data-raw/0-prep-understanding-society.R
In philmikejones/thesis: hrsim: A spatial microsimulation analysis of health inequalities and health resilience
##
## Warning!
## Requires ~8GB RAM to run the script
##
if (!dir.exists("inst/extdata/restricted_us")) {
if (Sys.info()["sysname"] != "Linux") {
warning("It looks like you're not using a Linux system\n
I cannot automatically unzip and place the
Understanding Society data files in the correct place.\n
Manually copy the *.tab files in UKDA-6614-tab/tab
to inst/extdata/restricted_us/\n
You can also move the documentation to
inst/doc/restricted_us [optional]")
} else {
dir.create("inst/extdata/restricted_us/", showWarnings = FALSE)
# Looks like filename is unique; use 6614TAB_ to obtain the filename
us_zip <- list.files("inst/extdata/", pattern = "6614TAB_",
full.names = TRUE)
unzip(us_zip, exdir = "inst/extdata/restricted_us/")
# Move documentation
dir.create("inst/doc/restricted_us",
recursive = TRUE, showWarnings = FALSE)
system2("mv", args = c("inst/extdata/restricted_us/UKDA-6614-tab/mrdoc",
"inst/doc/restricted_us"))
# Move data files
system2("mv",
args = c("inst/extdata/restricted_us/UKDA-6614-tab/tab/*.tab",
"inst/extdata/restricted_us/"))
system2("rm", c("-rf", "inst/extdata/restricted_us/UKDA-6614-tab/tab/"))
# Move README and file information
system2("mv", c("inst/extdata/restricted_us/UKDA-6614-tab/*.*",
"inst/doc/restricted_us/"))
system2("rm", c("-rf", "inst/extdata/restricted_us/UKDA-6614-tab/"))
}
rm(us_zip)
}
if (!file.exists("data/us.RData")) {
# Prepare wave f ====
usf <- readr::read_tsv("inst/extdata/restricted_us/f_indresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")
# -11 only available for IEMB
# -10 not available for IEMB
# -9 missing
# -8 Inapplicable
# -7 Proxy
# -2 Refused
# -1 Don't know
) %>%
select(pidp, f_hidp, # id and household id
f_hhorig, # sample origin (i.e. BHPS)
f_dvage, # derived age
f_sex, # sex corrected
f_racel, # ethnic group
f_hiqual_dv, # highest qualification
f_jbstat, # current economic activity
f_health, # long-standing illness or disability
f_scsf1, # self-reported general health
f_jbnssec_dv, # NS-SEC/social class
f_jbnssec8_dv, # NS-SEC 8
f_jbnssec5_dv, # NS-SEC 5
f_jbnssec3_dv, # NS-SEC 3
f_mlstat, # marital status
f_scopngbhe, # willing to improve neighbourhood
f_scopngbhg, # similar to others in neighbourhood
f_nbrcoh3, # people in neighbourhood can be trusted
f_scopngbha, # belong to neighbourhood
f_orga1, # active in organisation
f_orga2, # active in organisation
f_orga3, # active in organisation
f_orga4, # active in organisation
f_orga5, # active in organisation
f_orga6, # active in organisation
f_orga7, # active in organisation
f_orga8, # active in organisation
f_orga9, # active in organisation
f_orga10, # active in organisation
f_orga11, # active in organisation
f_orga12, # active in organisation
f_orga13, # active in organisation
f_orga14, # active in organisation
f_orga15, # active in organisation
f_orga16, # active in organisation
f_nbrcoh4, # people in neighbourhood don't get along
f_crdark, # feel safe walking alone at night
f_siminc, # Prop friends with similar income
f_simrace, # Friends similar ethnicity
f_poleff4, # don't have say in what g'ment does
f_finnow, # subjective financial status
f_closenum, # number of close friends
f_scghqa, # GHQ: concentration
f_scghqb, # GHQ: loss of sleep
f_scghqc, # GHQ: playing a useful role
f_scghqd, # GHQ: capable of making decisions
f_scghqe, # GHQ: constantly under strain
f_scghqf, # GHQ: problem overcoming difficulties
f_scghqg, # GHQ: enjoy day-to-day activities
f_scghqh, # GHQ: ability to face problems
f_scghqi, # GHQ: unhappy or depressed
f_scghqj, # GHQ: losing confidence
f_scghqk, # GHQ: believe worthless
f_scghql, # GHQ: general happiness
f_ncigs, # Cigs smoked
f_netlv_1, # distance best friend lives
f_netlv_2, # distance best friend lives
f_netlv_3, # distance best friend lives
f_netph_1, # best friend often in touch
f_netph_2, # best friend often in touch
f_netph_3, # best friend often in touch
f_save, # saves money
# received core benefit:
f_benbase1, # 1 == income support
f_benbase2, # 2 == job-seekers allowance
# other benefit:
f_othben5, # 5 == working tax credit;
f_othben8, # 8 == housing benefit
f_benctc, # received child tax credit
f_bendis2, # disability benefits (ESA == 2)
f_hubuys, # who does food shopping? (for cohabiting couples)
f_hcondno1,
f_hcondno2,
f_hcondno3,
f_hcondno4,
f_hcondno5,
f_hcondno6,
f_hcondno7,
f_hcondno8,
f_hcond17 # clinical depression
