16785  0.27

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

CL

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.30   -1.30    1.06    1.06    1.06  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)    0.288      0.255    1.13     0.26

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 86.046  on 62  degrees of freedom
Residual deviance: 86.046  on 62  degrees of freedom
  (17 observations deleted due to missingness)
AIC: 88.05

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 82 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - age - R.approximant - C8 - R.flap.tap - sex - C27 - C41 - C43 - C40 - R.trill - phonexp - English.proficiency - C12 - C37 

            post.mean l-95% CI u-95% CI eff.samp pMCMC   
(Intercept)     0.266   -0.360    0.902    16281 0.407   
C44             0.821    0.158    1.465    15127 0.009 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
C44                  1          1

Multivariate psrf

1

PC

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.30   -1.30    1.06    1.06    1.06  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)    0.288      0.255    1.13     0.26

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 86.046  on 62  degrees of freedom
Residual deviance: 86.046  on 62  degrees of freedom
  (17 observations deleted due to missingness)
AIC: 88.05

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 81 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - sex - PC10 - PC12 - R.flap.tap - PC9 - age - R.approximant - PC7 - PC5 - R.trill - PC13 - PC6 - PC2 - PC1 - PC3 - PC8 - PC4 - English.proficiency - phonexp 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC   
(Intercept)     0.392   -0.253    1.028    15993 0.2304   
PC11            0.871    0.232    1.551    14935 0.0059 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
PC11                 1          1

Multivariate psrf

1

Bunched vs. everything else

Coding: TRUE FALSE “bunched” “other”

ICC = 0.992

Real-time (multilevel)

AN

glmer

Length  Class   Mode 
     0   NULL   NULL 

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 81 

 G-structure:  ~ID

   post.mean l-95% CI u-95% CI eff.samp
ID       131     28.9      302     31.9

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - cWidth - pHeight - pm2Width - 
    pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - antPArea - ttAvailA - jTotPC2 - overbite - m2Width - R.retroflex - slMHeight - antPAreaR - C.M2w - pm2Len - upIA - pWid - R.trill - aRBulge - C.P2w - C.P2l - pTransA - aRA - pRoofA - overjet - phonexp - sex - m2Height - slMDepth - R.approximant - age - English.proficiency - jLowPC2 - lowIA - jLowPC3 - R.flap.tap 

            post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)     -5.21    -7.70    -2.78     1192 0.00011 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1       1.01

CL

glmer

Generalized linear mixed model fit by maximum likelihood (Laplace
  Approximation) [glmerMod]
 Family: binomial  ( logit )
Formula: dv ~ (1 | ID)
   Data: data

     AIC      BIC   logLik deviance df.resid 
     151      159      -74      147      467 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-2.1953 -0.0044 -0.0044 -0.0044  2.2674 

Random effects:
 Groups Name        Variance Std.Dev.
 ID     (Intercept) 399      20      
Number of obs: 469, groups:  ID, 79

Fixed effects:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)   -10.78       1.36   -7.93  2.2e-15 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 83 

 G-structure:  ~ID

   post.mean l-95% CI u-95% CI eff.samp
ID       109     32.2      217     80.1

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - C41 - C8 - C43 - C27 - R.trill - C37 - sex - C40 - phonexp - R.approximant - age - English.proficiency - C12 - R.flap.tap - C44 

            post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)     -5.14    -7.51    -2.85     1336 0.00011 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1       1.01

PC

glmer

Generalized linear mixed model fit by maximum likelihood (Laplace
  Approximation) [glmerMod]
 Family: binomial  ( logit )
Formula: dv ~ (1 | ID)
   Data: data

     AIC      BIC   logLik deviance df.resid 
     151      159      -74      147      467 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-2.1953 -0.0044 -0.0044 -0.0044  2.2674 

Random effects:
 Groups Name        Variance Std.Dev.
 ID     (Intercept) 399      20      
Number of obs: 469, groups:  ID, 79

Fixed effects:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)   -10.78       1.36   -7.93  2.2e-15 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 83 

 G-structure:  ~ID

   post.mean l-95% CI u-95% CI eff.samp
ID       116     29.6      253     24.9

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - sex - PC2 - PC6 - R.trill - PC9 - PC7 - PC1 - phonexp - PC10 - PC5 - PC3 - age - PC4 - R.approximant - English.proficiency - PC13 - PC11 - PC12 - PC8 - R.flap.tap 

            post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)     -5.09    -7.38    -2.71     1329 0.00011 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

Real-time (per-participant, ‘flat’)

AN

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-0.696  -0.696  -0.696  -0.696   1.753  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)   -1.294      0.274   -4.73  2.3e-06 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 82.279  on 78  degrees of freedom
Residual deviance: 82.279  on 78  degrees of freedom
  (1 observation deleted due to missingness)
AIC: 84.28

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 76 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - cWidth - pHeight - pm2Width - 
    pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - antPArea - ttAvailA - jTotPC2 - overbite - m2Width - R.retroflex - slMHeight - antPAreaR - C.M2w - pm2Len - upIA - pWid - R.trill - pTransA - C.P2l - pRoofA - R.approximant - m2Height - sex - C.P2w - slMDepth - phonexp - overjet - aRBulge - aRA - jLowPC2 - jLowPC3 - age - lowIA - English.proficiency 

              post.mean l-95% CI u-95% CI eff.samp  pMCMC   
(Intercept)      -1.057   -1.748   -0.371    16494 0.0014 **
R.flap.tapyes    -2.234   -4.048   -0.527     8507 0.0064 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

