Disparities in the ICT Skills: Lessons from MICS6 data

Questions in MICS6 related to the ICT skills

library(tidyverse)
library(haven)
library(sjPlot)
library(srvyr)
library(sjmisc)
library(knitr)
library(kableExtra)
library(DT)

MICS6 = readRDS(file = "mics6.rds")

MICS6 =
  MICS6 %>%
  mutate(Age_group = case_when(WB4 >=15 & WB4 <= 18 ~ "15-18",
                               TRUE ~ "else"))

The variables selected for the analysis include data on ICT skills from the female and male questionnaire, as well as variables on male/female weight, level of education, disability status, area of residence, and age group.

These data are complemented by the variables on possession of internet and PC at home thta comes from the household dataset.

All in all, only those 14 countries where ICT data were collected for female and male respondents were selected:

unique(MICS6$Country)

##  [1] "DR Congo"          "Ghana"             "Kiribati"         
##  [4] "Lao PDR"           "Lesotho"           "Madagascar"       
##  [7] "Mongolia"          "Pakistan (Punjab)" "Suriname"         
## [10] "The Gambia"        "Sierra Leone"      "Togo"             
## [13] "Tunisia"           "Zimbabwe"

All those files were merged to produce the analysis. Let’s start by exploring the set of questions that measures ICT skills in MICS by also looking at the basic patterns and descriptive statistics.

As the analysis of the ICT skills on the merged data on 14 countries suggests, those skills that require basic manipulations with documents or folders or file transfering, have the highest prevelence that accounts at least 50%. However, as we shift to more comprehend routines that require use of internet, or installation of devices, or calculus functions, the numbers gradually drop. Not surprisingly, writing of a computer program has a lower percentage which is equal to 10%. The numbers are not global estimates but rather summary of distribution on the set of countries with the available MICS6 data.

Therefore, the results of measurement by single variables are not as interesting as aggregation of the indicators on the ICT skills by countries is a major focus.

By sex

Let’s also produce visualization of ICT by gender with the confidence levels for our estimates.

As the data table and and graph show, there are wide disparities in the ICT skills across different countries by gender. Needs to be noticed, the parity ratio does not fall within the interval between 0.97 and 1.03, which means that for none of the 14 countries parity in ICT skills is achieved. Countries like DR Congo, Ghana, and Togo constitute top 3 of those where differences in ict skills by gender are particularly sharp.

Moreover, the disparities become wider when we shift to the group of children 15-18.

By education

ict_educ =   
  MICS6_srvyr %>%
  mutate(Education = case_when(welevel ==0 ~ "Pre-primary or none",
                               welevel ==1 ~ "Primary",
                               welevel ==2 ~ "JSS/JHS/Middle",
                               welevel ==3 ~ "SSS/SHS/ Secondary",
                               welevel ==4 ~ "Higher",
                               TRUE ~ as.character(NA))) %>%
  group_by(Country, Gender, Education) %>%
  summarise(ICT = survey_mean(ICT_combined, na.rm = T, vartype = "se")) %>%
  mutate(ICT = round(ICT*100, 1)) %>%
  select(-ICT_se) %>%  
  spread(Education, ICT) %>%
  mutate(`Disparity (Pre-primary or none/Higher)` = round(`Pre-primary or none`/`Higher`, 2)) %>%
  select(Country, Gender, `Pre-primary or none`, Primary, `JSS/JHS/Middle`, `SSS/SHS/ Secondary`, Higher, 
         `Disparity (Pre-primary or none/Higher)`)

datatable(ict_educ, extensions = 'Buttons', 
          options = list(
          dom = 'Bfrtip',
          pageLength = 15, #lengthMenu = c(5, 10, 14),
          buttons = c('copy', 'csv', 'excel', 'pdf', 'print')))

By wealth

ict_wealth =   
  MICS6_srvyr %>%
  group_by(Country, Gender, Wealth, Age_group) %>%
  cascade(ICT = survey_mean(ICT_combined, na.rm = T, vartype = "se"), .fill = "Total") %>%
  mutate(ICT = round(ICT*100, 1)) %>%
  select(-ICT_se) %>%
  filter(!Age_group %in% c("else")) %>%
  spread(Wealth, ICT) %>%
  select(-Total) %>%
  mutate(`Disparity (Poorest/Richest)` = round(Poorest/Richest, 2)) %>%
  select(Country, Gender, Age_group, Poorest, Poor, Middle, Rich, Richest, `Disparity (Poorest/Richest)`) %>%
  drop_na()

datatable(ict_wealth, extensions = 'Buttons', 
          options = list(
          dom = 'Bfrtip',
          pageLength = 15, #lengthMenu = c(5, 10, 14),
          buttons = c('copy', 'csv', 'excel', 'pdf', 'print')))

By area

ict_area =   
  MICS6_srvyr %>%
  group_by(Country, Gender, Area) %>%
  summarise(ICT = survey_mean(ICT_combined, na.rm = T, vartype = "se")) %>%
  mutate(ICT = round(ICT*100, 1)) %>%
  select(-ICT_se) %>%
  spread(Area, ICT) %>%
  mutate(`Disparity (Rural/Urban)` = round(Rural/Urban, 2)) 

datatable(ict_area, extensions = 'Buttons', 
          options = list(
          dom = 'Bfrtip',
          pageLength = 14, lengthMenu = c(5, 10, 14),
          buttons = c('copy', 'csv', 'excel', 'pdf', 'print')))

By computer possession

ict_comp =   
  MICS6_srvyr %>%
  group_by(Country, Gender,  Computer, Age_group) %>%
  cascade(ICT = survey_mean(ICT_combined, na.rm = T, vartype = "se"), .fill = "Total") %>%
  mutate(ICT = round(ICT*100, 1)) %>%
  select(-ICT_se) %>%  
  drop_na(Computer) %>%
  filter(!Age_group %in% c("else") & Gender != "Total") %>%
  spread(Computer, ICT) %>%
  select(-Total) %>%
  mutate(`Disparity (No PC/Have PC)` = round(`No computer at home`/`Has computer at home`, 2)) 

datatable(ict_comp, extensions = 'Buttons', 
          options = list(
          dom = 'Bfrtip',
          pageLength = 14, lengthMenu = c(5, 10, 14),
          buttons = c('copy', 'csv', 'excel', 'pdf', 'print')))

By disability

ict_dis =   
  MICS6_srvyr %>%
  mutate(disability = case_when(disability == 1 ~ "Has functional difficulty",
                                disability == 2 ~ "Has no functional difficulty")) %>%
  group_by(Country, disability) %>%
  summarise(ICT = survey_mean(ICT_combined, na.rm = T, vartype = "se")) %>%
  mutate(ICT = round(ICT*100, 1)) %>%
  drop_na(disability) %>%
  select(-ICT_se) %>%
  spread(disability, ICT) %>%
  mutate(`Disparity (Disable/Non-disable)` = round(`Has functional difficulty`/`Has no functional difficulty`, 2)) 

datatable(ict_dis, extensions = 'Buttons', 
          options = list(
          dom = 'Bfrtip',
          pageLength = 14, lengthMenu = c(5, 10, 14),
          buttons = c('copy', 'csv', 'excel', 'pdf', 'print')))