Tristan Mahr 2017-11-07
This script fits the growth curve models for the analyses.
library(dplyr)
library(littlelisteners)
library(ggplot2)
library(lme4)
source("./plotting-helpers.R", encoding = "UTF8")
looks <- readr::read_csv("./data/model.csv.gz") %>%
mutate(
Cond_Lab = Condition %>%
factor(c("real", "MP", "nonsense"),
c("Real word", "Mispronunciation", "Nonword")))looks <- looks %>%
polypoly::poly_add_columns(Time, degree = 3, prefix = "ot") For the models with vocabulary information, we need to remove any pairs of children in which one of the children is missing an EVT score. We identify those pairs.
no_vocab_pairs <- looks %>%
distinct(ChildStudyID, Matching_PairNumber, EVT_Standard) %>%
filter(is.na(EVT_Standard)) %>%
select(Matching_PairNumber) %>%
print()
#> # A tibble: 1 x 1
#> Matching_PairNumber
#> <int>
#> 1 15Prepare a dataset for each condition.
d_mp <- looks %>%
filter(Condition == "MP")
d_rw <- looks %>%
filter(Condition == "real")
d_ns <- looks %>%
filter(Condition == "nonsense")
d_mp_evt <- d_mp %>%
anti_join(no_vocab_pairs, by = "Matching_PairNumber")
d_rw_evt <- d_rw %>%
anti_join(no_vocab_pairs, by = "Matching_PairNumber")
d_ns_evt <- d_ns %>%
anti_join(no_vocab_pairs, by = "Matching_PairNumber")First, we just look at the group effects on the growth curve shapes.
glmer_controls <- glmerControl(
optimizer = "bobyqa",
optCtrl = list(maxfun = 2e5))
m_mp <- glmer(
cbind(Target, Distractor) ~
Group * (ot1 + ot2 + ot3) + (ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_mp)
summary(m_mp)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ Group * (ot1 + ot2 + ot3) + (ot1 +
#> ot2 + ot3 | ChildStudyID)
#> Data: d_mp
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11917.1 12020.4 -5940.6 11881.1 2276
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -2.40445 -0.41933 -0.00934 0.42238 2.39728
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.3694 0.6078
#> ot1 3.8897 1.9722 0.46
#> ot2 0.8285 0.9102 -0.32 -0.32
#> ot3 0.4418 0.6647 0.15 -0.43 -0.07
#> Number of obs: 2294, groups: ChildStudyID, 74
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) -0.004854 0.100360 -0.048 0.9614
#> GroupNormalHearing 0.154643 0.141891 1.090 0.2758
#> ot1 0.327820 0.328509 0.998 0.3183
#> ot2 0.296101 0.158533 1.868 0.0618 .
#> ot3 0.111082 0.120885 0.919 0.3581
#> GroupNormalHearing:ot1 0.734386 0.464265 1.582 0.1137
#> GroupNormalHearing:ot2 0.229322 0.223408 1.026 0.3047
#> GroupNormalHearing:ot3 -0.680799 0.169981 -4.005 6.2e-05 ***
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) GrpNrH ot1 ot2 ot3 GrNH:1 GrNH:2
#> GrpNrmlHrng -0.707
#> ot1 0.451 -0.319
#> ot2 -0.299 0.211 -0.298
#> ot3 0.132 -0.093 -0.386 -0.054
#> GrpNrmlHr:1 -0.319 0.452 -0.707 0.211 0.273
#> GrpNrmlHr:2 0.212 -0.299 0.211 -0.709 0.039 -0.298
#> GrpNrmlHr:3 -0.094 0.132 0.275 0.039 -0.711 -0.388 -0.052
m_rw <- glmer(
cbind(Target, Distractor) ~
Group * (ot1 + ot2 + ot3) + (ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_rw)
summary(m_rw)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ Group * (ot1 + ot2 + ot3) + (ot1 +
#> ot2 + ot3 | ChildStudyID)
#> Data: d_rw
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11558.8 11662.1 -5761.4 11522.8 2276
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -3.16093 -0.42522 0.00612 0.44104 2.44476
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.5225 0.7228
#> ot1 4.8612 2.2048 0.49
#> ot2 1.7769 1.3330 -0.24 -0.19
#> ot3 0.6677 0.8171 -0.28 -0.46 0.34
#> Number of obs: 2294, groups: ChildStudyID, 74
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) 0.7337 0.1194 6.146 7.96e-10 ***
#> GroupNormalHearing 0.4500 0.1688 2.665 0.007697 **
#> ot1 2.3730 0.3681 6.447 1.14e-10 ***
#> ot2 -0.7607 0.2278 -3.340 0.000838 ***
#> ot3 -0.1179 0.1468 -0.803 0.421818
#> GroupNormalHearing:ot1 0.2147 0.5203 0.413 0.679883
#> GroupNormalHearing:ot2 -0.5656 0.3216 -1.759 0.078560 .
