R/clean_economic_status.R

In STAPM/hseclean: Health Survey Data Wrangling

#' Economic status
#'
#' Process the data on employment and occupation.
#'
#' First apply the function [hseclean::clean_age()] to the data because this function uses the age variable.
#'
#' The issues around using occupation-based social classifications for social survey research are discussed by
#' \insertCite{Connelly2016a;textual}{hseclean}.
#' They advise using a range of alternative measures, but not creating new measures beyond what is already established.
#'
#' The classifications considered are:
#'
#' \itemize{
#' \item Employed / in paid work or not.
#' \item The \href{https://www.ons.gov.uk/methodology/classificationsandstandards/otherclassifications/thenationalstatisticssocioeconomicclassificationnssecrebasedonsoc2010}{NS-SEC measure}, which was constructed to measure the employment relations and conditions of occupations (i.e. it classifies people based on their employment occupation). It is
#' therefore not that good at classifying people who are not employed for various reasons.
#' \item The \href{https://humanrights.brightblue.org.uk/blog-1/2016/5/6/measuring-social-class}{NRS social grade system}. This measure is the one used in the
#' Tobacco and Alcohol Toolkit studies, but is not reported in the Health Survey for England. We create this variable by recategorising the NS-SEC 8 level variable.
#' This is important to facilitate the link of analysis to the Toolkit Study.
#' \item Manual vs. non-manual occupation. In the 2017 Tobacco control plan for England, there was a specific target to reduce the difference in rates of smoking
#' between people classified with a manual or non-manual occupation. We create this variable from the 3 level NS-SEC classification
#' by grouping Managerial and professional with intermediate occupations to give the non-manual group.
#' \item Economic status - retired / employed / unemployed.
#' \item Activity status for last week that adds more detail such as 'in education' and 'looking after home or family'.
#' }
#'
#' @param data Data table - the Health Survey for England dataset.
#' @importFrom data.table :=
#' @return
#' \itemize{
#' \item employ2cat: "employed", "unemployed" (in paid work or not)
#' \item nssec3_lab: "Managprof", "Intermediate", "Routinemanual", "Other".
#' Social class/occupation based on the NS-SEC system. A four level variable for
#' managerial and professional occupations, intermediate occupations and routine and manual occupations. The fourth level is
#' 'other' that includes anyone (young/old/unemployed) who do not have an occupation.
#' \item social_grade: "ABC1", "C2DE". 2 level social grade classification.
#' \item man_nonman: "Manual", "Non-manual", "Other"
#' \item activity_lstweek: "Managprof" (employed), "Intermediate" (employed), "Routinemanual" (employed), "Unemployed", "In education", "Looking after home or family", "Sickness, injury or disability", "Retired"
#' }
#'
#' @export
#'
#' @references
#' \insertRef{Connelly2016a}{hseclean}
#'
#' @examples
#'
#' \dontrun{
#'
#' data <- read_2016(root = "/Volumes/Shared/")
#'
#' data <- clean_economic_status(data = data)
#'
#' }
#'
clean_economic_status <- function(
    data
) {

  if(!("age" %in% colnames(data))) {
    warning("First apply the function [hseclean::clean_age()] to the data because this function uses the age variable.")
  }

  country <- unique(data[ , country][1])
  year <- unique(data[ , year][1])

  if(!("econact" %in% colnames(data))) {
    data[ , econact := NA_real_]
  }
  if(!("activb" %in% colnames(data))) {
    data[ , activb := NA_real_]
  }
  if(!("econac12" %in% colnames(data))) {
    data[ , econac12 := NA_real_]
  }
  if(!("nactiv" %in% colnames(data))) {
    data[ , nactiv := NA_real_]
  }
  if(!("econstat" %in% colnames(data))) {
    data[ , econstat := NA_real_]
  }