              Point est. Upper C.I.
(Intercept)            1          1
R.flap.tapyes          1          1

Multivariate psrf

1

CL

glmer

Call:
glm(formula = dv ~ R.flap.tap, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-0.848  -0.848  -0.280  -0.280   2.553  

Coefficients:
              Estimate Std. Error z value Pr(>|z|)   
(Intercept)     -0.838      0.299   -2.80   0.0051 **
R.flap.tapyes   -2.381      1.062   -2.24   0.0251 * 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 82.279  on 78  degrees of freedom
Residual deviance: 73.397  on 77  degrees of freedom
  (1 observation deleted due to missingness)
AIC: 77.4

Number of Fisher Scoring iterations: 5

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 77 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - C43 - sex - C41 - R.trill - R.approximant - C27 - C40 - C37 - C8 - phonexp - R.flap.tap - age - C12 - English.proficiency 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)    -1.701   -2.381   -1.017    13366 <6e-05 ***
C44             0.925    0.242    1.634    12183 0.0058 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
C44                  1          1

Multivariate psrf

1

PC

glmer

Call:
glm(formula = dv ~ R.flap.tap, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-0.848  -0.848  -0.280  -0.280   2.553  

Coefficients:
              Estimate Std. Error z value Pr(>|z|)   
(Intercept)     -0.838      0.299   -2.80   0.0051 **
R.flap.tapyes   -2.381      1.062   -2.24   0.0251 * 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 82.279  on 78  degrees of freedom
Residual deviance: 73.397  on 77  degrees of freedom
  (1 observation deleted due to missingness)
AIC: 77.4

Number of Fisher Scoring iterations: 5

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 76 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - sex - PC9 - PC6 - R.trill - PC2 - PC7 - R.approximant - PC10 - phonexp - PC5 - PC4 - PC13 - PC8 - PC1 - PC3 - PC12 - age - PC11 - English.proficiency 

              post.mean l-95% CI u-95% CI eff.samp  pMCMC   
(Intercept)      -1.057   -1.735   -0.369    15218 0.0027 **
R.flap.tapyes    -2.250   -4.052   -0.523     8653 0.0041 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

              Point est. Upper C.I.
(Intercept)            1          1
R.flap.tapyes          1          1

Multivariate psrf

1

Static (per participant)

AN

glmer

Call:
glm(formula = dv ~ jLowPC2, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-1.444  -0.901  -0.636   1.098   2.112  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)   
(Intercept)   -0.704      0.260    -2.7   0.0068 **
jLowPC2        0.847      0.292     2.9   0.0037 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 102.723  on 78  degrees of freedom
Residual deviance:  92.553  on 77  degrees of freedom
AIC: 96.55

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 66 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency + tongue.PC1 + tongue.PC2 + 
    tongue.PC3 + tongue.PC4 + tongue.PC5 + tongue.PC6 - cWidth - pHeight - m2Width - pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - jTotPC2 - antPArea - ttAvailA - overbite - pWid - R.retroflex - slMHeight - C.P2w - antPAreaR - pm2Len - tongue.PC4 - upIA - pm2Width - aRA - m2Height - English.proficiency - aRBulge - R.approximant - phonexp - pRoofA - R.trill - pTransA - C.P2l - overjet - C.M2w - slMDepth - R.flap.tap - tongue.PC3 - 
    tongue.PC5 - jLowPC3 - age - lowIA - tongue.PC2 - jLowPC2 

            post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)    -2.054   -3.096   -1.028     7881 < 6e-05 ***
sexF            1.802    0.428    3.140    12413 0.00856 ** 
tongue.PC1      2.047    1.168    2.984     7575 < 6e-05 ***
tongue.PC6     -1.272   -2.107   -0.480     7689 0.00078 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
sexF                 1          1
tongue.PC1           1          1
tongue.PC6           1          1

Multivariate psrf

1

CL

glmer

Call:
glm(formula = dv ~ C44, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-1.877  -0.855  -0.630   1.196   2.169  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)   
(Intercept)   -0.703      0.260   -2.70   0.0069 **
C44            0.905      0.308    2.94   0.0033 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 102.723  on 78  degrees of freedom
Residual deviance:  91.656  on 77  degrees of freedom
AIC: 95.66

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 56 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency + tongue.PC1 + tongue.PC2 + tongue.PC3 + tongue.PC4 + tongue.PC5 + tongue.PC6 - Retroflexes - R.uvular - group - R.retroflex - sex - C27 - R.trill - C8 - C41 - R.approximant - C12 - phonexp - English.proficiency - R.flap.tap - C37 - C43 - tongue.PC3 - C40 - tongue.PC5 - tongue.PC2 - age - C44 - tongue.PC6 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)    -0.958   -1.718   -0.245    11771 0.0098 ** 
tongue.PC1      2.017    1.140    2.948     8747 <6e-05 ***
tongue.PC4     -1.675   -2.453   -0.893    10526 <6e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
tongue.PC1           1          1
tongue.PC4           1          1

Multivariate psrf

1

PC

glmer

Call:
glm(formula = dv ~ R.uvular, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-1.177  -0.927  -0.678   1.177   1.780  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)   -1.355      0.397   -3.42  0.00064 ***
R.uvularyes    1.355      0.507    2.67  0.00757 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 102.723  on 78  degrees of freedom
Residual deviance:  95.031  on 77  degrees of freedom
AIC: 99.03

