Read in language and behavioral effect sizes.

LANG_ES_PATH <- here("analyses/study1c/parameter_exploration/es_params.csv")
lang_es <- read_csv(LANG_ES_PATH,
                    col_names = c("domain", "x",
                                  "lang_effect_size", "model",
                                  "min_word_count", "vector_size", "ngrams", "window_size")) %>%
  select(-x) 
tidy_lang_es <- lang_es %>% 
  filter(model == "coca") %>%
  select(-model) %>%
  full_join(lang_es %>% filter(model == "bnc") %>% select(-model), 
            by = c("domain", "min_word_count",
                   "vector_size", "ngrams", "window_size")) %>%
  rename(coca_lang_es = lang_effect_size.x,
         bnc_lang_es = lang_effect_size.y) %>%
  mutate(lang_diff = bnc_lang_es - coca_lang_es) %>%
  select(domain, min_word_count, vector_size, ngrams, window_size, everything(), lang_diff) 

BEHAVIORAL_PATH <- here("data/study1c/processed/tidy_behavioral_iat_data.csv")
tidy_behavioral_es <- read_csv(BEHAVIORAL_PATH) %>%
  rename(uk_behavior_es = uk,
         us_behavior_es = us)

Get correlations.

full_df <- tidy_lang_es %>%
  left_join(tidy_behavioral_es)

get_correlations <- function(current_df){
    cor1 <- tidy(cor.test(current_df$coca_lang_es, current_df$us_behavior_es)) %>%
                   mutate(corr = "coca_us")
    cor2 <- tidy(cor.test(current_df$coca_lang_es, current_df$uk_behavior_es)) %>%
                    mutate(corr = "coca_uk")
    cor3 <- tidy(cor.test(current_df$bnc_lang_es, current_df$us_behavior_es)) %>%
                    mutate(corr =  "bnc_us")
    cor4 <- tidy(cor.test(current_df$bnc_lang_es, current_df$uk_behavior_es)) %>%
                    mutate(corr = "bnc_uk")
    cor5 <- tidy(cor.test(current_df$behavioral_resid_diff, current_df$lang_diff)) %>%
                    mutate(corr = "diff_corr")
    
    bind_rows(list(cor1, cor2, cor3, cor4, cor5)) %>%
      select(corr, estimate, statistic, p.value,
             parameter, conf.low, conf.high) %>%
      list()
}

all_corrs <- full_df %>%
  group_by(vector_size, ngrams, window_size) %>%
  slice(1:31) %>%
  nest() %>% 
  mutate(model = map(data, ~get_correlations(.))) %>%
  select(-data) %>%
  unnest() %>%
  unnest()

Plot correlations.

200-500 (right) refer to vector size. 2 - 20 (top) refer to window size.

all_corrs %>%
  mutate(corr = fct_relevel(corr, c("coca_us", "bnc_uk"))) %>%
  filter(vector_size < 500,
         window_size < 20) %>%
  ggplot(aes(x = corr, y = estimate, color = ngrams, group = interaction(ngrams, window_size, vector_size))) +
  ylab("correlation coefficient") +
  geom_line() +
  geom_hline(aes(yintercept = 0), linetype = 2) +
  geom_pointrange(aes(ymin = conf.low, ymax = conf.high), size = .4) +
  facet_grid(ngrams + vector_size ~ window_size) +
  theme_classic() +
  theme(axis.text.x = element_text(angle = 90, hjust = 1))