#> GroupNormalHearing:ot3 0.1678 0.2067 0.811 0.417100
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) GrpNrH ot1 ot2 ot3 GrNH:1 GrNH:2
#> GrpNrmlHrng -0.707
#> ot1 0.488 -0.345
#> ot2 -0.230 0.163 -0.170
#> ot3 -0.255 0.180 -0.407 0.314
#> GrpNrmlHr:1 -0.345 0.490 -0.707 0.121 0.288
#> GrpNrmlHr:2 0.163 -0.231 0.121 -0.707 -0.220 -0.169
#> GrpNrmlHr:3 0.181 -0.256 0.289 -0.221 -0.707 -0.412 0.317
m_ns <- glmer(
cbind(Target, Distractor) ~
Group * (ot1 + ot2 + ot3) + (ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_ns)
summary(m_ns)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ Group * (ot1 + ot2 + ot3) + (ot1 +
#> ot2 + ot3 | ChildStudyID)
#> Data: d_ns
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11534.9 11638.2 -5749.4 11498.9 2276
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -2.74598 -0.46161 -0.02695 0.45878 2.74382
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.3753 0.6127
#> ot1 5.5977 2.3660 0.47
#> ot2 1.6033 1.2662 0.04 0.24
#> ot3 1.1856 1.0888 -0.16 -0.41 -0.30
#> Number of obs: 2294, groups: ChildStudyID, 74
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) -0.74499 0.10148 -7.341 2.11e-13 ***
#> GroupNormalHearing -0.55025 0.14341 -3.837 0.000125 ***
#> ot1 -1.62901 0.39598 -4.114 3.89e-05 ***
#> ot2 0.19217 0.21894 0.878 0.380103
#> ot3 0.05652 0.18919 0.299 0.765131
#> GroupNormalHearing:ot1 -0.48871 0.55840 -0.875 0.381464
#> GroupNormalHearing:ot2 0.30571 0.30795 0.993 0.320848
#> GroupNormalHearing:ot3 0.01835 0.26714 0.069 0.945227
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) GrpNrH ot1 ot2 ot3 GrNH:1 GrNH:2
#> GrpNrmlHrng -0.707
#> ot1 0.465 -0.329
#> ot2 0.047 -0.033 0.239
#> ot3 -0.147 0.104 -0.374 -0.245
#> GrpNrmlHr:1 -0.330 0.465 -0.708 -0.168 0.265
#> GrpNrmlHr:2 -0.033 0.044 -0.169 -0.709 0.175 0.235
#> GrpNrmlHr:3 0.104 -0.149 0.266 0.175 -0.707 -0.380 -0.251We can plot the growth curve fixed effects which describe how the average child in each group x condition performs.
Allow EVT to interact with time, time2 and time3 because these data are curvier.
m_mp_1a <- glmer(
cbind(Target, Distractor) ~
Group * (1 + ot1 + ot2 + ot3) + (ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_mp_evt)
summary(m_mp_1a)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ Group * (1 + ot1 + ot2 + ot3) + (ot1 +
#> ot2 + ot3 | ChildStudyID)
#> Data: d_mp_evt
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11579.8 11682.6 -5771.9 11543.8 2214
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -2.39938 -0.42304 -0.00919 0.42036 2.39557
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.3718 0.6098
#> ot1 3.8411 1.9599 0.47
#> ot2 0.8374 0.9151 -0.31 -0.35
#> ot3 0.4396 0.6630 0.17 -0.42 -0.07
#> Number of obs: 2232, groups: ChildStudyID, 72
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) -0.01067 0.10208 -0.105 0.917
#> GroupNormalHearing 0.18040 0.14433 1.250 0.211
#> ot1 0.41115 0.33111 1.242 0.214
#> ot2 0.32094 0.16159 1.986 0.047 *
#> ot3 0.09754 0.12240 0.797 0.425
#> GroupNormalHearing:ot1 0.69373 0.46795 1.482 0.138
#> GroupNormalHearing:ot2 0.18965 0.22769 0.833 0.405
#> GroupNormalHearing:ot3 -0.69506 0.17207 -4.039 5.36e-05 ***
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) GrpNrH ot1 ot2 ot3 GrNH:1 GrNH:2
#> GrpNrmlHrng -0.707
#> ot1 0.460 -0.325
#> ot2 -0.289 0.205 -0.320
#> ot3 0.157 -0.111 -0.370 -0.058
#> GrpNrmlHr:1 -0.325 0.461 -0.708 0.226 0.262
#> GrpNrmlHr:2 0.205 -0.290 0.227 -0.710 0.041 -0.320
#> GrpNrmlHr:3 -0.112 0.157 0.263 0.041 -0.711 -0.372 -0.056
m_mp_2a <- glmer(
cbind(Target, Distractor) ~
EVT_GSV_z * (1 + ot1 + ot2 + ot3) + (ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_mp_evt)
summary(m_mp_2a)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ EVT_GSV_z * (1 + ot1 + ot2 + ot3) +
#> (ot1 + ot2 + ot3 | ChildStudyID)
#> Data: d_mp_evt
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11583.2 11685.9 -5773.6 11547.2 2214
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -2.37925 -0.42658 -0.00838 0.41904 2.36470
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.3245 0.5696
#> ot1 3.8271 1.9563 0.45
#> ot2 0.8457 0.9196 -0.33 -0.33
#> ot3 0.5459 0.7388 0.17 -0.43 -0.11
#> Number of obs: 2232, groups: ChildStudyID, 72
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) 0.08006 0.06745 1.187 0.235242
#> EVT_GSV_z 0.23520 0.06749 3.485 0.000492 ***
#> ot1 0.75864 0.23354 3.248 0.001160 **
#> ot2 0.41665 0.11438 3.643 0.000270 ***
#> ot3 -0.25415 0.09427 -2.696 0.007016 **
#> EVT_GSV_z:ot1 0.36940 0.23370 1.581 0.113956
#> EVT_GSV_z:ot2 0.03004 0.11466 0.262 0.793306
#> EVT_GSV_z:ot3 -0.12938 0.09459 -1.368 0.171359
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) EVT_GSV_z ot1 ot2 ot3 EVT_GSV_:1 EVT_GSV_:2
#> EVT_GSV_z 0.003
#> ot1 0.445 0.002
#> ot2 -0.307 -0.001 -0.307
#> ot3 0.157 0.000 -0.387 -0.088
#> EVT_GSV_z:1 0.002 0.445 0.003 0.000 -0.001
#> EVT_GSV_z:2 -0.001 -0.307 0.000 0.000 0.002 -0.305
#> EVT_GSV_z:3 0.000 0.157 -0.001 0.002 -0.003 -0.386 -0.084
m_mp_3a <- glmer(
cbind(Target, Distractor) ~
Group * (1 + ot1 + ot2 + ot3) +
EVT_GSV_z * (1 + ot1 + ot2 + ot3) +
(ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_mp_evt)
summary(m_mp_3a)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ Group * (1 + ot1 + ot2 + ot3) + EVT_GSV_z *
#> (1 + ot1 + ot2 + ot3) + (ot1 + ot2 + ot3 | ChildStudyID)
#> Data: d_mp_evt
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11577.0 11702.7 -5766.5 11533.0 2210
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -2.40132 -0.42262 -0.00877 0.41963 2.38941
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.3244 0.5696
#> ot1 3.7784 1.9438 0.46
#> ot2 0.8369 0.9148 -0.33 -0.35
#> ot3 0.4396 0.6630 0.18 -0.42 -0.07
#> Number of obs: 2232, groups: ChildStudyID, 72
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) 0.08947 0.10034 0.892 0.372586
#> GroupNormalHearing -0.01859 0.14832 -0.125 0.900270
#> ot1 0.52447 0.34526 1.519 0.128742
#> ot2 0.31627 0.16972 1.863 0.062397 .