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 72 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency + tongue.PC1 + tongue.PC2 + tongue.PC3 + tongue.PC4 + tongue.PC5 + tongue.PC6 - Retroflexes - R.uvular - group - R.retroflex - sex - tongue.PC4 - PC9 - R.approximant - PC5 - phonexp - R.trill - PC7 - English.proficiency - PC6 - PC10 - tongue.PC5 - R.flap.tap - PC2 - tongue.PC3 - PC11 - 
    PC4 - PC13 - PC3 - PC8 - age - tongue.PC2 - PC12 - PC1 

            post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)    -1.235   -2.011   -0.533     9511 0.00056 ***
tongue.PC1      1.999    1.183    2.884     7995 < 6e-05 ***
tongue.PC6     -1.061   -1.821   -0.318     9098 0.00322 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
tongue.PC1           1          1
tongue.PC6           1          1

Multivariate psrf

1

Real-time /for/ context (per participant)

AN

glmer

Call:
glm(formula = dv ~ jTotPC2, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-2.062  -1.056   0.639   0.952   1.430  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)   
(Intercept)    0.171      0.267    0.64    0.522   
jTotPC2        0.880      0.332    2.65    0.008 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 89.354  on 64  degrees of freedom
Residual deviance: 80.355  on 63  degrees of freedom
  (15 observations deleted due to missingness)
AIC: 84.35

Number of Fisher Scoring iterations: 3

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 91 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - cWidth - pHeight - pm2Width - 
    pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - antPArea - ttAvailA - jTotPC2 - overbite - m2Width - R.retroflex - slMHeight - antPAreaR - C.M2w - pm2Len - upIA - pWid - jLowPC2 - R.flap.tap - pRoofA - pTransA - aRBulge - C.P2w - age - R.approximant - jLowPC3 - m2Height - C.P2l - R.trill - English.proficiency - phonexp - slMDepth - sex - overjet - lowIA - aRA 

            post.mean l-95% CI u-95% CI eff.samp pMCMC
(Intercept)     0.260   -0.336    0.841    16153  0.39

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

CL

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.27   -1.27    1.09    1.09    1.09  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)    0.216      0.250    0.87     0.39

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 89.354  on 64  degrees of freedom
Residual deviance: 89.354  on 64  degrees of freedom
  (15 observations deleted due to missingness)
AIC: 91.35

Number of Fisher Scoring iterations: 3

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 85 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - age - R.approximant - C8 - R.flap.tap - C27 - sex - C41 - C43 - C40 - R.trill - phonexp - English.proficiency - C12 - C37 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC   
(Intercept)     0.185   -0.438    0.801    16484 0.5641   
C44             0.851    0.195    1.522    14428 0.0079 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
C44                  1          1

Multivariate psrf

1

PC

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.27   -1.27    1.09    1.09    1.09  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)    0.216      0.250    0.87     0.39

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 89.354  on 64  degrees of freedom
Residual deviance: 89.354  on 64  degrees of freedom
  (15 observations deleted due to missingness)
AIC: 91.35

Number of Fisher Scoring iterations: 3

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 84 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - sex - PC10 - PC9 - PC12 - age - R.flap.tap - R.approximant - PC7 - PC5 - R.trill - PC13 - PC3 - PC8 - PC6 - PC2 - PC1 - PC4 - English.proficiency - phonexp 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC   
(Intercept)     0.316   -0.306    0.943    15210 0.3109   
PC11            0.869    0.211    1.512    14262 0.0054 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
PC11                 1          1

Multivariate psrf

1

Tip-up/retroflex vs. everything else

Coding: TRUE FALSE “tip-up/retroflex” “other”

ICC = 0.982

Real-time (multilevel)

AN

glmer

Generalized linear mixed model fit by maximum likelihood (Laplace
  Approximation) [glmerMod]
 Family: binomial  ( logit )
Formula: dv ~ (1 | ID)
   Data: data

     AIC      BIC   logLik deviance df.resid 
     261      270     -129      257      467 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-2.2915 -0.1118  0.0115  0.0115  2.1611 

Random effects:
 Groups Name        Variance Std.Dev.
 ID     (Intercept) 178      13.3    
Number of obs: 469, groups:  ID, 79

Fixed effects:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)     8.79       1.17     7.5  6.6e-14 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 159 

 G-structure:  ~ID

   post.mean l-95% CI u-95% CI eff.samp
ID        67     21.9      126      135

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - cWidth - pHeight - pm2Width - 
    pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - antPArea - ttAvailA - jTotPC2 - overbite - m2Width - R.retroflex - slMHeight - antPAreaR - C.M2w - pm2Len - upIA - pWid - pRoofA - R.trill - R.flap.tap - C.P2l - C.P2w - overjet - aRA - sex - pTransA - m2Height - aRBulge - phonexp - R.approximant - jLowPC2 - jLowPC3 - lowIA - slMDepth - age - English.proficiency 

            post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)      3.42     1.55     5.44     1622 0.00067 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

CL

glmer

Length  Class   Mode 
     0   NULL   NULL 

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 158 

 G-structure:  ~ID

   post.mean l-95% CI u-95% CI eff.samp
ID      68.9     24.9      131      126

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - C37 - C27 - R.trill - R.flap.tap - C41 - sex - phonexp - C40 - C43 - C8 - R.approximant - C12 - age - C44 - English.proficiency 

            post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)      3.46     1.48     5.40     1225 0.00056 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

PC

glmer

Generalized linear mixed model fit by maximum likelihood (Laplace
  Approximation) [glmerMod]
 Family: binomial  ( logit )
Formula: dv ~ (1 | ID)
   Data: data

     AIC      BIC   logLik deviance df.resid 
     261      270     -129      257      467 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-2.2915 -0.1118  0.0115  0.0115  2.1611 