# f_btype6, # tax credits
# f_disdif, # Type of impairment or disability
# f_hcondno1, # Diagnosed health condition(s)
# f_hcondno2, f_hcondno3, f_hcondno4, f_hcondno5, f_hcondno6,
# f_hcondno7,
# f_hcondns1, # still has condition
# f_hcondns2, f_hcondns3, f_hcondns4, f_hcondns5, f_hcondns6,
# f_hcondns7,
# f_scsf2a, # health limits moderate activities
# f_scsf2b, # health limits several flights of stairs
# f_scsf3a, # health limits work (last 4 wks)
# f_scsf3b, # health limits kind of work (last 4 wks)
# f_scsf4a, # mental health accomp. less (last 4 wks)
# f_scsf4b, # mental health worked less carefully (last 4 wks)
# f_scsf5, # pain interfered w/ work (last 4 wks)
# f_scsf6a, # felt calm & peaceful (last 4 wks)
# f_scsf6b, # Had lots of energy (last 4 wks)
# f_scsf6c, # Felt downhearted/depressed (last 4 wks)
# f_scsf7, # health (p or m) interfered w/ social life (4 wks)
# f_sclfsat1, # Satisfaction with health
# f_sclfsato, # Satisfaction with life overall
)
# Prepare wave e ====
use <- readr::read_tsv("inst/extdata/restricted_us/e_indresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(pidp, e_hidp, # id and household id
e_hhorig, # sample origin (i.e. BHPS)
e_dvage, # derived age
e_sex, # sex corrected
e_racel, # ethnic group
e_hiqual_dv, # highest qualification
e_jbstat, # current economic activity
e_health, # long-standing illness or disability
e_scsf1, # self-reported general health
e_jbnssec_dv, # NS-SEC/social class
e_jbnssec8_dv, # NS-SEC 8
e_jbnssec5_dv, # NS-SEC 5
e_jbnssec3_dv, # NS-SEC 3
e_mlstat, # marital status
e_finnow, # subjective financial status
e_scghqa, # GHQ: concentration
e_scghqb, # GHQ: loss of sleep
e_scghqc, # GHQ: playing a useful role
e_scghqd, # GHQ: capable of making decisions
e_scghqe, # GHQ: constantly under strain
e_scghqf, # GHQ: problem overcoming difficulties
e_scghqg, # GHQ: enjoy day-to-day activities
e_scghqh, # GHQ: ability to face problems
e_scghqi, # GHQ: unhappy or depressed
e_scghqj, # GHQ: losing confidence
e_scghqk, # GHQ: believe worthless
e_scghql, # GHQ: general happiness
e_ncigs, # cigs smoked per day
e_benunemp1, # job-seekers allowance
e_btype2, # income support
e_btype3, # sickness benefit, incl. ESA
# I think there's an error in the coding of Universal Credit
# it's labelled as e_btype10 but is in the wrong order
e_btype6, # tax credits
e_hcondno1,
e_hcondno2,
e_hcondno3,
e_hcondno4,
e_hcondno5,
e_hcondno6,
e_hcondno7,
e_hcond17 # clinical depression
)
# Prepare wave d ====
usd <- readr::read_tsv("inst/extdata/restricted_us/d_indresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(pidp, d_hidp,
d_hhorig,
d_dvage,
d_sex,
d_racel,
d_hiqual_dv,
d_jbstat, # economic activity
d_health,
d_scsf1,
d_jbnssec_dv, # NS-SEC/social class
d_jbnssec8_dv, # NS-SEC 8
d_jbnssec5_dv, # NS-SEC 5
d_jbnssec3_dv, # NS-SEC 3
d_mlstat, # marital status
d_finnow, # subjective financial situation
d_scghqa, # GHQ: concentration
d_scghqb, # GHQ: loss of sleep
d_scghqc, # GHQ: playing a useful role
d_scghqd, # GHQ: capable of making decisions
d_scghqe, # GHQ: constantly under strain
d_scghqf, # GHQ: problem overcoming difficulties
d_scghqg, # GHQ: enjoy day-to-day activities
d_scghqh, # GHQ: ability to face problems
d_scghqi, # GHQ: unhappy or depressed
d_scghqj, # GHQ: losing confidence
d_scghqk, # GHQ: believe worthless
d_scghql, # GHQ: general happiness
d_save, # saves money
d_benunemp1, # job-seekers allowance
d_btype2, # income support
d_btype3, # sickness benefit, incl. ESA
# I think there's an error in the coding of Universal Credit
# it's labelled as e_btype10 but is in the wrong order
d_btype6, # tax credits
d_hubuys, # who does food shopping? (for cohabiting couples)
d_hcondno1,
d_hcondno2,
d_hcondno3,
d_hcondno4,
d_hcondno5,
d_hcondno6,
d_hcondno7,
d_hcondno8,
d_hcondno9,
d_hcondno10,
d_hcond17 # clinical depression
)
# Prepare wave c ====
usc <- readr::read_tsv("inst/extdata/restricted_us/c_indresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(pidp, c_hidp,
c_hhorig,
c_dvage,
c_sex,
c_racel,
c_hiqual_dv,
c_jbstat,
c_health,
c_sf1,
c_jbnssec_dv, # NS-SEC/social class
c_jbnssec8_dv, # NS-SEC 8
c_jbnssec5_dv, # NS-SEC 5
c_jbnssec3_dv, # NS-SEC 3
c_mlstat, # marital status
c_scopngbhe, # willing improve neighbourhood
c_scopngbhg, # similar to others in neighbourhood
c_scopngbha, # belong neighbourhood
c_nbrcoh3, # people in neighbourhood can be trusted
c_crdark, # safe walking alone at night
c_orga1, # active in organisation
c_orga2, # active in organisation
c_orga3, # active in organisation
c_orga4, # active in organisation
c_orga5, # active in organisation
c_orga6, # active in organisation
c_orga7, # active in organisation
c_orga8, # active in organisation
c_orga9, # active in organisation
c_orga10, # active in organisation
c_orga11, # active in organisation
c_orga12, # active in organisation
c_orga13, # active in organisation
c_orga14, # active in organisation
c_orga15, # active in organisation