#> ot3 0.10543 0.12852 0.820 0.412007
#> EVT_GSV_z 0.23911 0.07420 3.222 0.001272 **
#> GroupNormalHearing:ot1 0.46829 0.51007 0.918 0.358569
#> GroupNormalHearing:ot2 0.19823 0.25006 0.793 0.427925
#> GroupNormalHearing:ot3 -0.71001 0.18902 -3.756 0.000172 ***
#> ot1:EVT_GSV_z 0.27172 0.25528 1.064 0.287145
#> ot2:EVT_GSV_z -0.01120 0.12538 -0.089 0.928809
#> ot3:EVT_GSV_z 0.01843 0.09493 0.194 0.846051
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) GrpNrH ot1 ot2 ot3 EVT_GS GrNH:1 GrNH:2 GrNH:3 o1:EVT o2:EVT
#> GrpNrmlHrng -0.740
#> ot1 0.448 -0.332
#> ot2 -0.306 0.227 -0.321
#> ot3 0.160 -0.119 -0.376 -0.058
#> EVT_GSV_z 0.310 -0.416 0.139 -0.095 0.050
#> GrpNrmlHr:1 -0.332 0.449 -0.740 0.238 0.279 -0.186
#> GrpNrmlHr:2 0.228 -0.307 0.238 -0.742 0.044 0.128 -0.322
#> GrpNrmlHr:3 -0.119 0.160 0.280 0.044 -0.743 -0.067 -0.378 -0.058
#> o1:EVT_GSV_ 0.139 -0.186 0.309 -0.099 -0.117 0.449 -0.415 0.134 0.158
#> o2:EVT_GSV_ -0.095 0.128 -0.099 0.307 -0.017 -0.307 0.134 -0.414 0.025 -0.320
#> o3:EVT_GSV_ 0.050 -0.067 -0.117 -0.017 0.305 0.159 0.157 0.025 -0.414 -0.377 -0.053
m_mp_3b <- glmer(
cbind(Target, Distractor) ~
Group * EVT_GSV_z * (1 + ot1 + ot2 + ot3) +
(ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_mp_evt)
summary(m_mp_3b)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ Group * EVT_GSV_z * (1 + ot1 + ot2 +
#> ot3) + (ot1 + ot2 + ot3 | ChildStudyID)
#> Data: d_mp_evt
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11582.9 11731.4 -5765.5 11530.9 2206
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -2.40132 -0.41913 -0.00741 0.41865 2.39107
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.3244 0.5696
#> ot1 3.7072 1.9254 0.46
#> ot2 0.8316 0.9119 -0.33 -0.34
#> ot3 0.4365 0.6607 0.18 -0.42 -0.08
#> Number of obs: 2232, groups: ChildStudyID, 72
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) 0.09196 0.10165 0.905 0.365658
#> GroupNormalHearing -0.01222 0.15422 -0.079 0.936843
#> EVT_GSV_z 0.24497 0.08374 2.925 0.003442 **
#> ot1 0.45941 0.34663 1.325 0.185052
#> ot2 0.33355 0.17144 1.946 0.051705 .
#> ot3 0.11899 0.12980 0.917 0.359287
#> GroupNormalHearing:EVT_GSV_z -0.02703 0.18064 -0.150 0.881045
#> GroupNormalHearing:ot1 0.29984 0.52558 0.570 0.568345
#> GroupNormalHearing:ot2 0.24291 0.25915 0.937 0.348584
#> GroupNormalHearing:ot3 -0.67521 0.19583 -3.448 0.000565 ***
#> EVT_GSV_z:ot1 0.11674 0.28550 0.409 0.682615
#> EVT_GSV_z:ot2 0.02992 0.14123 0.212 0.832229
#> EVT_GSV_z:ot3 0.05073 0.10706 0.474 0.635600
#> GroupNormalHearing:EVT_GSV_z:ot1 0.71998 0.61583 1.169 0.242351
#> GroupNormalHearing:EVT_GSV_z:ot2 -0.19018 0.30364 -0.626 0.531105
#> GroupNormalHearing:EVT_GSV_z:ot3 -0.14996 0.22917 -0.654 0.512874
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1The last two plots are from the Group + Vocab models.