Random effects:
 Groups Name        Variance Std.Dev.
 ID     (Intercept) 178      13.3    
Number of obs: 469, groups:  ID, 79

Fixed effects:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)     8.79       1.17     7.5  6.6e-14 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 160 

 G-structure:  ~ID

   post.mean l-95% CI u-95% CI eff.samp
ID      64.3     21.8      123      137

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - sex - PC6 - R.trill - PC7 - R.flap.tap - PC2 - PC1 - PC10 - PC9 - phonexp - R.approximant - PC3 - PC13 - PC8 - PC12 - PC4 - PC5 - PC11 - age - English.proficiency 

            post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)      3.39     1.45     5.30     1153 0.00033 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

Real-time (per-participant, ‘flat’)

AN

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-1.657  -0.447   0.764   0.764   0.764  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)    1.082      0.259    4.18  2.9e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 89.394  on 78  degrees of freedom
Residual deviance: 89.394  on 78  degrees of freedom
  (1 observation deleted due to missingness)
AIC: 91.39

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 91 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - cWidth - pHeight - pm2Width - 
    pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - antPArea - ttAvailA - jTotPC2 - overbite - m2Width - R.retroflex - slMHeight - antPAreaR - C.M2w - pm2Len - upIA - pWid - R.approximant - R.flap.tap - pRoofA - C.P2l - m2Height - R.trill - sex - pTransA - aRA - C.P2w - jLowPC3 - jLowPC2 - aRBulge - overjet - phonexp - slMDepth - age - lowIA - English.proficiency 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)     1.278    0.698    1.891    15564 <6e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

CL

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-1.657  -0.447   0.764   0.764   0.764  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)    1.082      0.259    4.18  2.9e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 89.394  on 78  degrees of freedom
Residual deviance: 89.394  on 78  degrees of freedom
  (1 observation deleted due to missingness)
AIC: 91.39

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 91 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - C37 - C27 - R.trill - C41 - R.flap.tap - sex - R.approximant - C8 - C40 - phonexp - C43 - C12 - age - C44 - English.proficiency 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)     1.281    0.679    1.864    15775 <6e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

PC

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-1.657  -0.447   0.764   0.764   0.764  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)    1.082      0.259    4.18  2.9e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 89.394  on 78  degrees of freedom
Residual deviance: 89.394  on 78  degrees of freedom
  (1 observation deleted due to missingness)
AIC: 91.39

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 91 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - sex - R.flap.tap - PC10 - PC9 - PC2 - R.trill - R.approximant - PC6 - PC7 - PC3 - PC13 - PC8 - PC4 - phonexp - PC12 - PC5 - PC11 - PC1 - age - English.proficiency 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)     1.285    0.717    1.889    15007 <6e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

Static (per participant)

AN

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.32   -1.32    1.04    1.04    1.04  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)    0.332      0.228    1.46     0.15

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 107.37  on 78  degrees of freedom
Residual deviance: 107.37  on 78  degrees of freedom
AIC: 109.4

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 72 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency + tongue.PC1 + tongue.PC2 + 
    tongue.PC3 + tongue.PC4 + tongue.PC5 + tongue.PC6 - cWidth - pHeight - m2Width - pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - jTotPC2 - antPArea - ttAvailA - overbite - pWid - R.retroflex - slMHeight - C.P2w - antPAreaR - pm2Len - tongue.PC4 - upIA - pm2Width - aRA - phonexp - C.M2w - tongue.PC3 - R.trill - tongue.PC5 - pTransA - English.proficiency - slMDepth - R.approximant - m2Height - pRoofA - C.P2l - overjet - aRBulge - 
    lowIA - R.flap.tap - jLowPC3 - jLowPC2 - tongue.PC2 - sex - age - tongue.PC6 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)     0.529   -0.133    1.163    14413   0.11    
tongue.PC1     -2.140   -2.995   -1.267     9808 <6e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
tongue.PC1           1          1

Multivariate psrf

1

CL

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.32   -1.32    1.04    1.04    1.04  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)    0.332      0.228    1.46     0.15

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 107.37  on 78  degrees of freedom
Residual deviance: 107.37  on 78  degrees of freedom
AIC: 109.4

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 72 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency + tongue.PC1 + tongue.PC2 + tongue.PC3 + tongue.PC4 + tongue.PC5 + tongue.PC6 - Retroflexes - R.uvular - group - R.retroflex - sex - C12 - R.trill - phonexp - C8 - C43 - tongue.PC5 - C27 - tongue.PC3 - R.approximant - C41 - English.proficiency - R.flap.tap - C40 - tongue.PC2 - C37 - tongue.PC4 - tongue.PC6 - age - C44 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)     0.535   -0.133    1.176    14359   0.11    
tongue.PC1     -2.136   -2.994   -1.291    10508 <6e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
tongue.PC1           1          1

Multivariate psrf

1

PC

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.32   -1.32    1.04    1.04    1.04  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)    0.332      0.228    1.46     0.15

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 107.37  on 78  degrees of freedom
Residual deviance: 107.37  on 78  degrees of freedom
AIC: 109.4

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 72 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency + tongue.PC1 + tongue.PC2 + tongue.PC3 + tongue.PC4 + tongue.PC5 + tongue.PC6 - Retroflexes - R.uvular - group - R.retroflex - sex - tongue.PC4 - PC9 - phonexp - tongue.PC3 - PC2 - PC7 - R.trill - PC8 - tongue.PC5 - PC5 - R.approximant - PC11 - English.proficiency - R.flap.tap - PC12 - 
    PC10 - PC6 - PC13 - PC4 - tongue.PC2 - PC3 - PC1 - age - tongue.PC6 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC    
(Intercept)     0.528   -0.126    1.180    12517   0.11    
tongue.PC1     -2.138   -3.043   -1.298     9800 <6e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
tongue.PC1           1          1