c_orga16, # active in organisation
c_nbrcoh4, # people in neighbourhood don't get along
c_poleff4, # don't have say in what g'ment does
c_siminc, # Prop. friends with similar income
c_simrace, # Prop. friends similar ethnicity
c_finnow, # subjective financial situation
c_benunemp1, # job-seekers allowance
c_btype2, # income support
c_btype3, # sickness benefit (inc. ESA)
# I think there's an error in the coding of Universal Credit
# it's labelled as e_btype10 but is in the wrong order
c_btype6, # tax credits
c_closenum, # Number of close friends
c_scghqa, # GHQ: concentration
c_scghqb, # GHQ: loss of sleep
c_scghqc, # GHQ: playing a useful role
c_scghqd, # GHQ: capable of making decisions
c_scghqe, # GHQ: constantly under strain
c_scghqf, # GHQ: problem overcoming difficulties
c_scghqg, # GHQ: enjoy day-to-day activities
c_scghqh, # GHQ: ability to face problems
c_scghqi, # GHQ: unhappy or depressed
c_scghqj, # GHQ: losing confidence
c_scghqk, # GHQ: believe worthless
c_scghql, # GHQ: general happiness
c_netlv_1, # distance best friend lives
c_netlv_2, # distance best friend lives
c_netlv_3, # distance best friend lives
c_netph_1, # best friend often in touch
c_netph_2, # best friend often in touch
c_netph_3, # best friend often in touch
c_hcondno1,
c_hcondno2,
c_hcondno3,
c_hcondno4,
c_hcondno5,
c_hcondno6,
c_hcondno7,
c_hcondno8,
c_hcond17 # clinical depression
)
# Prepare wave b ====
usb <- readr::read_tsv("inst/extdata/restricted_us/b_indresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(pidp, b_hidp,
b_hhorig,
b_dvage,
b_sex,
b_racel,
b_hiqual_dv,
b_jbstat,
b_health,
b_sf1,
b_jbnssec_dv, # NS-SEC/social class
b_jbnssec8_dv, # NS-SEC 8
b_jbnssec5_dv, # NS-SEC 5
b_jbnssec3_dv, # NS-SEC 3
b_mlstat, # marital status
b_finnow, # subjective financial situation
b_scghqa, # GHQ: concentration
b_scghqb, # GHQ: loss of sleep
b_scghqc, # GHQ: playing a useful role
b_scghqd, # GHQ: capable of making decisions
b_scghqe, # GHQ: constantly under strain
b_scghqf, # GHQ: problem overcoming difficulties
b_scghqg, # GHQ: enjoy day-to-day activities
b_scghqh, # GHQ: ability to face problems
b_scghqi, # GHQ: unhappy or depressed
b_scghqj, # GHQ: losing confidence
b_scghqk, # GHQ: believe worthless
b_scghql, # GHQ: general happiness
b_ncigs, # cigs smoked per day
b_save, # saves money
b_benunemp1, # job-seekers allowance
b_btype2, # income support
b_btype3, # sickness benefit (inc. ESA)
# I think there's an error in the coding of Universal Credit
# it's labelled as e_btype10 but is in the wrong order
b_btype6, # tax credits
b_hubuys, # who does food shopping? (for cohabiting couples)
b_hcondno1,
b_hcondno2,
b_hcondno3,
b_hcondno4,
b_hcondno5,
b_hcondno6,
b_hcondno7,
b_hcondno8,
b_hcondn17 # clinical depression
)
# Prepare wave a ====
usa <- readr::read_tsv("inst/extdata/restricted_us/a_indresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(pidp, a_hidp,
a_hhorig,
a_dvage,
a_sex,
a_racel,
a_hiqual_dv,
a_jbstat,
a_health,
a_sf1,
a_jbnssec_dv, # NS-SEC/social class
a_jbnssec8_dv, # NS-SEC 8
a_jbnssec5_dv, # NS-SEC 5
a_jbnssec3_dv, # NS-SEC 3
a_mlstat, # marital status
a_scopngbhe, # willing to improve neighbourhood
a_scopngbhg, # similar to others in neighbourhood
a_scopngbha, # belong to neighbourhood
a_sctrust, # trust in others (neighbourhood)
a_finnow, # subjective financial situation
a_benunemp1, # job-seekers allowance
a_btype2, # income support
a_btype3, # sickness benefit (inc. ESA)
# I think there's an error in the coding of Universal Credit
# it's labelled as e_btype10 but is in the wrong order
a_btype6, # tax credits
a_scghqa, # GHQ: concentration
a_scghqb, # GHQ: loss of sleep
a_scghqc, # GHQ: playing a useful role
a_scghqd, # GHQ: capable of making decisions
a_scghqe, # GHQ: constantly under strain
a_scghqf, # GHQ: problem overcoming difficulties
a_scghqg, # GHQ: enjoy day-to-day activities
a_scghqh, # GHQ: ability to face problems
a_scghqi, # GHQ: unhappy or depressed
a_scghqj, # GHQ: losing confidence
a_scghqk, # GHQ: believe worthless
a_scghql, # GHQ: general happiness
a_hcond17 # clinical depression
)
# Merge waves ====
us <- usf %>%
full_join(use, by = "pidp") %>%
full_join(usd, by = "pidp") %>%
full_join(usc, by = "pidp") %>%
full_join(usb, by = "pidp") %>%
full_join(usa, by = "pidp")
rm(usa, usb, usc, usd, use, usf)
stopifnot(
us$pidp == unique(us$pidp),
!(any(us == -9, na.rm = TRUE))
)
# Prepare household responses ====
hhf <- readr::read_tsv("inst/extdata/restricted_us/f_hhresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(f_hidp,
f_hsownd, f_tenure_dv,
f_ncars,
f_heatch,
f_gor_dv,
f_hsbeds, # no. bedrooms
f_hsrooms, # no. other rooms
f_hhsize, # no. people currently in household
f_nkids015, # no. children 0-15 in hhold