Model comparison here should be taken with a grain of salt because we are adding predictors in batches of four coefficients (the predictor's main effect and interactions with time). AIC and BIC penalize fit measures based on number of parameters, so we are adding big penalities with each additional predictor. Consider this more of coarse look at how batches of coefficients improve model fit.
do.call(anova, unname(models))
#> Data: d_mp_evt
#> Models:
#> MODEL1: cbind(Target, Distractor) ~ Group * (1 + ot1 + ot2 + ot3) + (ot1 +
#> MODEL1: ot2 + ot3 | ChildStudyID)
#> MODEL2: cbind(Target, Distractor) ~ EVT_GSV_z * (1 + ot1 + ot2 + ot3) +
#> MODEL2: (ot1 + ot2 + ot3 | ChildStudyID)
#> MODEL3: cbind(Target, Distractor) ~ Group * (1 + ot1 + ot2 + ot3) + EVT_GSV_z *
#> MODEL3: (1 + ot1 + ot2 + ot3) + (ot1 + ot2 + ot3 | ChildStudyID)
#> MODEL4: cbind(Target, Distractor) ~ Group * EVT_GSV_z * (1 + ot1 + ot2 +
#> MODEL4: ot3) + (ot1 + ot2 + ot3 | ChildStudyID)
#> Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
#> MODEL1 18 11580 11683 -5771.9 11544
#> MODEL2 18 11583 11686 -5773.6 11547 0.0000 0 1.000000
#> MODEL3 22 11577 11703 -5766.5 11533 14.1251 4 0.006906 **
#> MODEL4 26 11583 11731 -5765.5 11531 2.1158 4 0.714474
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Allow EVT to interact with time1, time2 and time3.
m_rw_1a <- glmer(
cbind(Target, Distractor) ~
Group * (1 + ot1 + ot2 + ot3) + (ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_rw %>% anti_join(no_vocab_pairs))
summary(m_rw_1a)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ Group * (1 + ot1 + ot2 + ot3) + (ot1 +
#> ot2 + ot3 | ChildStudyID)
#> Data: d_rw %>% anti_join(no_vocab_pairs)
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11229.5 11332.3 -5596.7 11193.5 2214
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -3.1823 -0.4226 0.0039 0.4436 2.4463
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.5384 0.7337
#> ot1 4.7604 2.1818 0.51
#> ot2 1.8305 1.3530 -0.24 -0.18
#> ot3 0.6799 0.8245 -0.28 -0.44 0.35
#> Number of obs: 2232, groups: ChildStudyID, 72
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) 0.7343 0.1228 5.978 2.26e-09 ***
#> GroupNormalHearing 0.4530 0.1737 2.607 0.009124 **
#> ot1 2.4761 0.3695 6.701 2.07e-11 ***
#> ot2 -0.7779 0.2343 -3.321 0.000899 ***
#> ot3 -0.1233 0.1502 -0.821 0.411515
#> GroupNormalHearing:ot1 0.1734 0.5224 0.332 0.739879
#> GroupNormalHearing:ot2 -0.5336 0.3307 -1.613 0.106676
#> GroupNormalHearing:ot3 0.1441 0.2114 0.682 0.495441
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) GrpNrH ot1 ot2 ot3 GrNH:1 GrNH:2
#> GrpNrmlHrng -0.707
#> ot1 0.501 -0.354
#> ot2 -0.230 0.163 -0.167
#> ot3 -0.254 0.180 -0.397 0.326
#> GrpNrmlHr:1 -0.354 0.502 -0.707 0.119 0.281
#> GrpNrmlHr:2 0.163 -0.231 0.118 -0.707 -0.229 -0.166
#> GrpNrmlHr:3 0.180 -0.256 0.283 -0.230 -0.707 -0.402 0.330
m_rw_2a <- glmer(
cbind(Target, Distractor) ~
EVT_GSV_z * (1 + ot1 + ot2 + ot3) + ot2 + ot3 + (ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_rw %>% anti_join(no_vocab_pairs))
summary(m_rw_2a)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ EVT_GSV_z * (1 + ot1 + ot2 + ot3) +
#> ot2 + ot3 + (ot1 + ot2 + ot3 | ChildStudyID)
#> Data: d_rw %>% anti_join(no_vocab_pairs)
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11227.5 11330.3 -5595.8 11191.5 2214
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -3.1859 -0.4273 0.0032 0.4436 2.4542
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.5300 0.7280
#> ot1 4.6949 2.1668 0.48
#> ot2 1.7957 1.3400 -0.22 -0.16
#> ot3 0.6842 0.8271 -0.28 -0.45 0.35
#> Number of obs: 2232, groups: ChildStudyID, 72
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) 0.96168 0.08622 11.154 < 2e-16 ***
#> EVT_GSV_z 0.24461 0.08612 2.840 0.00451 **
#> ot1 2.56292 0.25962 9.872 < 2e-16 ***
#> ot2 -1.04557 0.16429 -6.364 1.96e-10 ***
#> ot3 -0.05153 0.10644 -0.484 0.62827
#> EVT_GSV_z:ot1 0.26832 0.25862 1.038 0.29949
#> EVT_GSV_z:ot2 -0.33198 0.16292 -2.038 0.04158 *
#> EVT_GSV_z:ot3 0.06182 0.10485 0.590 0.55545
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) EVT_GSV_z ot1 ot2 ot3 EVT_GSV_:1 EVT_GSV_:2
#> EVT_GSV_z 0.003
#> ot1 0.480 0.002
#> ot2 -0.213 -0.002 -0.145
#> ot3 -0.257 -0.001 -0.407 0.334
#> EVT_GSV_z:1 0.002 0.479 0.002 0.001 -0.003
#> EVT_GSV_z:2 -0.002 -0.215 0.001 0.002 0.005 -0.152
#> EVT_GSV_z:3 -0.001 -0.259 -0.002 0.005 0.000 -0.414 0.326
m_rw_3a <- glmer(
cbind(Target, Distractor) ~
Group * (1 + ot1 + ot2 + ot3) +
EVT_GSV_z * (1 + ot1 + ot2 + ot3) +
(ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_rw %>% anti_join(no_vocab_pairs))
summary(m_rw_3a)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ Group * (1 + ot1 + ot2 + ot3) + EVT_GSV_z *
#> (1 + ot1 + ot2 + ot3) + (ot1 + ot2 + ot3 | ChildStudyID)
#> Data: d_rw %>% anti_join(no_vocab_pairs)
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11230.8 11356.5 -5593.4 11186.8 2210
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -3.1835 -0.4238 0.0012 0.4436 2.4479
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.5108 0.7147
#> ot1 4.6931 2.1664 0.50
#> ot2 1.7721 1.3312 -0.21 -0.17
#> ot3 0.6816 0.8256 -0.30 -0.45 0.36
#> Number of obs: 2232, groups: ChildStudyID, 72
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) 0.81056 0.12590 6.438 1.21e-10 ***
#> GroupNormalHearing 0.30150 0.18622 1.619 0.105438
#> ot1 2.59386 0.38612 6.718 1.85e-11 ***
#> ot2 -0.88859 0.24305 -3.656 0.000256 ***
#> ot3 -0.10784 0.15827 -0.681 0.495628
#> EVT_GSV_z 0.18188 0.09303 1.955 0.050569 .