Multivariate psrf

1

Real-time /for/ context (per participant)

AN

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.04   -1.04   -1.04    1.33    1.33  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)   -0.342      0.252   -1.36     0.17

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 88.239  on 64  degrees of freedom
Residual deviance: 88.239  on 64  degrees of freedom
  (15 observations deleted due to missingness)
AIC: 90.24

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 90 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - cWidth - pHeight - pm2Width - 
    pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - antPArea - ttAvailA - jTotPC2 - overbite - m2Width - R.retroflex - slMHeight - antPAreaR - C.M2w - pm2Len - upIA - pWid - jLowPC2 - C.P2w - R.flap.tap - aRBulge - R.approximant - pRoofA - age - pTransA - jLowPC3 - m2Height - C.P2l - sex - English.proficiency - R.trill - phonexp - overjet - slMDepth - aRA - lowIA 

            post.mean l-95% CI u-95% CI eff.samp pMCMC
(Intercept)    -0.406   -0.991    0.187    16462  0.18

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

CL

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.04   -1.04   -1.04    1.33    1.33  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)   -0.342      0.252   -1.36     0.17

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 88.239  on 64  degrees of freedom
Residual deviance: 88.239  on 64  degrees of freedom
  (15 observations deleted due to missingness)
AIC: 90.24

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 90 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - R.approximant - age - R.flap.tap - sex - C27 - C8 - C41 - C43 - C40 - phonexp - R.trill - English.proficiency - C12 - C37 - C44 

            post.mean l-95% CI u-95% CI eff.samp pMCMC
(Intercept)    -0.407   -1.011    0.166    16566  0.17

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

PC

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.04   -1.04   -1.04    1.33    1.33  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)   -0.342      0.252   -1.36     0.17

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 88.239  on 64  degrees of freedom
Residual deviance: 88.239  on 64  degrees of freedom
  (15 observations deleted due to missingness)
AIC: 90.24

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 90 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - sex - age - PC12 - R.approximant - R.flap.tap - PC10 - PC9 - PC7 - PC6 - PC5 - PC2 - PC1 - PC8 - PC3 - PC4 - R.trill - PC13 - English.proficiency - phonexp - PC11 

            post.mean l-95% CI u-95% CI eff.samp pMCMC
(Intercept)    -0.407   -1.008    0.170    16332  0.17

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

Lip rounding

Coding: TRUE FALSE “lip rounding” “no lip rounding”

ICC = 0.995

Real-time (multilevel)

AN

glmer

Length  Class   Mode 
     0   NULL   NULL 

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 77 

 G-structure:  ~ID

   post.mean l-95% CI u-95% CI eff.samp
ID      1170     86.2     2905     3.46

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - cWidth - pHeight - pm2Width - 
    pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - antPArea - ttAvailA - jTotPC2 - overbite - m2Width - R.retroflex - slMHeight - antPAreaR - C.M2w - pm2Len - upIA - pWid - C.P2w - aRBulge - overjet - m2Height - jLowPC2 - slMDepth - pTransA - age - sex - R.trill - phonexp - C.P2l - R.approximant - lowIA - pRoofA - aRA - English.proficiency - jLowPC3 - R.flap.tap 

            post.mean l-95% CI u-95% CI eff.samp pMCMC
(Intercept)     -0.31    -3.79     3.28     1236  0.86

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

CL

glmer

Generalized linear mixed model fit by maximum likelihood (Laplace
  Approximation) [glmerMod]
 Family: binomial  ( logit )
Formula: dv ~ R.flap.tap + (1 | ID)
   Data: data

     AIC      BIC   logLik deviance df.resid 
     182      194      -88      176      466 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-2.2598 -0.0030 -0.0030  0.0053  1.4026 

Random effects:
 Groups Name        Variance Std.Dev.
 ID     (Intercept) 501      22.4    
Number of obs: 469, groups:  ID, 79

Fixed effects:
              Estimate Std. Error z value Pr(>|z|)    
(Intercept)      10.40       1.42    7.34  2.1e-13 ***
R.flap.tapyes   -21.99       2.95   -7.46  8.6e-14 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Correlation of Fixed Effects:
            (Intr)
R.flap.tpys -0.688

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 81 

 G-structure:  ~ID

   post.mean l-95% CI u-95% CI eff.samp
ID       459      115      894     16.7

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - C40 - C41 - age - C8 - sex - C43 - phonexp - R.trill - R.approximant - C37 - C12 - C27 - English.proficiency - C44 - R.flap.tap 

            post.mean l-95% CI u-95% CI eff.samp pMCMC
(Intercept)    -0.286   -3.617    2.883     1381  0.86

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

PC

glmer

Generalized linear mixed model fit by maximum likelihood (Laplace
  Approximation) [glmerMod]
 Family: binomial  ( logit )
Formula: dv ~ PC1 + PC10 + (1 | ID)
   Data: data

     AIC      BIC   logLik deviance df.resid 
     220      237     -106      212      465 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-2.2633 -0.0292  0.0000  0.0461  1.4171 

Random effects:
 Groups Name        Variance Std.Dev.
 ID     (Intercept) 466      21.6    
Number of obs: 469, groups:  ID, 79