f_fihhmngrs_dv, # gross hh income month before interview
f_rent, # last rental payment
f_rentwc, # weeks covered by last rent payment f_rent
f_xpmg, # last mortgage payment
f_xphsdb, # problems paying housing costs
f_xpaltob_g3 # £ spent on alcohol
)
hhe <- readr::read_tsv("inst/extdata/restricted_us/e_hhresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(e_hidp,
e_hsownd, e_tenure_dv,
e_ncars,
e_heatch,
e_gor_dv,
e_hsbeds, # no. bedrooms
e_hsrooms, # no. other rooms
e_hhsize, # no. people currently in household
e_nkids_dv, # no. children in hhold
e_fihhmngrs_dv, # gross hh income month before interview
e_rent, # last rent payment £
e_rentwc, # weeks covered by e_rent
e_xpmg, # last mortgage payment
e_xphsdb, # behind with rent/mortgage
e_xpaltob_g3 # £ spent on alcohol
)
hhd <- readr::read_tsv("inst/extdata/restricted_us/d_hhresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(d_hidp,
d_hsownd, d_tenure_dv,
d_ncars,
d_heatch,
d_gor_dv,
d_hsbeds,
d_hsrooms,
d_hhsize,
d_nkids015,
d_fihhmngrs_dv, # gross hh income month before interview
d_rent, # last rent payment £
d_rentwc, # weeks covered by d_rent
d_xphsdb, # problems paying for housing costs
d_xpmg, # last mortgage payment
d_xpaltob_g3 # £ spent on alcohol
)
hhc <- readr::read_tsv("inst/extdata/restricted_us/c_hhresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(c_hidp,
c_hsownd, c_tenure_dv,
c_ncars,
c_heatch,
c_gor_dv,
c_hsbeds,
c_hsrooms,
c_hhsize,
c_nkids015, # no. children (<16) in hhold
c_fihhmngrs_dv, # gross hh income month before interview
c_rent, # last rent payment £
c_rentwc, # period covered c_rent
c_xpmg, # last mortgage payment
c_xphsdb, # behind with rent/mortgage
c_xpaltob_g3 # £ spent on alcohol
)
hhb <- readr::read_tsv("inst/extdata/restricted_us/b_hhresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(b_hidp,
b_hsownd, b_tenure_dv,
b_ncars,
b_heatch,
b_gor_dv,
b_hsbeds,
b_hsrooms,
b_hhsize,
b_nkids015, # no. children in hhold (<16)
b_fihhmngrs_dv, # gross hh income month before interview
b_xphsdb, # problems paying housing costs
b_xpmg, # last mortgage payment
b_rent, # last rent payment £
b_rentwc, # weeks covered b_rent
b_xpaltob_g3 # £ spent on alcohol
)
hha <- readr::read_tsv("inst/extdata/restricted_us/a_hhresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(a_hidp,
a_hsownd, a_tenure_dv,
a_ncars,
a_heatch,
a_gor_dv,
a_hsbeds,
a_hsrooms,
a_hhsize,
a_numchild, # no. children in hhold
a_fihhmngrs_dv, # gross hh income month before interview
a_rent, # last rent payment £
a_rentwc, # weeks covered last rent payment
a_xpmg, # last mortgage payment
a_xphsdb, # problems paying housing costs
a_xpaltob_g3 # £ spent on alcohol
)
stopifnot(
!(any(hhe == -9, na.rm = TRUE)),
!(any(hhd == -9, na.rm = TRUE)),
!(any(hhc == -9, na.rm = TRUE)),
!(any(hhb == -9, na.rm = TRUE)),
!(any(hha == -9, na.rm = TRUE))
)
# Merge household responses ====
us <- us %>%
left_join(hhf, by = "f_hidp") %>%
left_join(hhe, by = "e_hidp") %>%
left_join(hhd, by = "d_hidp") %>%
left_join(hhc, by = "c_hidp") %>%
left_join(hhb, by = "b_hidp") %>%
left_join(hha, by = "a_hidp")
rm(hhf, hhe, hhd, hhc, hhb, hha)
# # Nurse health assessments ====
#
# nhac <- readr::read_tsv("inst/extdata/restricted_us_nha/c_indresp_ns.tab")
#
# nhac <- select(nhac, -c_waist3) # All NA
#
# nhac[nhac == -9] <- NA # missing
# nhac[nhac == -8] <- NA # Inapplicable
# nhac[nhac == -7] <- NA # Proxy
# nhac[nhac == -2] <- NA # Refused
# nhac[nhac == -1] <- NA # Don't know
#
# stopifnot(
# !(any(nhac == -9, na.rm = TRUE))
# )
#
# nhab <- readr::read_tsv("inst/extdata/restricted_us_nha/b_indresp_ns.tab",
# col_types = cols(
# b_wtpc = col_double()
# )
# )
#
# nhab[nhab == -9] <- NA # missing
# nhab[nhab == -8] <- NA # Inapplicable
# nhab[nhab == -7] <- NA # Proxy
# nhab[nhab == -2] <- NA # Refused
# nhab[nhab == -1] <- NA # Don't know
#
# stopifnot(
# !(any(nhab == -9, na.rm = TRUE))
# )
#
#
# # Merge nurse health assessments ====
#
# us <- left_join(us, nhac, by = "pidp")
# us <- left_join(us, nhab, by = "pidp")
#
# rm(nhac, nhab)
# Test ====
stopifnot( # make sure variables from all waves are present
any(grepl("a_", colnames(us))),
any(grepl("b_", colnames(us))),
any(grepl("c_", colnames(us))),
any(grepl("d_", colnames(us))),
any(grepl("e_", colnames(us))),
any(grepl("f_", colnames(us))),
any(grepl("hh", colnames(us)))
)
# Code region ====
us$region <- NA
us$region <- us$f_gor_dv
us$region[is.na(us$region)] <- us$e_gor_dv[is.na(us$region)]
us$region[is.na(us$region)] <- us$d_gor_dv[is.na(us$region)]
us$region[is.na(us$region)] <- us$c_gor_dv[is.na(us$region)]
us$region[is.na(us$region)] <- us$b_gor_dv[is.na(us$region)]
us$region[is.na(us$region)] <- us$a_gor_dv[is.na(us$region)]
us$region <- factor(us$region, levels = 1:12,
