Accessing the Wordbank database
Mika Braginsky
2022-09-08
The wordbankr package allows you to access data in the
Wordbank database from
R. This vignette shows some examples of how to use the data
loading functions and what the resulting data look like.
There are three different data views that you can pull out of
Wordbank: by-administration, by-item, and administration-by-item.
Additionally, you can get metadata about the datasets and instruments
underlying the data. Advanced functionality let’s you get estimates of
words’ age of acquisition and word mappings across languages.
Administrations
The get_administration_data() function gives
by-administration information, either for a specific language and/or
form or for all instruments.
get_administration_data(language = "English (American)", form = "WS")
## # A tibble: 7,601 × 12
## data_id date_of_test age comprehension production is_norming dataset_name
## <dbl> <chr> <int> <int> <int> <lgl> <chr>
## 1 245518 1996-11-03 28 497 497 TRUE Marchman
## 2 245519 1996-10-14 22 369 369 TRUE Marchman
## 3 245520 1996-10-15 26 190 190 TRUE Marchman
## 4 245521 1996-11-01 27 264 264 TRUE Marchman
## 5 245522 1996-10-27 19 159 159 TRUE Marchman
## 6 245523 1996-10-16 30 513 513 TRUE Marchman
## 7 245524 1996-10-28 25 444 444 TRUE Marchman
## 8 245525 1996-11-04 24 582 582 TRUE Marchman
## 9 245526 1996-10-23 28 558 558 TRUE Marchman
## 10 245527 1991-10-13 18 7 7 TRUE Marchman
## # … with 7,591 more rows, and 5 more variables: dataset_origin_name <chr>,
## # language <chr>, form <chr>, form_type <chr>, child_id <int>
get_administration_data()
## # A tibble: 90,897 × 12
## data_id date_of_test age comprehension production is_norming dataset_name
## <dbl> <chr> <int> <int> <int> <lgl> <chr>
## 1 26372 <NA> 13 293 88 TRUE CLEX
## 2 26373 <NA> 16 122 12 TRUE CLEX
## 3 26374 <NA> 9 3 0 TRUE CLEX
## 4 26375 <NA> 12 0 0 TRUE CLEX
## 5 26376 <NA> 12 44 0 TRUE CLEX
## 6 26377 <NA> 8 14 5 TRUE CLEX
## 7 26378 <NA> 9 2 1 TRUE CLEX
## 8 26379 <NA> 10 44 1 TRUE CLEX
## 9 26380 <NA> 13 172 51 TRUE CLEX
## 10 26381 <NA> 16 241 68 TRUE CLEX
## # … with 90,887 more rows, and 5 more variables: dataset_origin_name <chr>,
## # language <chr>, form <chr>, form_type <chr>, child_id <int>
Items
The get_item_data() function gives by-item information,
either for a specific language and/or form or for all instruments.
get_item_data(language = "Italian", form = "WG")
## # A tibble: 505 × 11
## item_id language form form_type item_kind category item_definition
## <chr> <chr> <chr> <chr> <chr> <chr> <chr>
## 1 item_1 Italian WG WG first_signs <NA> Risponde quando è chia…
## 2 item_2 Italian WG WG first_signs <NA> Risponde ad un No
## 3 item_3 Italian WG WG first_signs <NA> Reagisce ad un C'è la …
## 4 item_4 Italian WG WG phrases <NA> Vuoi la pappa
## 5 item_5 Italian WG WG phrases <NA> Hai sonno? Sei stanco
## 6 item_6 Italian WG WG phrases <NA> Vuoi bere?
## 7 item_7 Italian WG WG phrases <NA> Stai attento
## 8 item_8 Italian WG WG phrases <NA> Stai buono
## 9 item_9 Italian WG WG phrases <NA> Batti le manine
## 10 item_10 Italian WG WG phrases <NA> Cambiamo il pannolino
## # … with 495 more rows, and 4 more variables: english_gloss <chr>,
## # uni_lemma <chr>, lexical_category <chr>, complexity_category <chr>
## # A tibble: 40,959 × 11
## item_id language form form_type item_kind category item_definition
## <chr> <chr> <chr> <chr> <chr> <chr> <chr>
## 1 item_1 British Sign Lang… WG WG phrases <NA> be careful
## 2 item_2 British Sign Lang… WG WG phrases <NA> bring me
## 3 item_3 British Sign Lang… WG WG phrases <NA> change nappy
## 4 item_4 British Sign Lang… WG WG phrases <NA> come here
## 5 item_5 British Sign Lang… WG WG phrases <NA> daddy/mummy ho…
## 6 item_6 British Sign Lang… WG WG phrases <NA> donttouch
## 7 item_7 British Sign Lang… WG WG phrases <NA> finish
## 8 item_8 British Sign Lang… WG WG phrases <NA> get up
## 9 item_9 British Sign Lang… WG WG phrases <NA> give me hug
## 10 item_10 British Sign Lang… WG WG phrases <NA> give me kiss
## # … with 40,949 more rows, and 4 more variables: english_gloss <chr>,
## # uni_lemma <chr>, lexical_category <chr>, complexity_category <chr>
Administrations x Items
If you are only looking at total vocabulary size, admins
is all you need, since it has both productive and receptive vocabulary
sizes calculated. If you are looking at specific items or subsets of
items, you need to load instrument data, using the
get_instrument_data() function. Pass it an instrument
language and form, along with a list of items you want to extract (by
item_id).