  ####################################################################
  # NS-SEC

  if("nssec3" %in% colnames(data)){
    # Label 3 level variable
    data[nssec3 == 1, nssec3_lab := "managprof"]
    data[nssec3 == 2 , nssec3_lab := "intermediate"]
    data[nssec3 == 3 , nssec3_lab := "routinemanual"]
    data[nssec3 == 99 , nssec3_lab := "other"]
  }

  ####################################################################
  # Employment status

  # For years after 2010, base initially on whether in paid work
  if(country == "England" & max(data[ , year]) >= 2010) {

    data[paidwk == 1, employ2cat := "employed"]
    data[paidwk == 2, employ2cat := "unemployed"]

    #data[ , paidwk := NULL]

  }

  data[year < 2010 | country %in% c("Wales","Scotland"), employ2cat := NA_character_]
  data[country == "England" & year >= 2015, econact := NA_integer_]

  # Fill missing with info from other variables that indicate employment or not
  data[is.na(employ2cat) & (econstat %in% 2:5 | econact == 1 | activb == 2 | econac12 == 2 | nactiv == 2), employ2cat := "employed"]

  data[country == "England" & year < 2015 & is.na(employ2cat) & (econact %in% 2:4 | activb %in% c(1, 3:10)), employ2cat := "unemployed"]
  data[country == "England" & year >= 2015 & is.na(employ2cat) & (econact %in% 2:4 | activb %in% c(1, 3:9)), employ2cat := "unemployed"]

  data[country == "Scotland" & (econac12 %in% c(1, 3:7) | nactiv %in% c(1, 3:11)), employ2cat := "unemployed"]

  data[country == "Wales" & year > 2015 & (econstat %in% c(1, 6:11)), employ2cat := "unemployed"]

  # WHS
  if(country == "Wales" & year <= 2015) {

    data[country == "Wales" & year >= 2009 & year <= 2015 & nssec3 == 8, employ2cat := "unemployed"]
    data[is.na(employ2cat) & country == "Wales" & year >= 2009 & year <= 2015 & employ == 1, employ2cat := "employed"]
    data[is.na(employ2cat) & country == "Wales" & year >= 2009 & year <= 2015 & employ == 2, employ2cat := "unemployed"]
    data[is.na(employ2cat) & country == "Wales" & year >= 2009 & year <= 2015 & ecstat == 1, employ2cat := "employed"]
    data[is.na(employ2cat) & country == "Wales" & year >= 2009 & year <= 2015 & ecstat == 2, employ2cat := "unemployed"]

  }

  #data[country == "Wales" & year >= 2009 & year <= 2015, `:=`(ecstat = NULL, employ = NULL, nssec3 = NULL)]

  # Fill missing for children with 'unemployed'
  data[is.na(employ2cat) & age < 16, employ2cat := "unemployed"]


  ####################################################################
  # Activity status for last week combined with economic activity variable

  if(!(country == "Wales" & year <= 2015)) { # i.e. not WHS

    data[year < 2015  & (econstat == 10 | econac12 == 6 | activb == 10), activity_lstweek := "home_or_family"]
    data[year >= 2015 & (econstat == 10 | econac12 == 6 | activb == 9), activity_lstweek := "home_or_family"]

    data[year < 2015  & (econstat == 1 | econac12 == 1 |activb %in% c(1, 3)), activity_lstweek := "education"]
    data[year >= 2015 & (econstat == 1 | econac12 == 1  |activb == 1), activity_lstweek := "education"]

    data[year < 2015  & (econstat == 9 | econac12 == 5 | econact == 3 | activb == 9), activity_lstweek := "retired"]
    data[year >= 2015 & (econstat == 9 | econac12 == 5 | econact == 3 | activb == 8), activity_lstweek := "retired"]

    data[year < 2015  & (econstat == 6 | econac12 == 4 | econact == 2 | activb %in% c(5, 6)), activity_lstweek := "unemployed"]
    data[year >= 2015 & (econstat == 6 | econac12 == 4 | econact == 2 | activb %in% c(4, 5)), activity_lstweek := "unemployed"]