Fixed effects:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)   -0.862      2.028   -0.42  0.67087    
PC1            7.469      2.314    3.23  0.00125 ** 
PC10          -7.967      2.379   -3.35  0.00081 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Correlation of Fixed Effects:
     (Intr) PC1   
PC1  -0.501       
PC10  0.441 -0.532

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 82 

 G-structure:  ~ID

   post.mean l-95% CI u-95% CI eff.samp
ID       436     87.4      867     17.9

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - sex - PC2 - PC4 - PC5 - PC7 - age - PC6 - PC9 - R.trill - phonexp - R.approximant - PC13 - PC12 - PC8 - PC1 - PC11 - English.proficiency - PC10 - PC3 - R.flap.tap 

            post.mean l-95% CI u-95% CI eff.samp pMCMC
(Intercept)    -0.349   -3.495    2.813     1762  0.83

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1       1.01

Real-time (per-participant, ‘flat’)

AN

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.19   -1.19    1.17    1.17    1.17  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)   0.0253     0.2250    0.11     0.91

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 109.5  on 78  degrees of freedom
Residual deviance: 109.5  on 78  degrees of freedom
  (1 observation deleted due to missingness)
AIC: 111.5

Number of Fisher Scoring iterations: 3

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 100 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - cWidth - pHeight - pm2Width - 
    pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - antPArea - ttAvailA - jTotPC2 - overbite - m2Width - R.retroflex - slMHeight - antPAreaR - C.M2w - pm2Len - upIA - pWid - C.P2w - sex - phonexp - lowIA - C.P2l - m2Height - age - overjet - jLowPC2 - aRBulge - aRA - slMDepth - English.proficiency - R.trill - R.approximant - pRoofA - pTransA - jLowPC3 

              post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)     0.65279 -0.00588  1.31632    16557 0.05300 .  
R.flap.tapyes  -1.99240 -3.19829 -0.77751    16281 0.00067 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

              Point est. Upper C.I.
(Intercept)            1          1
R.flap.tapyes          1          1

Multivariate psrf

1

CL

glmer

Call:
glm(formula = dv ~ R.flap.tap, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-1.432  -0.724   0.942   0.942   1.712  

Coefficients:
              Estimate Std. Error z value Pr(>|z|)   
(Intercept)      0.582      0.286    2.03   0.0422 * 
R.flap.tapyes   -1.786      0.547   -3.27   0.0011 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 109.50  on 78  degrees of freedom
Residual deviance:  97.26  on 77  degrees of freedom
  (1 observation deleted due to missingness)
AIC: 101.3

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 100 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - phonexp - C41 - age - C40 - sex - C8 - C12 - C43 - R.approximant - English.proficiency - C27 - R.trill - C37 - C44 

              post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)      0.6497  -0.0135   1.3234    16635 0.05411 .  
R.flap.tapyes   -1.9976  -3.2202  -0.8409    15548 0.00078 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

              Point est. Upper C.I.
(Intercept)            1          1
R.flap.tapyes          1          1

Multivariate psrf

1

PC

glmer

Call:
glm(formula = dv ~ R.flap.tap, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-1.432  -0.724   0.942   0.942   1.712  

Coefficients:
              Estimate Std. Error z value Pr(>|z|)   
(Intercept)      0.582      0.286    2.03   0.0422 * 
R.flap.tapyes   -1.786      0.547   -3.27   0.0011 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 109.50  on 78  degrees of freedom
Residual deviance:  97.26  on 77  degrees of freedom
  (1 observation deleted due to missingness)
AIC: 101.3

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 100 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - sex - PC4 - age - PC2 - phonexp - PC7 - PC8 - PC6 - PC10 - PC11 - PC5 - PC1 - R.approximant - PC12 - PC13 - R.trill - English.proficiency - PC9 - PC3 

              post.mean l-95% CI u-95% CI eff.samp   pMCMC    
(Intercept)      0.6546  -0.0111   1.3047    17042 0.05111 .  
R.flap.tapyes   -1.9982  -3.2173  -0.8280    16320 0.00067 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

              Point est. Upper C.I.
(Intercept)            1          1
R.flap.tapyes          1          1

Multivariate psrf

1

Static (per participant)

AN

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-0.957  -0.957  -0.957   1.416   1.416  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)  
(Intercept)   -0.545      0.233   -2.33     0.02 *
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 103.87  on 78  degrees of freedom
Residual deviance: 103.87  on 78  degrees of freedom
AIC: 105.9

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 99 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency + tongue.PC1 + tongue.PC2 + 
    tongue.PC3 + tongue.PC4 + tongue.PC5 + tongue.PC6 - cWidth - pHeight - m2Width - pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - jTotPC2 - antPArea - ttAvailA - overbite - pWid - R.retroflex - slMHeight - C.P2w - antPAreaR - pm2Len - tongue.PC4 - upIA - pm2Width - aRA - phonexp - tongue.PC6 - R.approximant - tongue.PC1 - English.proficiency - pTransA - age - overjet - tongue.PC3 - m2Height - jLowPC2 - R.trill - tongue.PC2 - 
    lowIA - aRBulge - C.P2l - tongue.PC5 - pRoofA - C.M2w - sex - slMDepth - jLowPC3 

              post.mean l-95% CI u-95% CI eff.samp  pMCMC   
(Intercept)      -0.210   -0.857    0.414    16802 0.5200   
R.flap.tapyes    -1.660   -2.990   -0.411    14721 0.0082 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

              Point est. Upper C.I.
(Intercept)            1          1
R.flap.tapyes          1          1

Multivariate psrf

1

CL

glmer

Call:
glm(formula = dv ~ R.flap.tap, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-1.115  -1.115  -0.591   1.241   1.915  

