Mean hypernym at t as predictor of vocab change at t-1
CDI
2019-05-17
library(knitr)
opts_chunk$set(echo = F, message = F, warning = F,
error = F, cache = T, tidy = F)
library(tidyverse)
library(langcog)
library(feather)
library(lme4)
theme_set(theme_classic(base_size = 10))
Each point corresponds to a child. Y-axis shows number of new words learned at t relative to t-1, after residualizing out number of words at t-1, mean frequency of words at t-1, and time between t and t-1. X-axis shows mean hypernym level of vocabulary at t-1. Red line shows linear fit; blue line shows loess fit.
hyponyms_all (old measure)
## Linear mixed model fit by REML ['lmerMod']
## Formula: delta_words_spoken ~ mean_hypernyms + delta_age + mean_freq +
## words_spoken + session_num + (session_num | child_id)
## Data: full_df
##
## REML criterion at convergence: 12541.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -7.1376 -0.4680 -0.1295 0.3069 9.4132
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## child_id (Intercept) 308.026 17.551
## session_num 8.339 2.888 -1.00
## Residual 1947.459 44.130
## Number of obs: 1201, groups: child_id, 195
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 449.54035 40.71426 11.041
## mean_hypernyms -9.34622 2.08265 -4.488
## delta_age 31.95301 1.92041 16.639
## mean_freq -43.18709 4.56591 -9.459
## words_spoken -0.09758 0.01200 -8.134
## session_num 0.95050 0.62094 1.531
##
## Correlation of Fixed Effects:
## (Intr) mn_hyp delt_g mn_frq wrds_s
## mn_hyprnyms -0.420
## delta_age -0.077 0.003
## mean_freq -0.875 -0.064 0.033
## words_spokn -0.664 0.507 0.039 0.455
## session_num 0.033 0.015 -0.129 -0.080 -0.457
## convergence code: 0
## boundary (singular) fit: see ?isSingular## mean_hypernyms delta_age mean_freq words_spoken session_num
## 1.918449 1.021184 1.657444 3.108289 1.604757hypernyms_first
## Linear mixed model fit by REML ['lmerMod']
## Formula:
## delta_words_spoken ~ mean_hypernyms_first + delta_age + mean_freq +
## words_spoken + session_num + (session_num | child_id)
## Data: full_df
##
## REML criterion at convergence: 12549.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -7.1308 -0.4688 -0.1232 0.2990 9.3283
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## child_id (Intercept) 319.240 17.867
## session_num 9.378 3.062 -1.00
## Residual 1957.975 44.249
## Number of obs: 1201, groups: child_id, 195
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 452.87693 44.57847 10.159
## mean_hypernyms_first -8.21987 2.44601 -3.361
## delta_age 32.03955 1.92627 16.633
## mean_freq -45.71765 4.58652 -9.968
## words_spoken -0.08972 0.01213 -7.394
## session_num 0.94895 0.62941 1.508
##
## Correlation of Fixed Effects:
## (Intr) mn_hy_ delt_g mn_frq wrds_s
## mn_hyprnym_ -0.558
## delta_age -0.067 -0.004
## mean_freq -0.881 0.106 0.033
## words_spokn -0.700 0.517 0.035 0.539
## session_num 0.039 -0.003 -0.131 -0.081 -0.458
## convergence code: 0
## boundary (singular) fit: see ?isSingular## mean_hypernyms_first delta_age mean_freq
## 1.692766 1.021513 1.668289
## words_spoken session_num
## 3.123790 1.584606hyponyms_leaf
## Linear mixed model fit by REML ['lmerMod']
## Formula:
## delta_words_spoken ~ log(mean_hypo_leaf) + +delta_age + mean_freq +
## words_spoken + session_num + (session_num | child_id)
## Data: full_df
##
## REML criterion at convergence: 12557.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -7.1171 -0.4718 -0.1245 0.2961 9.2018
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## child_id (Intercept) 348.45 18.667
## session_num 11.08 3.328 -1.00
## Residual 1964.45 44.322
## Number of obs: 1201, groups: child_id, 195
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 338.83430 40.97255 8.270
## log(mean_hypo_leaf) 3.87560 2.55135 1.519
## delta_age 31.99950 1.93189 16.564
## mean_freq -41.97660 4.69588 -8.939
## words_spoken -0.06924 0.01056 -6.560
## session_num 0.98001 0.64394 1.522
##
## Correlation of Fixed Effects:
## (Intr) lg(__) delt_g mn_frq wrds_s
## lg(mn_hyp_) -0.425
## delta_age -0.068 -0.018
## mean_freq -0.975 0.228 0.028
## words_spokn -0.460 -0.144 0.045 0.517
## session_num 0.003 0.086 -0.134 -0.059 -0.530
## convergence code: 0
## boundary (singular) fit: see ?isSingular## log(mean_hypo_leaf) delta_age mean_freq
## 1.184595 1.022195 1.737472
## words_spoken session_num
## 2.306006 1.570062hypernyms_first + hyponyms_leaf
## Linear mixed model fit by REML ['lmerMod']
## Formula:
## delta_words_spoken ~ log(mean_hypo_leaf) + mean_hypernyms_first +
## delta_age + mean_freq + words_spoken + session_num + (session_num |
## child_id)
## Data: full_df
##
## REML criterion at convergence: 12538.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -7.1293 -0.4797 -0.1268 0.3054 9.4546
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## child_id (Intercept) 281.492 16.778
## session_num 8.061 2.839 -1.00
## Residual 1956.889 44.237
## Number of obs: 1201, groups: child_id, 195
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 427.47601 45.54015 9.387
## log(mean_hypo_leaf) 7.07100 2.64397 2.674
## mean_hypernyms_first -10.27685 2.54381 -4.040
## delta_age 31.93948 1.92331 16.607
## mean_freq -43.39452 4.66193 -9.308
## words_spoken -0.09955 0.01256 -7.925
## session_num 1.10541 0.62382 1.772
##
## Correlation of Fixed Effects:
## (Intr) lg(__) mn_hy_ delt_g mn_frq wrds_s
## lg(mn_hyp_) -0.230
## mn_hyprnym_ -0.452 -0.289
## delta_age -0.062 -0.017 0.002
## mean_freq -0.886 0.204 0.038 0.029
## words_spokn -0.592 -0.274 0.555 0.039 0.450
## session_num 0.020 0.089 -0.031 -0.131 -0.061 -0.469
## convergence code: 0
## boundary (singular) fit: see ?isSingular## log(mean_hypo_leaf) mean_hypernyms_first delta_age
## 1.288051 1.855596 1.021751
## mean_freq words_spoken session_num
## 1.742856 3.413037 1.616611