labels = c("NE", "NW", "YH", "EM", "WM", "EE",
"LN", "SE", "SW", "WA", "SC", "NI"),
ordered = FALSE)
us <- us %>%
select(-f_gor_dv, -e_gor_dv, -d_gor_dv, -c_gor_dv, -b_gor_dv, -a_gor_dv)
# Code respondent origin ====
# Was going to remove BHPS sample members because of survey fatigue
# us$origin <- NA
# us$origin <- us$f_hhorig
# us$origin[is.na(us$origin)] <- us$e_hhorig[is.na(us$origin)]
# us$origin[is.na(us$origin)] <- us$d_hhorig[is.na(us$origin)]
# us$origin[is.na(us$origin)] <- us$c_hhorig[is.na(us$origin)]
# us$origin[is.na(us$origin)] <- us$b_hhorig[is.na(us$origin)]
# us$origin[is.na(us$origin)] <- us$a_hhorig[is.na(us$origin)]
# us <- us %>%
# filter(origin != 3) %>%
# filter(origin != 4) %>%
# filter(origin != 5) %>%
# filter(origin != 6)
us <- us %>%
select(-f_hhorig, -e_hhorig, -d_hhorig, -c_hhorig, -b_hhorig, -a_hhorig)
# Recode age ====
# cut us$c_dvage to match census
us$age_full <- NA
us$age_full <- us$f_dvage
us$age_full[is.na(us$age_full)] <- us$e_dvage[is.na(us$age_full)] + 1
us$age_full[is.na(us$age_full)] <- us$d_dvage[is.na(us$age_full)] + 2
us$age_full[is.na(us$age_full)] <- us$c_dvage[is.na(us$age_full)] + 3
us$age_full[is.na(us$age_full)] <- us$b_dvage[is.na(us$age_full)] + 4
us$age_full[is.na(us$age_full)] <- us$a_dvage[is.na(us$age_full)] + 5
us$age_full[is.na(us$age_full)] <- round(mean(us$age_full, na.rm = TRUE))
us$age <- cut(us$age_full,
breaks = c(min(us$age_full),
17, 19, 24, 29, 44, 59, 64, 74, 84, 89,
max(us$age_full)),
include.lowest = TRUE,
labels = c("age_16_17", "age_18_19", "age_20_24",
"age_25_29", "age_30_44", "age_45_59",
"age_60_64", "age_65_74", "age_75_84",
"age_85_89", "age_90_plus"))
stopifnot(
all(!is.na(us$age))
)
us <- us %>%
select(-f_dvage, -e_dvage, -d_dvage, -c_dvage, -b_dvage, -a_dvage)
# Recode sex ====
us$sex <- NA
us$sex <- us$f_sex
us$sex[is.na(us$sex)] <- us$e_sex[is.na(us$sex)]
us$sex[is.na(us$sex)] <- us$d_sex[is.na(us$sex)]
us$sex[is.na(us$sex)] <- us$c_sex[is.na(us$sex)]
us$sex[is.na(us$sex)] <- us$b_sex[is.na(us$sex)]
us$sex[is.na(us$sex)] <- us$a_sex[is.na(us$sex)]
us$sex <- factor(us$sex,
levels = 1:2,
labels = c("sex_male", "sex_female"))
us$sex <- forcats::fct_relevel(us$sex, "sex_female", "sex_male")
us <- us %>%
select(-f_sex, -e_sex, -d_sex, -c_sex, -b_sex, -a_sex)
# Recode ethnicity ====
us$eth_full <- NA
us$eth_full <- us$f_racel
us$eth_full[is.na(us$eth_full)] <- us$e_racel[is.na(us$eth_full)]
us$eth_full[is.na(us$eth_full)] <- us$d_racel[is.na(us$eth_full)]
us$eth_full[is.na(us$eth_full)] <- us$c_racel[is.na(us$eth_full)]
us$eth_full[is.na(us$eth_full)] <- us$b_racel[is.na(us$eth_full)]
us$eth_full[is.na(us$eth_full)] <- us$a_racel[is.na(us$eth_full)]
us$eth_full <- factor(us$eth_full, levels = c(1:17, 97),
labels = c("White British",
"Irish",
"Gypsy or Irish Traveller",
"other White",
"White and Black Caribbean",
"White and Black African",
"White and Asian",
"other mixed",
"Indian",
"Pakistani",
"Bangladeshi",
"Chinese",
"other Asian",
"Caribbean",
"African",
"other black",
"Arab",
"other ethnic group"),
ordered = FALSE)
us <- us %>%
select(-f_racel, -e_racel, -d_racel, -c_racel, -b_racel, -a_racel)
# Code ethnic minority group
us$eth_minority <- NA
us$eth_minority[us$eth_full == "White British"] <- FALSE
us$eth_minority[us$eth_full != "White British"] <- TRUE
# Code reduced ethnic groups
# Reduced ethnic group classes - to match census aggregate files
us$eth <- NA
us$eth[us$eth_full == "White British"] <- 1
us$eth[us$eth_full == "Irish"] <- 2
us$eth[us$eth_full == "Gypsy or Irish Traveller"] <- 3
us$eth[us$eth_full == "other White"] <- 3
us$eth[us$eth_full == "White and Black Caribbean"] <- 4
us$eth[us$eth_full == "White and Black African"] <- 4
us$eth[us$eth_full == "White and Asian"] <- 4
us$eth[us$eth_full == "other mixed"] <- 4
us$eth[us$eth_full == "Indian"] <- 5
us$eth[us$eth_full == "Pakistani"] <- 5
us$eth[us$eth_full == "Bangladeshi"] <- 5
us$eth[us$eth_full == "Chinese"] <- 5
us$eth[us$eth_full == "other Asian"] <- 5
us$eth[us$eth_full == "Caribbean"] <- 6
us$eth[us$eth_full == "African"] <- 6
us$eth[us$eth_full == "other black"] <- 6
us$eth[us$eth_full == "Arab"] <- 7
us$eth[us$eth_full == "other ethnic group"] <- 7
us$eth <- factor(us$eth,
levels = 1:7,
labels = c("eth_british",
"eth_irish",
"eth_other_white",
"eth_mixed_multiple_ethnic",
"eth_asian_asian_british",
"eth_black_african_caribbean_british",
"eth_other_ethnicity"),
ordered = FALSE)
# Recode education ====
# Create most recent known hiqual
us$tmp <- NA
us$tmp <- us$f_hiqual_dv
us$tmp[is.na(us$tmp)] <- us$e_hiqual_dv[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$d_hiqual_dv[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$c_hiqual_dv[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$b_hiqual_dv[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$a_hiqual_dv[is.na(us$tmp)]