Coefficients:
              Estimate Std. Error z value Pr(>|z|)  
(Intercept)     -0.148      0.273   -0.54    0.587  
R.flap.tapyes   -1.510      0.610   -2.48    0.013 *
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 103.867  on 78  degrees of freedom
Residual deviance:  96.547  on 77  degrees of freedom
AIC: 100.5

Number of Fisher Scoring iterations: 3

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 99 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency + tongue.PC1 + tongue.PC2 + tongue.PC3 + tongue.PC4 + tongue.PC5 + tongue.PC6 - Retroflexes - R.uvular - group - R.retroflex - sex - C43 - English.proficiency - tongue.PC1 - tongue.PC3 - R.trill - C8 - R.approximant - C27 - tongue.PC5 - C12 - C44 - C41 - phonexp - age - C37 - tongue.PC6 - C40 - tongue.PC4 - tongue.PC2 

              post.mean l-95% CI u-95% CI eff.samp  pMCMC   
(Intercept)      -0.210   -0.835    0.444    19118 0.5152   
R.flap.tapyes    -1.653   -2.935   -0.339    14904 0.0093 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

              Point est. Upper C.I.
(Intercept)            1          1
R.flap.tapyes          1          1

Multivariate psrf

1

PC

glmer

Call:
glm(formula = dv ~ R.flap.tap, family = binomial(link = "logit"), 
    data = data, na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
-1.115  -1.115  -0.591   1.241   1.915  

Coefficients:
              Estimate Std. Error z value Pr(>|z|)  
(Intercept)     -0.148      0.273   -0.54    0.587  
R.flap.tapyes   -1.510      0.610   -2.48    0.013 *
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 103.867  on 78  degrees of freedom
Residual deviance:  96.547  on 77  degrees of freedom
AIC: 100.5

Number of Fisher Scoring iterations: 3

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 99 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency + tongue.PC1 + tongue.PC2 + tongue.PC3 + tongue.PC4 + tongue.PC5 + tongue.PC6 - Retroflexes - R.uvular - group - R.retroflex - sex - tongue.PC4 - tongue.PC3 - phonexp - tongue.PC1 - R.approximant - PC8 - tongue.PC6 - PC5 - age - PC2 - PC12 - English.proficiency - R.trill - PC7 - PC11 - 
    PC1 - tongue.PC5 - PC4 - PC9 - PC13 - tongue.PC2 - PC6 - PC3 - R.flap.tap 

            post.mean l-95% CI u-95% CI eff.samp  pMCMC   
(Intercept)    -0.709   -1.285   -0.132    15381 0.0136 * 
PC10           -0.782   -1.389   -0.203    15769 0.0079 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1
PC10                 1          1

Multivariate psrf

1

Real-time /for/ context (per participant)

AN

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.01   -1.01   -1.01    1.35    1.35  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)   -0.405      0.253    -1.6     0.11

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 87.492  on 64  degrees of freedom
Residual deviance: 87.492  on 64  degrees of freedom
  (15 observations deleted due to missingness)
AIC: 89.49

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 89 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + pRoofA + pTransA + aRA + upIA + pHeight + pLen + pWid + pWHRat + pHLRat + aRHeight + aRDepth + antPArea + antPAreaR + overjet + overbite + slMHeight + slMDepth + ttAvailA + lowIA + jTotPC1 + jTotPC2 + jTotPC3 + jLowPC2 + jLowPC3 + cHeight + pm2Height + m2Height + pm2Len + cWidth + pm2Width + m2Width + C.P2w + C.M2w + C.P2l + aRBulge + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - cWidth - pHeight - pm2Width - 
    pLen - pHLRat - pWHRat - Retroflexes - aRDepth - aRHeight - jTotPC1 - R.uvular - cHeight - jTotPC3 - pm2Height - group - antPArea - ttAvailA - jTotPC2 - overbite - m2Width - R.retroflex - slMHeight - antPAreaR - C.M2w - pm2Len - upIA - pWid - C.P2w - overjet - English.proficiency - m2Height - aRBulge - lowIA - aRA - jLowPC2 - pRoofA - jLowPC3 - pTransA - phonexp - slMDepth - C.P2l - R.flap.tap - sex - R.approximant - R.trill - age 

            post.mean l-95% CI u-95% CI eff.samp pMCMC
(Intercept)    -0.483   -1.053    0.128    16578  0.11

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

CL

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.01   -1.01   -1.01    1.35    1.35  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)   -0.405      0.253    -1.6     0.11

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 87.492  on 64  degrees of freedom
Residual deviance: 87.492  on 64  degrees of freedom
  (15 observations deleted due to missingness)
AIC: 89.49

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 89 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + C43 + C41 + C44 + C40 + C27 + C37 + C12 + C8 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - C40 - English.proficiency - C8 - sex - C27 - C12 - R.approximant - C41 - C37 - C43 - phonexp - R.trill - C44 - R.flap.tap - age 

            post.mean l-95% CI u-95% CI eff.samp pMCMC
(Intercept)    -0.484   -1.071    0.113    15183  0.11

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

PC

glmer

Call:
glm(formula = dv ~ 1, family = binomial(link = "logit"), data = data, 
    na.action = na.omit)

Deviance Residuals: 
   Min      1Q  Median      3Q     Max  
 -1.01   -1.01   -1.01    1.35    1.35  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)
(Intercept)   -0.405      0.253    -1.6     0.11

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 87.492  on 64  degrees of freedom
Residual deviance: 87.492  on 64  degrees of freedom
  (15 observations deleted due to missingness)
AIC: 89.49