#> GroupNormalHearing:ot1 -0.06123 0.57137 -0.107 0.914664
#> GroupNormalHearing:ot2 -0.31230 0.35934 -0.869 0.384799
#> GroupNormalHearing:ot3 0.11041 0.23354 0.473 0.636381
#> ot1:EVT_GSV_z 0.28151 0.28485 0.988 0.323025
#> ot2:EVT_GSV_z -0.26686 0.17859 -1.494 0.135092
#> ot3:EVT_GSV_z 0.03871 0.11550 0.335 0.737502
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) GrpNrH ot1 ot2 ot3 EVT_GS GrNH:1 GrNH:2 GrNH:3 o1:EVT o2:EVT
#> GrpNrmlHrng -0.740
#> ot1 0.490 -0.363
#> ot2 -0.198 0.146 -0.148
#> ot3 -0.271 0.200 -0.404 0.340
#> EVT_GSV_z 0.310 -0.417 0.153 -0.061 -0.083
#> GrpNrmlHr:1 -0.363 0.492 -0.740 0.109 0.299 -0.206
#> GrpNrmlHr:2 0.146 -0.197 0.109 -0.741 -0.252 0.081 -0.145
#> GrpNrmlHr:3 0.201 -0.272 0.300 -0.252 -0.741 0.113 -0.406 0.343
#> o1:EVT_GSV_ 0.153 -0.206 0.312 -0.043 -0.125 0.491 -0.420 0.058 0.168
#> o2:EVT_GSV_ -0.061 0.082 -0.043 0.314 0.111 -0.199 0.058 -0.422 -0.147 -0.150
#> o3:EVT_GSV_ -0.085 0.114 -0.127 0.111 0.313 -0.273 0.170 -0.147 -0.423 -0.411 0.338
m_rw_3b <- glmer(
cbind(Target, Distractor) ~
Group * EVT_GSV_z * (1 + ot1 + ot2 + ot3) +
(ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_rw %>% anti_join(no_vocab_pairs))
summary(m_rw_3b)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ Group * EVT_GSV_z * (1 + ot1 + ot2 +
#> ot3) + (ot1 + ot2 + ot3 | ChildStudyID)
#> Data: d_rw %>% anti_join(no_vocab_pairs)
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11229.8 11378.3 -5588.9 11177.8 2206
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -3.14537 -0.42363 0.00127 0.44243 2.44166
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.4898 0.6998
#> ot1 4.4008 2.0978 0.47
#> ot2 1.7476 1.3220 -0.24 -0.20
#> ot3 0.6570 0.8105 -0.27 -0.43 0.39
#> Number of obs: 2232, groups: ChildStudyID, 72
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) 0.84577 0.12493 6.770 1.29e-11 ***
#> GroupNormalHearing 0.39437 0.18976 2.078 0.037687 *
#> EVT_GSV_z 0.26639 0.10278 2.592 0.009546 **
#> ot1 2.72474 0.37928 7.184 6.77e-13 ***
#> ot2 -0.85015 0.24477 -3.473 0.000514 ***
#> ot3 -0.14386 0.15799 -0.911 0.362527
#> GroupNormalHearing:EVT_GSV_z -0.39406 0.22210 -1.774 0.076026 .
#> GroupNormalHearing:ot1 0.28480 0.57711 0.493 0.621664
#> GroupNormalHearing:ot2 -0.20813 0.37205 -0.559 0.575889
#> GroupNormalHearing:ot3 0.01236 0.23981 0.052 0.958903
#> EVT_GSV_z:ot1 0.59507 0.31120 1.912 0.055858 .
#> EVT_GSV_z:ot2 -0.17237 0.19999 -0.862 0.388754
#> EVT_GSV_z:ot3 -0.05080 0.12810 -0.397 0.691673
#> GroupNormalHearing:EVT_GSV_z:ot1 -1.46614 0.67440 -2.174 0.029705 *
#> GroupNormalHearing:EVT_GSV_z:ot2 -0.43681 0.43398 -1.007 0.314162
#> GroupNormalHearing:EVT_GSV_z:ot3 0.41554 0.27846 1.492 0.135633
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1The last two plots are from the Group + Vocab model.
This final plot is from the Group x Vocab model. I think the Normal Hearing side is overly sensitive to some outlying behavior in the high EVT normal hearing group.