us$qual <- NA
us$qual[us$tmp == 9] <- 0 # No qualification
# GCSE A*-C Level 2; GCSE D-G Level 1. US not split, so all GCSEs L2
us$qual[us$tmp == 4] <- 2 # L1 or L2: GCSE
us$qual[us$tmp == 3] <- 3 # L3: AS/A Level
us$qual[us$tmp == 2] <- 4 # L4: Degree or above
us$qual[us$tmp == 1] <- 4 # L4: Degree or above
us$qual[us$tmp == 5] <- 5 # other qualification
us$qual <- factor(us$qual,
levels = c(0, 2:5),
labels = c("qual_0",
"qual_2",
"qual_3",
"qual_4_plus",
"qual_other"),
ordered = FALSE)
us <- us %>%
select(-tmp,
-f_hiqual_dv, -e_hiqual_dv, -d_hiqual_dv, -c_hiqual_dv, -b_hiqual_dv,
-a_hiqual_dv)
# Recode car or van availability ====
us$car <- NA
us$car <- us$f_ncars
us$car[is.na(us$car)] <- us$e_ncars[is.na(us$car)]
us$car[is.na(us$car)] <- us$d_ncars[is.na(us$car)]
us$car[is.na(us$car)] <- us$c_ncars[is.na(us$car)]
us$car[is.na(us$car)] <- us$b_ncars[is.na(us$car)]
us$car[is.na(us$car)] <- us$a_ncars[is.na(us$car)]
us$car <- replace(us$car, us$car > 2 & !is.na(us$car), 2)
us$car <- factor(us$car, levels = 0:2,
labels = c("car_0", "car_1", "car_2_plus"),
ordered = FALSE)
us <- us %>%
select(-f_ncars, -e_ncars, -d_ncars, -c_ncars, -b_ncars, -a_ncars)
# Recode heat ====
us$heat <- NA
us$heat <- us$f_heatch
us$heat[is.na(us$heat)] <- us$e_heatch[is.na(us$heat)]
us$heat[is.na(us$heat)] <- us$d_heatch[is.na(us$heat)]
us$heat[is.na(us$heat)] <- us$c_heatch[is.na(us$heat)]
us$heat[is.na(us$heat)] <- us$b_heatch[is.na(us$heat)]
us$heat[is.na(us$heat)] <- us$a_heatch[is.na(us$heat)]
us$heat[us$heat == 2] <- 0
us$heat <- factor(us$heat, levels = 0:1,
labels = c("heat_no", "heat_yes"),
ordered = FALSE)
us <- us %>%
select(-f_heatch, -e_heatch, -d_heatch, -c_heatch, -b_heatch, -a_heatch)
# Recode tenure ====
us$ten_full <- NA
us$ten_full <- us$f_hsownd
us$ten_full[is.na(us$ten_full)] <- us$e_hsownd[is.na(us$ten_full)]
us$ten_full[is.na(us$ten_full)] <- us$d_hsownd[is.na(us$ten_full)]
us$ten_full[is.na(us$ten_full)] <- us$c_hsownd[is.na(us$ten_full)]
us$ten_full[is.na(us$ten_full)] <- us$b_hsownd[is.na(us$ten_full)]
us$ten_full[is.na(us$ten_full)] <- us$a_hsownd[is.na(us$ten_full)]
us$ten_full <- factor(us$ten_full, levels = c(1:5, 97),
labels = c("owned_outright",
"owned_mortgage_loan",
"shared_ownership",
"rented",
"living_rent_free",
"other_tenure"),
ordered = FALSE)
# Reduced categories of tenure to match census aggregates
us$ten <- NA
us$ten[us$ten_full == "owned_outright"] <- 1
us$ten[us$ten_full == "owned_mortgage_loan"] <- 2
us$ten[us$ten_full == "shared_ownership"] <- 2
us$ten[us$ten_full == "rented"] <- 3
us$ten[us$ten_full == "living_rent_free"] <- 3
us$ten[us$ten_full == "other_tenure"] <- NA
us$ten <- factor(us$ten,
levels = 1:3,
labels = c("ten_owned_outright",
"ten_owned_mortgage_shared",
"ten_rented"),
ordered = FALSE)
us <- us %>%
select(-f_tenure_dv, -e_tenure_dv, -d_tenure_dv, -c_tenure_dv,
-b_tenure_dv, -a_tenure_dv) %>%
select(-f_hsownd, -e_hsownd, -d_hsownd, -c_hsownd, -b_hsownd, -a_hsownd)
# Recode economic activity ====
us$econ_act <- NA
us$econ_act <- us$f_jbstat
us$econ_act[is.na(us$econ_act)] <- us$e_jbstat[is.na(us$econ_act)]
us$econ_act[is.na(us$econ_act)] <- us$d_jbstat[is.na(us$econ_act)]
us$econ_act[is.na(us$econ_act)] <- us$c_jbstat[is.na(us$econ_act)]
us$econ_act[is.na(us$econ_act)] <- us$b_jbstat[is.na(us$econ_act)]
us$econ_act[is.na(us$econ_act)] <- us$a_jbstat[is.na(us$econ_act)]
# levels checked as correct 2018-01-12
us$econ_act <- factor(us$econ_act, levels = c(1:11, 97),
labels = c("self_emp",
"emp",
"unemp",
"retired",
"maternity",
"home_fam",
"stu_ft",
"lt_sick",
"gment_train",
"unpaid_fam",
"apprent",
"other"))
# Reaggregate some variables to match census
us$econ_act <- us$econ_act %>%
forcats::fct_collapse(home_fam = c("home_fam", "maternity"),
other = c("other", "gment_train", "unpaid_fam",
"apprent"))
us$econ_act <- us$econ_act %>%
forcats::fct_recode(
eca_emp = "emp",
eca_homefam = "home_fam",
eca_ltsick = "lt_sick",
eca_other = "other",
eca_retired = "retired",
eca_selfemp = "self_emp",
eca_student = "stu_ft",
eca_unemp = "unemp"
) %>%
forcats::fct_relevel( # order alphabetically for simulation
"eca_emp", "eca_homefam", "eca_ltsick", "eca_other", "eca_retired",
"eca_selfemp", "eca_student", "eca_unemp"
)
us <- us %>%
select(-f_jbstat, -e_jbstat, -d_jbstat, -c_jbstat, -b_jbstat, -a_jbstat)
# Overcrowding (> 1 persons per room) ====
us$rooms <- NA
# add bedrooms + other accommodation rooms
us$rooms <- us$f_hsbeds + us$f_hsrooms
# make sure bedrooms and other rooms are from the same wave, i.e. the
# same property
# e.g. e_hsrooms + d_hsbeds might be two different properties
us$rooms[is.na(us$rooms)] <-