Number of Fisher Scoring iterations: 4

MCMCglmm

 Iterations = 100001:999951
 Thinning interval  = 50
 Sample size  = 18000 

 DIC: 89 

 R-structure:  ~units

      post.mean l-95% CI u-95% CI eff.samp
units         1        1        1        0

 Location effects: dv ~ 1 + sex + group + phonexp + age + PC1 + PC2 + PC3 + PC4 + PC5 + PC6 + PC7 + PC8 + PC9 + PC10 + PC11 + PC12 + PC13 + Retroflexes + R.trill + R.retroflex + R.flap.tap + R.approximant + R.uvular + English.proficiency - Retroflexes - R.uvular - group - R.retroflex - sex - PC3 - PC7 - PC9 - phonexp - PC8 - PC4 - R.flap.tap - PC12 - PC11 - English.proficiency - PC5 - PC10 - PC13 - PC6 - PC2 - R.trill - R.approximant - PC1 - age 

            post.mean l-95% CI u-95% CI eff.samp pMCMC
(Intercept)    -0.486   -1.079    0.114    16460  0.11

Gelman-Rubin criterion:

Potential scale reduction factors:

            Point est. Upper C.I.
(Intercept)          1          1

Appendix II: R session info

This document was generated with:

R version 3.4.1 (2017-06-30)

**Platform:** x86_64-pc-linux-gnu (64-bit)

locale: LC_CTYPE=en_US.UTF-8, LC_NUMERIC=C, LC_TIME=en_US.UTF-8, LC_COLLATE=en_US.UTF-8, LC_MONETARY=en_US.UTF-8, LC_MESSAGES=en_US.UTF-8, LC_PAPER=en_US.UTF-8, LC_NAME=C, LC_ADDRESS=C, LC_TELEPHONE=C, LC_MEASUREMENT=en_US.UTF-8 and LC_IDENTIFICATION=C

attached base packages: compiler, grid, parallel, stats, graphics, grDevices, utils, datasets, methods and base

other attached packages: cluster(v.2.0.7-1), fpc(v.2.1-11.1), doBy(v.4.6-2), MASS(v.7.3-50), Hmisc(v.4.1-1), Formula(v.1.2-3), survival(v.2.42-6), lattice(v.0.20-35), factoextra(v.1.0.5), rmarkdown(v.1.10), MCMCglmm(v.2.26), ape(v.5.2), coda(v.0.19-2), lmerTest(v.3.0-1), lme4(v.1.1-18-1), Matrix(v.1.2-14), car(v.3.0-2), carData(v.3.0-2), plyr(v.1.8.4), gplots(v.3.0.1), ggthemes(v.4.0.1), ggplot2(v.3.0.0), reshape2(v.1.4.3), knitr(v.1.20), pander(v.0.6.2), psych(v.1.8.4) and RhpcBLASctl(v.0.18-205)

loaded via a namespace (and not attached): nlme(v.3.1-137), bitops(v.1.0-6), RColorBrewer(v.1.1-2), prabclus(v.2.2-6), rprojroot(v.1.3-2), numDeriv(v.2016.8-1), tensorA(v.0.36.1), tools(v.3.4.1), backports(v.1.1.2), R6(v.2.3.0), rpart(v.4.1-13), KernSmooth(v.2.23-15), lazyeval(v.0.2.1), colorspace(v.1.3-2), trimcluster(v.0.1-2.1), nnet(v.7.3-12), withr(v.2.1.2), tidyselect(v.0.2.4), gridExtra(v.2.3), mnormt(v.1.5-5), curl(v.3.2), htmlTable(v.1.12), labeling(v.0.3), diptest(v.0.75-7), caTools(v.1.17.1.1), scales(v.1.0.0), checkmate(v.1.8.5), DEoptimR(v.1.0-8), robustbase(v.0.93-3), mvtnorm(v.1.0-8), stringr(v.1.3.1), digest(v.0.6.17), foreign(v.0.8-71), minqa(v.1.2.4), rio(v.0.5.10), base64enc(v.0.1-3), pkgconfig(v.2.0.2), htmltools(v.0.3.6), highr(v.0.7), htmlwidgets(v.1.3), rlang(v.0.2.2), readxl(v.1.1.0), rstudioapi(v.0.8), bindr(v.0.1.1), mclust(v.5.4.1), gtools(v.3.8.1), acepack(v.1.4.1), dplyr(v.0.7.6), zip(v.1.0.0), magrittr(v.1.5), modeltools(v.0.2-22), Rcpp(v.0.12.19), munsell(v.0.5.0), abind(v.1.4-5), stringi(v.1.2.4), yaml(v.2.2.0), flexmix(v.2.3-14), gdata(v.2.18.0), ggrepel(v.0.8.0), forcats(v.0.3.0), crayon(v.1.3.4), haven(v.1.1.2), splines(v.3.4.1), hms(v.0.4.2), pillar(v.1.3.0), cubature(v.1.4), corpcor(v.1.6.9), stats4(v.3.4.1), codetools(v.0.2-15), glue(v.1.3.0), evaluate(v.0.11), latticeExtra(v.0.6-28), data.table(v.1.11.8), nloptr(v.1.2.1), cellranger(v.1.1.0), gtable(v.0.2.0), purrr(v.0.2.5), kernlab(v.0.9-27), assertthat(v.0.2.0), openxlsx(v.4.1.0), class(v.7.3-14), tibble(v.1.4.2) and bindrcpp(v.0.2.2)

The compilation took 0.04 minutes.