ggplot(fits %>% filter(Model == "Group x Vocab.")) +
aes(x = Time, y = fitted, color = quantile) +
hline_chance() +
geom_line(
aes(x = Time, y = fitted, group = interaction(Group, quantile)),
size = 1) +
facet_wrap("Group_Lab") +
viridis::scale_color_viridis(discrete = TRUE, end = .85, option = "B") +
labs(x = plot_text$x_time,
y = plot_text$y_fits,
color = "Within group EVT quantiles") +
legend_bottom(legend.justification = "right") +
align_axis_right() +
ggtitle("Vocabulary effects in correct production condition")
d_rw %>%
anti_join(no_vocab_pairs) %>%
left_join(condition_labels) %>%
group_by(Group) %>%
mutate(EVT_Bin = ntile(EVT_GSV, 5) %>% factor()) %>%
ungroup() %>%
ggplot() +
aes(x = Time, y = Prop, color = EVT_Bin) +
hline_chance() +
stat_mean_se() +
facet_wrap("Group_Lab") +
viridis::scale_color_viridis(discrete = TRUE, end = .85, option = "B") +
labs(x = plot_text$x_time,
y = plot_text$y_fits,
color = "Within group EVT bins") +
legend_bottom(legend.justification = "right") +
align_axis_right() +
ggtitle("Vocabulary effects in correct production condition")
do.call(anova, unname(models))
#> Data: d_rw %>% anti_join(no_vocab_pairs)
#> Models:
#> MODEL1: cbind(Target, Distractor) ~ Group * (1 + ot1 + ot2 + ot3) + (ot1 +
#> MODEL1: ot2 + ot3 | ChildStudyID)
#> MODEL2: cbind(Target, Distractor) ~ EVT_GSV_z * (1 + ot1 + ot2 + ot3) +
#> MODEL2: ot2 + ot3 + (ot1 + ot2 + ot3 | ChildStudyID)
#> MODEL3: cbind(Target, Distractor) ~ Group * (1 + ot1 + ot2 + ot3) + EVT_GSV_z *
#> MODEL3: (1 + ot1 + ot2 + ot3) + (ot1 + ot2 + ot3 | ChildStudyID)
#> MODEL4: cbind(Target, Distractor) ~ Group * EVT_GSV_z * (1 + ot1 + ot2 +
#> MODEL4: ot3) + (ot1 + ot2 + ot3 | ChildStudyID)
#> Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
#> MODEL1 18 11230 11332 -5596.7 11194
#> MODEL2 18 11228 11330 -5595.8 11192 1.9524 0 < 2e-16 ***
#> MODEL3 22 11231 11356 -5593.4 11187 4.6992 4 0.31957
#> MODEL4 26 11230 11378 -5588.9 11178 9.0176 4 0.06066 .
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Allow EVT to interact with time1.
m_ns_1a <- glmer(
cbind(Target, Distractor) ~
Group * (1 + ot1) + ot2 + ot3 + (ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_ns %>% anti_join(no_vocab_pairs))
summary(m_ns_1a)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ Group * (1 + ot1) + ot2 + ot3 + (ot1 +
#> ot2 + ot3 | ChildStudyID)
#> Data: d_ns %>% anti_join(no_vocab_pairs)
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11234.3 11325.6 -5601.1 11202.3 2216
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -2.75401 -0.45857 -0.02161 0.46262 2.79496
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.3713 0.6093
#> ot1 5.6498 2.3769 0.46
#> ot2 1.6375 1.2797 0.06 0.24
#> ot3 1.0864 1.0423 -0.11 -0.38 -0.30
#> Number of obs: 2232, groups: ChildStudyID, 72
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) -0.73810 0.10243 -7.206 5.77e-13 ***
#> GroupNormalHearing -0.53385 0.14506 -3.680 0.000233 ***
#> ot1 -1.58575 0.38863 -4.080 4.50e-05 ***
#> ot2 0.31183 0.15808 1.973 0.048545 *
#> ot3 0.04995 0.13045 0.383 0.701774
#> GroupNormalHearing:ot1 -0.53819 0.52867 -1.018 0.308679
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) GrpNrH ot1 ot2 ot3
#> GrpNrmlHrng -0.708
#> ot1 0.449 -0.303
#> ot2 0.040 0.004 0.176
#> ot3 -0.073 0.003 -0.256 -0.244
#> GrpNrmlHr:1 -0.316 0.447 -0.681 0.004 0.002
m_ns_2a <- glmer(
cbind(Target, Distractor) ~
EVT_GSV_z * (1 + ot1) + ot2 + ot3 + (ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_ns %>% anti_join(no_vocab_pairs))
summary(m_ns_2a)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ EVT_GSV_z * (1 + ot1) + ot2 + ot3 +
#> (ot1 + ot2 + ot3 | ChildStudyID)
#> Data: d_ns %>% anti_join(no_vocab_pairs)
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11227.4 11318.7 -5597.7 11195.4 2216
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -2.75440 -0.45644 -0.02429 0.46163 2.70958
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.3415 0.5844
#> ot1 5.5813 2.3625 0.44
#> ot2 1.6299 1.2767 0.02 0.23
#> ot3 1.0829 1.0406 -0.18 -0.42 -0.30
#> Number of obs: 2232, groups: ChildStudyID, 72
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) -1.00615 0.06940 -14.498 < 2e-16 ***
#> EVT_GSV_z -0.32791 0.06892 -4.758 1.95e-06 ***
#> ot1 -1.85436 0.28294 -6.554 5.61e-11 ***
#> ot2 0.31633 0.15765 2.007 0.0448 *
#> ot3 0.05420 0.13021 0.416 0.6772
#> EVT_GSV_z:ot1 -0.46303 0.25986 -1.782 0.0748 .