get_instrument_data(
language = "English (American)",
form = "WS",
items = c("item_26", "item_46")
)
## # A tibble: 17,098 × 5
## data_id item_id value produces understands
## <dbl> <chr> <chr> <lgl> <lgl>
## 1 245518 item_26 "produces" TRUE NA
## 2 245519 item_26 "produces" TRUE NA
## 3 245520 item_26 "produces" TRUE NA
## 4 245521 item_26 "produces" TRUE NA
## 5 245522 item_26 "" FALSE NA
## 6 245523 item_26 "produces" TRUE NA
## 7 245524 item_26 "produces" TRUE NA
## 8 245525 item_26 "produces" TRUE NA
## 9 245526 item_26 "produces" TRUE NA
## 10 245527 item_26 "" FALSE NA
## # … with 17,088 more rows
By default get_instrument_table() returns a data frame
with columns of the administration’s data_id, the item’s
num_item_id (numerical item_id), and the
corresponding value. To include administration information, you can set
the administrations argument to TRUE, or pass
the result of get_administration_data() as
administrations (that way you can prevent the
administration data from being loaded multiple times). Similarly, you
can set the iteminfo argument to TRUE, or pass
it result of get_item_data().
Loading the data is fast if you need only a handful of items, but the
time scales about linearly with the number of items, and can get quite
slow if you need many or all of them. So, it’s a good idea to filter
down to only the items you need before calling
get_instrument_data().
As an example, let’s say we want to look at the production of animal
words on English Words & Sentences over age. First we get the items
we want:
animals <- get_item_data(language = "English (American)", form = "WS") %>%
filter(category == "animals")
Then we get the instrument data for those items:
animal_data <- get_instrument_data(language = "English (American)",
form = "WS",
items = animals$item_id,
administration_info = TRUE,
item_info = TRUE)
Finally, we calculate how many animals words each child produces and
the median number of animals of each age bin:
animal_summary <- animal_data %>%
group_by(age, data_id) %>%
summarise(num_animals = sum(produces, na.rm = TRUE)) %>%
group_by(age) %>%
summarise(median_num_animals = median(num_animals, na.rm = TRUE))
ggplot(animal_summary, aes(x = age, y = median_num_animals)) +
geom_point() +
labs(x = "Age (months)", y = "Median animal words producing")
Advanced functionality: Age of acquisition
The fit_aoa() function computes estimates of items’ age
of acquisition (AoA). It needs to be provided with a data frame returned
by get_instrument_data() – one row per administration x
item combination, and minimally the columns age and
num_item_id. It returns a data frame with one row per item
and an aoa column with the estimate, preserving and
item-level columns in the input data. The AoA is estimated by computing
the proportion of administrations for which the child
understands/produces (measure) each word, smoothing the
proportion using method, and taking the age at which the
smoothed value is greater than proportion.
## # A tibble: 43 × 8
## aoa item_id item_kind item_definition category lexical_category uni_lemma
## <dbl> <chr> <chr> <chr> <chr> <chr> <chr>
## 1 26 item_13 word alligator animals nouns alligator
## 2 25 item_14 word animal animals nouns animal
## 3 26 item_15 word ant animals nouns ant
## 4 20 item_16 word bear animals nouns bear
## 5 22 item_17 word bee animals nouns bee
## 6 18 item_18 word bird animals nouns bird
## 7 23 item_19 word bug animals nouns bug
## 8 22 item_20 word bunny animals nouns bunny
## 9 24 item_21 word butterfly animals nouns butterfly
## 10 19 item_22 word cat animals nouns cat
## # … with 33 more rows, and 1 more variable: complexity_category <chr>
fit_aoa(animal_data, method = "glmrob", proportion = 1/3)
## # A tibble: 43 × 8
## aoa item_id item_kind item_definition category lexical_category uni_lemma
## <dbl> <chr> <chr> <chr> <chr> <chr> <chr>
## 1 24 item_13 word alligator animals nouns alligator
## 2 23 item_14 word animal animals nouns animal
## 3 24 item_15 word ant animals nouns ant
## 4 18 item_16 word bear animals nouns bear
## 5 19 item_17 word bee animals nouns bee
## 6 NA item_18 word bird animals nouns bird
## 7 20 item_19 word bug animals nouns bug
## 8 19 item_20 word bunny animals nouns bunny
## 9 22 item_21 word butterfly animals nouns butterfly
## 10 NA item_22 word cat animals nouns cat
## # … with 33 more rows, and 1 more variable: complexity_category <chr>
Advanced functionality: Cross-linguistic data
One of the item-level fields is uni_lemma (“universal
lemma”), which is intended to be an approximate semantic mapping between
words across the languages in Wordbank. The function
get_crossling_items() simply gives all the available
uni_lemma values.
## # A tibble: 2,552 × 2
## id uni_lemma
## <int> <chr>
## 1 1739 (hair)brush
## 2 1552 (play)pen
## 3 1494 (sheep)
## 4 1783 (to be in) pain
## 5 1777 (to be) hungry
## 6 1775 (to be) thirsty
## 7 1769 (to have) breakfast
## 8 1272 [possessive]
## 9 1593 [to splash in the water?]
## 10 1951 1PL
## # … with 2,542 more rows
The function get_crossling_data() takes a vector of
uni_lemmas and returns a data frame of summary statistics
for each item mapped to that uni_lemma in any language (on
WG forms). Each row is combination of item and age, and the
columns indicate the number of children (n_children), means
(comprehension, production), standard
deviations (comprehension_sd, production_sd),
and item-level fields.
get_crossling_data(uni_lemmas = c("hat", "nose")) %>%
select(language, uni_lemma, item_definition, age, n_children, comprehension,
production, comprehension_sd, production_sd) %>%
arrange(uni_lemma)