MTLD words by usefuleness
defined as trigrams from cocoa
2018-05-02
library(knitr)
opts_chunk$set(echo = T, message = F, warning = F,
error = F, cache = F, tidy = F)
library(tidyverse)
library(langcog)
library(data.table)
library(feather)
library(broom)
theme_set(theme_classic(base_size = 10))Do kids that know words at T2 (using mtld measure) know words that are more “useful” at T1? In particular, we define useful words to be words that allow you to make more trigrams in adult speech using coca spoken corpus (https://corpus.byu.edu/coca/).
FINDINGS:
- no evidence for more probable trigrams for high vs. low kids at either group or individual level
- Evidence that high kids have words for more frequent trigrams at group but not individual level
- effect somewhat sensitive to number of words cutoff (effect bigger with fewer words)
- but, lots of trigrams missing
Group level
# Get all unique words for two groups at t1.
groups_info <- read_csv("../1_exploration/groups_info.csv")
target_types <- read_csv("../1_exploration/target_types_delta_450_1150.csv") %>%
mutate(tbin = fct_recode(tbin,
t1 = "low",
t2 = "high")) %>%
mutate(gloss = tolower(gloss))
TOPNWORDS_PER_GROUP <- 500
t1_words <- target_types %>%
filter(tbin == "t1") %>% # #t1 only
left_join(groups_info %>%
mutate(target_child_id = as.numeric(target_child_id))) %>% # merge in group info
group_by(delta_resid_group, gloss) %>% # get glosses by group
summarize(n = n()) %>% # get gloss counts by group
top_n(TOPNWORDS_PER_GROUP) %>%
group_by(delta_resid_group) %>%
slice(1:TOPNWORDS_PER_GROUP) %>%
arrange(delta_resid_group, -n)coca_trigrams <- read_csv("results_freq_prob_coca.csv") %>%
select(w1, w2, w3, prob3, freq3) %>%
data.table()
t1_words_high <- t1_words %>%
filter(delta_resid_group == "high") %>%
select(gloss) %>%
unlist(use.names = F)
t1_words_low <- t1_words %>%
filter(delta_resid_group == "low") %>%
select(gloss) %>%
unlist(use.names = F)
high_trigrams <- coca_trigrams[w1 %in% t1_words_high &
w2 %in% t1_words_high &
w3 %in% t1_words_high] %>%
mutate(vocab_group = "high")
low_trigrams <- coca_trigrams[w1 %in% t1_words_low &
w2 %in% t1_words_low &
w3 %in% t1_words_low] %>%
mutate(vocab_group = "low")
all_trigrams <- bind_rows(high_trigrams, low_trigrams) %>%
mutate(log_freq = log(freq3))ggplot(all_trigrams, aes(x = prob3,
fill = vocab_group,
group = vocab_group)) +
geom_density(alpha = .4)
all_trigrams %>%
group_by(vocab_group) %>%
multi_boot_standard(col = "prob3", na.rm = T) %>%
ggplot(aes(x = vocab_group, y = mean, group = vocab_group,
fill = vocab_group)) +
geom_bar(position = "dodge", stat = "identity") +
geom_linerange(aes(ymin = ci_lower, ymax = ci_upper),
position=position_dodge(width=0.9)) +
theme_classic()
ggplot(all_trigrams, aes(x = log_freq,
fill = vocab_group,
group = vocab_group)) +
geom_density(alpha = .4)
all_trigrams %>%
group_by(vocab_group) %>%
multi_boot_standard(col = "log_freq", na.rm = T) %>%
ggplot(aes(x = vocab_group, y = mean, group = vocab_group,
fill = vocab_group)) +
geom_bar(position = "dodge", stat = "identity") +
geom_linerange(aes(ymin = ci_lower, ymax = ci_upper),
position=position_dodge(width=0.9)) +
theme_classic()
Individual Kids
good_types_t1 <- target_types %>%
left_join(groups_info %>%
mutate(target_child_id = as.numeric(target_child_id))) %>%
filter(tbin == "t1")
get_trigrams_by_kid <- function(df, all_trigrams){
current_trigrams <- all_trigrams[w1 %in% df$gloss &
w2 %in% df$gloss &
w3 %in% df$gloss]
mean(current_trigrams$prob3, na.rm = T)
}
group_ids <- good_types_t1 %>%
distinct(target_child_id,
delta_resid_group)
trigram_freqs_by_kid <- good_types_t1 %>%
nest(-target_child_id, -tbin) %>%
mutate(mean_trigram_probability =
map(data, get_trigrams_by_kid,
coca_trigrams)) %>%
select(-data) %>%
unnest() %>%
left_join(groups_info %>% select(target_child_id, delta_resid_group))
ggplot(trigram_freqs_by_kid,
aes(x = mean_trigram_probability,
fill = delta_resid_group,
group = delta_resid_group)) +
geom_density(alpha = .4)
trigram_freqs_by_kid %>%
group_by(delta_resid_group) %>%
multi_boot_standard(col = "mean_trigram_probability",
na.rm = T) %>%
ggplot(aes(x = delta_resid_group, y = mean, group = delta_resid_group, fill = delta_resid_group)) +
geom_bar(position = "dodge", stat = "identity") +
geom_linerange(aes(ymin = ci_lower, ymax = ci_upper),
position=position_dodge(width=0.9)) +
theme_classic()
good_types_t1 <- target_types %>%
left_join(groups_info %>%
mutate(target_child_id = as.numeric(target_child_id))) %>%
filter(tbin == "t1")
get_trigrams_by_kid <- function(df, all_trigrams){
current_trigrams <- all_trigrams[w1 %in% df$gloss &
w2 %in% df$gloss &
w3 %in% df$gloss]
mean(log(current_trigrams$freq3), na.rm = T)
}
group_ids <- good_types_t1 %>%
distinct(target_child_id,
delta_resid_group)
trigram_freqs_by_kid <- good_types_t1 %>%
nest(-target_child_id, -tbin) %>%
mutate(log_mean_trigram_freq=
map(data, get_trigrams_by_kid,
coca_trigrams)) %>%
select(-data) %>%
unnest() %>%
left_join(groups_info %>% select(target_child_id, delta_resid_group))
ggplot(trigram_freqs_by_kid,
aes(x = log_mean_trigram_freq,
fill = delta_resid_group,
group = delta_resid_group)) +
geom_density(alpha = .4)
trigram_freqs_by_kid %>%
group_by(delta_resid_group) %>%
multi_boot_standard(col = "log_mean_trigram_freq",
na.rm = T) %>%
ggplot(aes( x = delta_resid_group, y = mean, group = delta_resid_group, fill = delta_resid_group)) +
geom_bar(position = "dodge", stat = "identity") +
geom_linerange(aes(ymin = ci_lower, ymax = ci_upper),
position=position_dodge(width=0.9)) +
theme_classic()