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) EVT_GSV_z ot1 ot2 ot3
#> EVT_GSV_z 0.005
#> ot1 0.438 0.005
#> ot2 0.023 0.001 0.234
#> ot3 -0.169 -0.001 -0.379 -0.249
#> EVT_GSV_z:1 0.005 0.416 0.008 0.005 0.000
m_ns_3a <- glmer(
cbind(Target, Distractor) ~
Group * (1 + ot1 + ot2 + ot3) +
EVT_GSV_z * (1 + ot1) +
(ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_ns %>% anti_join(no_vocab_pairs))
summary(m_ns_3a)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ Group * (1 + ot1 + ot2 + ot3) + EVT_GSV_z *
#> (1 + ot1) + (ot1 + ot2 + ot3 | ChildStudyID)
#> Data: d_ns %>% anti_join(no_vocab_pairs)
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11229.3 11343.5 -5594.6 11189.3 2212
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -2.74976 -0.45788 -0.02647 0.46007 2.70574
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.3203 0.5659
#> ot1 5.5747 2.3611 0.45
#> ot2 1.5988 1.2644 0.05 0.24
#> ot3 1.0802 1.0393 -0.18 -0.41 -0.30
#> Number of obs: 2232, groups: ChildStudyID, 72
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) -0.85358 0.10018 -8.521 < 2e-16 ***
#> GroupNormalHearing -0.30431 0.14792 -2.057 0.039660 *
#> ot1 -1.84677 0.41933 -4.404 1.06e-05 ***
#> ot2 0.13780 0.22178 0.621 0.534375
#> ot3 0.10082 0.18408 0.548 0.583900
#> EVT_GSV_z -0.26232 0.07359 -3.565 0.000364 ***
#> GroupNormalHearing:ot1 -0.01545 0.61441 -0.025 0.979935
#> GroupNormalHearing:ot2 0.35369 0.31171 1.135 0.256504
#> GroupNormalHearing:ot3 -0.09483 0.25966 -0.365 0.714964
#> ot1:EVT_GSV_z -0.42565 0.28620 -1.487 0.136956
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) GrpNrH ot1 ot2 ot3 EVT_GS GrNH:1 GrNH:2 GrNH:3
#> GrpNrmlHrng -0.741
#> ot1 0.447 -0.328
#> ot2 0.057 -0.038 0.232
#> ot3 -0.158 0.108 -0.355 -0.244
#> EVT_GSV_z 0.312 -0.417 0.126 0.001 -0.001
#> GrpNrmlHr:1 -0.331 0.444 -0.738 -0.159 0.242 -0.168
#> GrpNrmlHr:2 -0.040 0.051 -0.165 -0.709 0.175 -0.002 0.220
#> GrpNrmlHr:3 0.112 -0.153 0.252 0.175 -0.708 -0.002 -0.347 -0.251
#> o1:EVT_GSV_ 0.136 -0.179 0.294 0.008 0.001 0.423 -0.393 -0.007 -0.001
m_ns_3b <- glmer(
cbind(Target, Distractor) ~
Group * (1 + ot1 + ot2 + ot3) +
Group * EVT_GSV_z * (1 + ot1) +
(ot1 + ot2 + ot3 | ChildStudyID),
family = binomial,
control = glmer_controls,
data = d_ns %>% anti_join(no_vocab_pairs))
summary(m_ns_3b)
#> Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
#> Family: binomial ( logit )
#> Formula: cbind(Target, Distractor) ~ Group * (1 + ot1 + ot2 + ot3) + Group *
#> EVT_GSV_z * (1 + ot1) + (ot1 + ot2 + ot3 | ChildStudyID)
#> Data: d_ns %>% anti_join(no_vocab_pairs)
#> Control: glmer_controls
#>
#> AIC BIC logLik deviance df.resid
#> 11231.3 11356.9 -5593.6 11187.3 2210
#>
#> Scaled residuals:
#> Min 1Q Median 3Q Max
#> -2.74983 -0.45753 -0.02828 0.46153 2.75570
#>
#> Random effects:
#> Groups Name Variance Std.Dev. Corr
#> ChildStudyID (Intercept) 0.3177 0.5637
#> ot1 5.3901 2.3217 0.44
#> ot2 1.5982 1.2642 0.05 0.24
#> ot3 1.0808 1.0396 -0.17 -0.40 -0.30
#> Number of obs: 2232, groups: ChildStudyID, 72
#>
#> Fixed effects:
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) -0.86416 0.10106 -8.551 < 2e-16 ***
#> GroupNormalHearing -0.33296 0.15322 -2.173 0.029769 *
#> ot1 -1.93199 0.41765 -4.626 3.73e-06 ***
#> ot2 0.13760 0.22176 0.621 0.534925
#> ot3 0.10026 0.18414 0.544 0.586115
#> EVT_GSV_z -0.28762 0.08236 -3.492 0.000479 ***
#> GroupNormalHearing:ot1 -0.24636 0.62593 -0.394 0.693876
#> GroupNormalHearing:ot2 0.35362 0.31170 1.135 0.256580
#> GroupNormalHearing:ot3 -0.09416 0.25977 -0.362 0.716983
#> GroupNormalHearing:EVT_GSV_z 0.12031 0.17741 0.678 0.497670
#> ot1:EVT_GSV_z -0.62835 0.31711 -1.982 0.047534 *
#> GroupNormalHearing:ot1:EVT_GSV_z 0.96888 0.68391 1.417 0.156575
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> Correlation of Fixed Effects:
#> (Intr) GrpNrH ot1 ot2 ot3 EVT_GS GrNH:1 GrNH:2 GrNH:3 GNH:EV o1:EVT
#> GrpNrmlHrng -0.661
#> ot1 0.439 -0.290
#> ot2 0.055 -0.034 0.229
#> ot3 -0.151 0.101 -0.343 -0.244
#> EVT_GSV_z 0.346 -0.233 0.139 0.003 0.000
#> GrpNrmlHr:1 -0.294 0.435 -0.668 -0.151 0.231 -0.095
#> GrpNrmlHr:2 -0.039 0.047 -0.162 -0.709 0.175 -0.003 0.210
#> GrpNrmlHr:3 0.107 -0.142 0.244 0.175 -0.708 -0.002 -0.330 -0.252
#> GNH:EVT_GSV -0.157 -0.273 -0.064 -0.005 -0.003 -0.454 -0.105 0.003 0.001
#> o1:EVT_GSV_ 0.149 -0.100 0.329 0.011 0.004 0.418 -0.224 -0.008 -0.003 -0.190
#> GNH:1:EVT_G -0.068 -0.112 -0.151 -0.008 -0.007 -0.189 -0.256 0.004 0.004 0.415 -0.454The last two plots are from the Group + Vocab model.