    data[year < 2015  & (econstat %in% 7:8 | econac12 == 3 | activb %in% 7:8), activity_lstweek := "sick_ill_disab"]
    data[year >= 2015 & (econstat %in% 7:8 | econac12 == 3 | activb %in% 6:7), activity_lstweek := "sick_ill_disab"]

    data[year < 2015  & (econstat %in% 2:5 | econac12 == 2 | econact == 1 | activb %in% c(2, 4)), activity_lstweek := "employed"]
    data[year >= 2015 & (econstat %in% 2:5 | econac12 == 2 | econact == 1 | activb %in% c(2, 3)), activity_lstweek := "employed"]

  }

  # WHS
  if(country == "Wales" & year <= 2015) {

    data[work == 10, activity_lstweek := "home_or_family"]
    data[work == 1, activity_lstweek := "education"]
    data[work == 9, activity_lstweek := "retired"]
    data[work %in% 7:8, activity_lstweek := "sick_ill_disab"]
    data[is.na(activity_lstweek) & employ2cat == "unemployed", activity_lstweek := "unemployed"]
    data[is.na(activity_lstweek) & employ2cat == "employed", activity_lstweek := "employed"]
  }

  # Fill missing using information on age
  data[is.na(activity_lstweek) & age < 18, activity_lstweek := "education"]
  data[is.na(activity_lstweek) & age >= 18 & age < 65 & (employ2cat == "unemployed" | is.na(employ2cat)), activity_lstweek := "unemployed"]
  data[is.na(activity_lstweek) & age >= 65, activity_lstweek := "retired"]


  ####################################################################
  # Fixes to flll some missing values

  if("nssec3" %in% colnames(data) ){
    data[is.na(nssec3_lab) & activity_lstweek == "home_or_family", nssec3_lab := "other"]
    data[is.na(nssec3_lab) & activity_lstweek == "sick_ill_disab", nssec3_lab := "other"]
    data[is.na(nssec3_lab) & activity_lstweek == "education", nssec3_lab := "not applicable"]
    data[is.na(nssec3_lab) & age < 18, nssec3_lab := "not applicable"]

    data[is.na(employ2cat) & nssec3_lab %in% c("routinemanual", "intermediate", "managprof"), employ2cat := "employed"]
    data[is.na(employ2cat) & activity_lstweek %in% c("education", "unemployed", "home_or_family", "retired", "sick_ill_disab"), employ2cat := "unemployed"]
    data[is.na(activity_lstweek) & employ2cat == "employed", activity_lstweek := "employed"]
  }

  ####################################################################
  # Manual vs. non-manual occupation

  if("nssec3" %in% colnames(data) ){
    data[ , man_nonman := nssec3_lab]
    data[nssec3_lab %in% c("managprof", "intermediate"), man_nonman := "nonmanual"]
    data[nssec3_lab == "routinemanual", man_nonman := "manual"]
  }

  ####################################################################
  # NRS social grade

  # Match nssec8 to the NRS social grade classification used in the Toolkit study
  if("nssec8" %in% colnames(data)){
    data[nssec8 %in% 1:3, social_grade := "ABC1"]
    data[nssec8 %in% 4:99 , social_grade := "C2DE"]


    # Fill some missing data
    data[is.na(social_grade) & activity_lstweek == "home_or_family", social_grade := "other"]
    data[is.na(social_grade) & activity_lstweek == "sick_ill_disab", social_grade := "other"]
    data[is.na(social_grade) & activity_lstweek == "education", social_grade := "not applicable"]
    data[is.na(social_grade) & age < 18, social_grade := "not applicable"]
    data[nssec3_lab == "other", social_grade := "other"]
  }

  # Remove variables not needed
  #data[ , `:=`(econact = NULL, activb = NULL, econac12 = NULL, nssec8 = NULL, nssec3 = NULL)]


  return(data[])
}