Measuring visual ability in linguistically diverse populations

[joshgilbert1994]

Article: https://link.springer.com/content/pdf/10.3758/s13428-024-02579-x.pdf
Data: https://figshare.com/articles/dataset/Data_for_Measuring_Visual_Ability_in_Linguistically_Diverse_Populations/24395098/2

CC BY 4.0

[KingArthur0205]

Study 1: 119 participants responded to the three novel object memory tests (NOMTs). Each participant "completed three NOMTs (with Ziggerins, then Greebles, then Sheinbugs), each including 72 trials(216 trials in total).

The 3 NOMTs are reflected in the item_family column:

1. Items 101-172: Ziggerins, item_family=1
2. Items 201-272: Greebles, item_family=2
3. Items 301-372: Sheinbugs, item_family=3.

Study 2: 393 participants responded to 2 NOMTs. The questions are merged and shuffled into one big survey

Note: In study 1, the response data are given as True/False. In study2, the responses are given with a finer granularity of 1-3.

[KingArthur0205]

@ben-domingue

I merged data from the 2 studies into one. However, many of the demographic data(covariates) of participants in study 2 are not given in study 1. This makes a lot of columns NA. Would this be a problem?

Data:
mva_hooper_2024.csv

Code:

# Paper: https://link.springer.com/content/pdf/10.3758/s13428-024-02579-x.pdf
# Data: https://figshare.com/articles/dataset/Data_for_Measuring_Visual_Ability_in_Linguistically_Diverse_Populations/24395098/2
library(haven)
library(dplyr)
library(tidyr)

study1_df <- read.csv("Part1_individual trials.csv")
study2_df <- read.csv("Part2_data.csv")

# ---------- Process Study1 Data ----------
study1_df <- study1_df |>
  rename(id=S)
study1_df <- pivot_longer(study1_df, cols=-id, names_to="item", values_to="resp")
study1_df$item_family <- with(study1_df, as.numeric(substr(gsub("X", "", item), 1, 1)))
study1_df$id <- paste0("study1_", study1_df$id)
study1_df$cov_sex <- NA
study1_df$cov_country_of_birth <- NA
study1_df$cov_current_country_of_residence <- NA
study1_df$cov_first_language <- NA
study1_df$cov_nationality <- NA
study1_df$cov_group <- NA
study1_df$cov_age <- NA

# ---------- Process Study2 Data ----------
study2_df <- study2_df |>
  rename(cov_group=GROUP, id=SS, cov_sex=Sex, cov_age=age, cov_country_of_birth=Country.of.Birth, cov_current_country_of_residence=Current.Country.of.Residence, cov_first_language=First.Language, cov_nationality=Nationality) |>
  select(starts_with("cov"), id, starts_with("Selection"))
study2_df$cov_sex <- ifelse(study2_df$cov_sex %in% c("Male", "Female"), study2_df$cov_sex, NA)
study2_df$cov_country_of_birth <- ifelse(study2_df$cov_country_of_birth %in% c("•", "DATE EXPIRED"), NA, study2_df$cov_country_of_birth)
study2_df$cov_current_country_of_residence <- ifelse(study2_df$cov_current_country_of_residence %in% c("•", "DATE EXPIRED"), NA, study2_df$cov_current_country_of_residence)
study2_df$cov_first_language <- ifelse(study2_df$cov_first_language %in% c("•", "DATE EXPIRED"), NA, study2_df$cov_first_language)
study2_df$cov_nationality <- ifelse(study2_df$cov_nationality %in% c("•", "DATE EXPIRED"), NA, study2_df$cov_nationality)
study2_df$id <- paste0("study2_", study2_df$id)

study2_df <- pivot_longer(study2_df, cols=-c(id, starts_with("cov")), names_to="item", values_to="resp")
study2_df$item_family <- NA

df <- rbind(study1_df, study2_df)

save(df, file="mva_hooper_2024.rdata")
write.csv(df, "mva_hooper_2024.csv", row.names=FALSE)

[ben-domingue]

@KingArthur0205 i need to look at this more but my initial reaction is that we might want to think about this from the 'trials' perspective, but i have to admit it's a weird edge case. more soon.