do.call(anova, unname(models))
#> Data: d_ns %>% anti_join(no_vocab_pairs)
#> Models:
#> MODEL1: cbind(Target, Distractor) ~ Group * (1 + ot1) + ot2 + ot3 + (ot1 +
#> MODEL1: ot2 + ot3 | ChildStudyID)
#> MODEL2: cbind(Target, Distractor) ~ EVT_GSV_z * (1 + ot1) + ot2 + ot3 +
#> MODEL2: (ot1 + ot2 + ot3 | ChildStudyID)
#> MODEL3: cbind(Target, Distractor) ~ Group * (1 + ot1 + ot2 + ot3) + EVT_GSV_z *
#> MODEL3: (1 + ot1) + (ot1 + ot2 + ot3 | ChildStudyID)
#> MODEL4: cbind(Target, Distractor) ~ Group * (1 + ot1 + ot2 + ot3) + Group *
#> MODEL4: EVT_GSV_z * (1 + ot1) + (ot1 + ot2 + ot3 | ChildStudyID)
#> Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
#> MODEL1 16 11234 11326 -5601.1 11202
#> MODEL2 16 11227 11319 -5597.7 11195 6.9087 0 <2e-16 ***
#> MODEL3 20 11229 11344 -5594.6 11189 6.0965 4 0.1921
#> MODEL4 22 11231 11357 -5593.6 11187 1.9973 2 0.3684
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1d_ns %>%
anti_join(no_vocab_pairs) %>%
distinct(Group_Lab, ResearchID, ChildStudyID, EVT_GSV) %>%
ggplot() +
theme(axis.text.x = element_text(size = rel(1))) +
aes(x = Group_Lab, y = EVT_GSV) +
geom_boxplot(width = .4) +
geom_point(position = position_jitter(.1), shape = 1) +
labs(x = NULL, y = "EVT-2 GSV") +
ggtitle("Children with CIs have more variable vocabularies") +
align_axis_right()
sessioninfo::session_info()
#> - Session info -----------------------------------------------------------------------------------
#> setting value
#> version R version 3.4.1 (2017-06-30)
#> os Windows 7 x64 SP 1
#> system x86_64, mingw32
#> ui RTerm
#> language (EN)
#> collate English_United States.1252
#> tz America/Chicago
#> date 2017-11-07
#>
#> - Packages ---------------------------------------------------------------------------------------
#> package * version date source
#> assertthat 0.2.0 2017-04-11 CRAN (R 3.3.2)
#> backports 1.1.1 2017-09-25 CRAN (R 3.4.1)
#> bindr 0.1 2016-11-13 CRAN (R 3.4.0)
#> bindrcpp * 0.2 2017-06-17 CRAN (R 3.4.0)
#> clisymbols 1.2.0 2017-08-04 Github (gaborcsardi/clisymbols@e49b4f5)
#> colorspace 1.3-2 2016-12-14 CRAN (R 3.3.2)
#> digest 0.6.12 2017-01-27 CRAN (R 3.3.2)
#> dplyr * 0.7.4 2017-09-28 CRAN (R 3.4.2)
#> evaluate 0.10.1 2017-06-24 CRAN (R 3.4.1)
#> ggplot2 * 2.2.1 2016-12-30 CRAN (R 3.4.1)
#> glue 1.2.0 2017-10-29 CRAN (R 3.4.2)
#> gridExtra 2.3 2017-09-09 CRAN (R 3.4.1)
#> gtable 0.2.0 2016-02-26 CRAN (R 3.2.3)
#> hms 0.3 2016-11-22 CRAN (R 3.3.2)
#> htmltools 0.3.6 2017-04-28 CRAN (R 3.4.0)
#> knitr * 1.17 2017-08-10 CRAN (R 3.4.2)
#> labeling 0.3 2014-08-23 CRAN (R 3.1.1)
#> lattice 0.20-35 2017-03-25 CRAN (R 3.3.3)
#> lazyeval 0.2.1 2017-10-29 CRAN (R 3.4.2)
#> littlelisteners * 0.0.0.9000 2017-09-22 Github (tjmahr/littlelisteners@44e87a4)
#> lme4 * 1.1-14 2017-09-27 CRAN (R 3.4.2)
#> magrittr 1.5 2014-11-22 CRAN (R 3.1.2)
#> MASS 7.3-47 2017-02-26 CRAN (R 3.4.1)
#> Matrix * 1.2-11 2017-08-16 CRAN (R 3.4.2)
#> minqa 1.2.4 2014-10-09 CRAN (R 3.1.1)
#> munsell 0.4.3 2016-02-13 CRAN (R 3.2.3)
#> nlme 3.1-131 2017-02-06 CRAN (R 3.4.1)
#> nloptr 1.0.4 2014-08-04 CRAN (R 3.1.1)
#> pkgconfig 2.0.1 2017-03-21 CRAN (R 3.3.3)
#> plyr 1.8.4 2016-06-08 CRAN (R 3.3.0)
#> polypoly 0.0.2 2017-05-30 Github (tjmahr/polypoly@4370bb5)
#> R6 2.2.2 2017-06-17 CRAN (R 3.4.0)
#> Rcpp 0.12.13 2017-09-28 CRAN (R 3.4.2)
#> readr 1.1.1 2017-05-16 CRAN (R 3.4.0)
#> rlang * 0.1.4 2017-11-05 CRAN (R 3.4.2)
#> rmarkdown 1.6 2017-06-15 CRAN (R 3.4.2)
#> rprojroot 1.2 2017-01-16 CRAN (R 3.3.2)
#> scales 0.5.0 2017-08-24 CRAN (R 3.4.1)
#> sessioninfo 1.0.1 2017-09-13 Github (r-lib/sessioninfo@e813de4)
#> stringi 1.1.5 2017-04-07 CRAN (R 3.3.3)
#> stringr 1.2.0 2017-02-18 CRAN (R 3.3.2)
#> tibble 1.3.4 2017-08-22 CRAN (R 3.4.1)
#> viridis 0.4.0 2017-03-27 CRAN (R 3.3.3)
#> viridisLite 0.2.0 2017-03-24 CRAN (R 3.3.2)
#> withr 2.1.0.9000 2017-11-02 Github (jimhester/withr@8ba5e46)
#> yaml 2.1.14 2016-11-12 CRAN (R 3.4.2)