us$e_hsbeds[is.na(us$rooms)] + us$e_hsrooms[is.na(us$rooms)]
us$rooms[is.na(us$rooms)] <-
us$d_hsbeds[is.na(us$rooms)] + us$d_hsrooms[is.na(us$rooms)]
us$rooms[is.na(us$rooms)] <-
us$c_hsbeds[is.na(us$rooms)] + us$c_hsrooms[is.na(us$rooms)]
us$rooms[is.na(us$rooms)] <-
us$b_hsbeds[is.na(us$rooms)] + us$b_hsrooms[is.na(us$rooms)]
us$rooms[is.na(us$rooms)] <-
us$a_hsbeds[is.na(us$rooms)] + us$a_hsrooms[is.na(us$rooms)]
us$ppr <- NA
us$ppr <- us$f_hhsize
us$ppr[is.na(us$ppr)] <- us$e_hhsize[is.na(us$ppr)]
us$ppr[is.na(us$ppr)] <- us$d_hhsize[is.na(us$ppr)]
us$ppr[is.na(us$ppr)] <- us$c_hhsize[is.na(us$ppr)]
us$ppr[is.na(us$ppr)] <- us$b_hhsize[is.na(us$ppr)]
us$ppr[is.na(us$ppr)] <- us$a_hhsize[is.na(us$ppr)]
us$ppr <- us$ppr / us$rooms
us$ocrowd <- NA
us$ocrowd[us$ppr > 1] <- 1
us$ocrowd[us$ppr <= 1] <- 0
us$ocrowd <- factor(us$ocrowd, levels = c(0, 1),
labels = c("ocrowd_no",
"ocrowd_yes"),
ordered = FALSE)
stopifnot( # check ocrowd coded properly
all.equal(nrow(filter(us, ppr > 1, ocrowd == "ocrowd_no")), 0),
all.equal(nrow(filter(us, ppr <= 1, ocrowd == "ocrowd_yes")), 0)
)
# hhsize needed for alcohol consumption
us <- us %>%
select(-rooms, -ppr) %>%
select(-f_hsbeds, -e_hsbeds, -d_hsbeds, -c_hsbeds, -b_hsbeds, -a_hsbeds) %>%
select(-f_hsrooms, -e_hsrooms, -d_hsrooms, -c_hsrooms, -b_hsrooms,
-a_hsrooms)
# Limiting long-term illness or disability ====
us$llid <- NA
us$llid <- us$f_health
us$llid[is.na(us$llid)] <- us$e_health[is.na(us$llid)]
us$llid[is.na(us$llid)] <- us$d_health[is.na(us$llid)]
us$llid[is.na(us$llid)] <- us$c_health[is.na(us$llid)]
us$llid[is.na(us$llid)] <- us$b_health[is.na(us$llid)]
us$llid[is.na(us$llid)] <- us$a_health[is.na(us$llid)]
us$llid <- replace(us$llid, us$llid == 2, 0)
us$llid <- factor(us$llid, levels = 0:1,
labels = c("llid_no", "llid_yes"))
us <- us %>%
select(-f_health, -e_health, -d_health, -c_health, -b_health, -a_health)
# Recode self-reported general health ====
us$tmp <- NA
us$tmp <- us$f_scsf1
us$tmp[is.na(us$tmp)] <- us$e_scsf1[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$d_scsf1[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$c_sf1[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$b_sf1[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$a_sf1[is.na(us$tmp)]
us$srgh <- NA
us$srgh[us$tmp == 1] <- 5 # excellent
us$srgh[us$tmp == 2] <- 4 # very good
us$srgh[us$tmp == 3] <- 3 # good
us$srgh[us$tmp == 4] <- 2 # fair
us$srgh[us$tmp == 5] <- 1 # poor
us$srgh <- factor(us$srgh, levels = 5:1,
labels = c("srgh_excellent",
"srgh_verygood",
"srgh_good",
"srgh_fair",
"srgh_poor"),
ordered = FALSE)
us <- us %>%
select(-f_scsf1, -e_scsf1, -d_scsf1, -c_sf1, -b_sf1, -a_sf1,
-tmp)
# Recode social class ====
us$class8 <- NA
us$class8 <- us$f_jbnssec8_dv
us$class8[is.na(us$class8)] <- us$e_jbnssec8_dv[is.na(us$class8)]
us$class8[is.na(us$class8)] <- us$d_jbnssec8_dv[is.na(us$class8)]
us$class8[is.na(us$class8)] <- us$c_jbnssec8_dv[is.na(us$class8)]
us$class8[is.na(us$class8)] <- us$b_jbnssec8_dv[is.na(us$class8)]
us$class8[is.na(us$class8)] <- us$a_jbnssec8_dv[is.na(us$class8)]
# Full NS-SEC, NS-SEC5 and NS-SEC3 have same number of NAs as NSSEC8
# us$class <- NA
# us$class <- us$f_jbnssec_dv
# us$class[is.na(us$class)] <- us$e_jbnssec_dv[is.na(us$class)]
# us$class[is.na(us$class)] <- us$d_jbnssec_dv[is.na(us$class)]
# us$class[is.na(us$class)] <- us$c_jbnssec_dv[is.na(us$class)]
# us$class[is.na(us$class)] <- us$b_jbnssec_dv[is.na(us$class)]
# us$class[is.na(us$class)] <- us$a_jbnssec_dv[is.na(us$class)]
#
# us$class5 <- NA
# us$class5 <- us$f_jbnssec5_dv
# us$class5[is.na(us$class5)] <- us$e_jbnssec5_dv[is.na(us$class5)]
# us$class5[is.na(us$class5)] <- us$d_jbnssec5_dv[is.na(us$class5)]
# us$class5[is.na(us$class5)] <- us$c_jbnssec5_dv[is.na(us$class5)]
# us$class5[is.na(us$class5)] <- us$b_jbnssec5_dv[is.na(us$class5)]
# us$class5[is.na(us$class5)] <- us$a_jbnssec5_dv[is.na(us$class5)]
#
# us$class3 <- NA
# us$class3 <- us$f_jbnssec3_dv
# us$class3[is.na(us$class3)] <- us$e_jbnssec3_dv[is.na(us$class3)]
# us$class3[is.na(us$class3)] <- us$d_jbnssec3_dv[is.na(us$class3)]
# us$class3[is.na(us$class3)] <- us$c_jbnssec3_dv[is.na(us$class3)]
# us$class3[is.na(us$class3)] <- us$b_jbnssec3_dv[is.na(us$class3)]
# us$class3[is.na(us$class3)] <- us$a_jbnssec3_dv[is.na(us$class3)]
us$class8 <- factor(us$class8, levels = 1:9,
labels = c("Large employers & higher management",
"Higher professional",
"Lower management & professional",
"Intermediate",
"Small employers & own account",
"Lower supervisory & technical",
"Semi-routine",
"Routine",
"Never worked & LT unemployed"),
ordered = FALSE)
us <- us %>%
select(-f_jbnssec_dv, -f_jbnssec8_dv, -f_jbnssec5_dv, -f_jbnssec3_dv) %>%
select(-e_jbnssec_dv, -e_jbnssec8_dv, -e_jbnssec5_dv, -e_jbnssec3_dv) %>%
select(-d_jbnssec_dv, -d_jbnssec8_dv, -d_jbnssec5_dv, -d_jbnssec3_dv) %>%
select(-c_jbnssec_dv, -c_jbnssec8_dv, -c_jbnssec5_dv, -c_jbnssec3_dv) %>%
select(-b_jbnssec_dv, -b_jbnssec8_dv, -b_jbnssec5_dv, -b_jbnssec3_dv) %>%
select(-a_jbnssec_dv, -a_jbnssec8_dv, -a_jbnssec5_dv, -a_jbnssec3_dv)
# # Tax credits ====
# us$taxcred <- NA
# us$taxcred <- us$f_btype6
# us$taxcred[is.na(us$taxcred)] <- us$e_btype6[is.na(us$taxcred)]
# us$taxcred[is.na(us$taxcred)] <- us$d_btype6[is.na(us$taxcred)]
# us$taxcred[is.na(us$taxcred)] <- us$c_btype6[is.na(us$taxcred)]
# us$taxcred[is.na(us$taxcred)] <- us$b_btype6[is.na(us$taxcred)]
# us$taxcred[is.na(us$taxcred)] <- us$a_btype6[is.na(us$taxcred)]
# Marital status ====
us$marital <- NA
us$marital <- us$f_mlstat
us$marital[is.na(us$marital)] <- us$e_mlstat[is.na(us$marital)]
us$marital[is.na(us$marital)] <- us$d_mlstat[is.na(us$marital)]
us$marital[is.na(us$marital)] <- us$c_mlstat[is.na(us$marital)]
us$marital[is.na(us$marital)] <- us$b_mlstat[is.na(us$marital)]
us$marital[is.na(us$marital)] <- us$a_mlstat[is.na(us$marital)]
# checked these levels are correct 2018-01-18
us$marital <- factor(us$marital, levels = 1:9,
labels = c("single",
"married",
"civil_part",
"separated",
"divorced",
"widowed",
"sep_civil_part",
"former_civil_part",
"surviving_civil_part")) %>%
forcats::fct_collapse(
separated = c("separated", "sep_civil_part"),
divorced = c("divorced", "former_civil_part"),
widowed = c("widowed", "surviving_civil_part")) %>%
forcats::fct_recode(
mar_civil_part = "civil_part",
mar_divorced = "divorced",
mar_married = "married",
mar_separated = "separated",
mar_single = "single",
mar_widowed = "widowed"
) %>%
forcats::fct_relevel( # make alphabetical for simulation
"mar_civil_part", "mar_divorced", "mar_married", "mar_separated",
"mar_single", "mar_widowed"
)
us <- us %>%
select(-f_mlstat, -e_mlstat, -d_mlstat, -c_mlstat, -b_mlstat, -a_mlstat)
# Couple ====
# This is couple, married or not
# In waves f_, d_ and b_ asked about domestic chores
# if there's any answer (or they're married) they're part of a couple
# I have to assume the remaining NAs are single
us$couple <- NA
us$couple[us$f_hubuys > 0 & !is.na(us$f_hubuys)] <- 1
us$couple[is.na(us$couple) & us$marital == "mar_married"] <- 1
us$couple[is.na(us$couple) & us$d_hubuys > 0 & !is.na(us$d_hubuys)] <- 1
us$couple[is.na(us$couple) & us$b_hubuys > 0 & !is.na(us$b_hubuys)] <- 1
us <- us %>%
select(-f_hubuys, -d_hubuys, -b_hubuys)
# Neighbourhood cohesion ====
# Paper 13 neighbourhood cohesion:
# extent to which neighbours pull together to improve neighbourhood
# No direct analogy but proxy:
# willing to improve neighbourhood, and
# similar to others in neighbourhood
# Willing to improve neighbourhood
us$imp_nhood <- NA
us$imp_nhood <- us$f_scopngbhe
us$imp_nhood[is.na(us$imp_nhood)] <- us$c_scopngbhe[is.na(us$imp_nhood)]
us$imp_nhood[is.na(us$imp_nhood)] <- us$a_scopngbhe[is.na(us$imp_nhood)]
# Similar to others in neighbourhood
us$sim_nhood <- NA
us$sim_nhood <- us$f_scopngbhg
us$sim_nhood[is.na(us$sim_nhood)] <- us$c_scopngbhg[is.na(us$sim_nhood)]
us$sim_nhood[is.na(us$sim_nhood)] <- us$a_scopngbhg[is.na(us$sim_nhood)]
# dplyr::case_when() doesn't yet work inside mutate()
# <= 2 == strongly agree + agree
# >= 3 == Neither + disagree + strongly disagree
us <- us %>%
mutate(nhood_coh = if_else(imp_nhood <= 2 & sim_nhood <= 2,
"ncoh_yes", "ncoh_no"))
stopifnot(
# more agree + strongly agree than cohesive neighbourhoods
nrow(us[us$imp_nhood <= 2 & !is.na(us$imp_nhood), ]) >
nrow(us[us$nhood_coh == "ncoh_yes" & !is.na(us$nhood_coh), ]),
nrow(us[us$sim_nhood <= 2 & !is.na(us$sim_nhood), ]) >
nrow(us[us$nhood_coh == "ncoh_yes" & !is.na(us$nhood_coh), ])
)
us <- us %>%
select(-imp_nhood, -sim_nhood,
-f_scopngbhe, -c_scopngbhe, -a_scopngbhe,
-f_scopngbhg, -c_scopngbhg, -a_scopngbhg)
# Neighbourhood trust ====
# Waves f and c are different to how wave a is coded
# Wave f and c: strongly agree + agree == trust
# wave a: most people can be trusted == trust
us$nhood_trust <- NA
us$nhood_trust <- if_else(us$f_nbrcoh3 <= 2, "trust_yes", "trust_no")
us$nhood_trust[is.na(us$nhood_trust)] <-
if_else(us$c_nbrcoh3[is.na(us$nhood_trust)] <= 2,
"trust_yes",
"trust_no")
us$nhood_trust[is.na(us$nhood_trust)] <-
if_else(us$a_sctrust[is.na(us$nhood_trust)] == 1,
"trust_yes",
"trust_no")
us <- us %>%
select(-f_nbrcoh3, -c_nbrcoh3, -a_sctrust)
# Neighbourhood belonging ====
us$nhood_belong <- NA
us$nhood_belong <- us$f_scopngbha
us$nhood_belong[is.na(us$nhood_belong)] <-
us$c_scopngbha[is.na(us$nhood_belong)]
us$nhood_belong[is.na(us$nhood_belong)] <-
us$a_scopngbha[is.na(us$nhood_belong)]
us$nhood_belong <- if_else(us$nhood_belong <= 2, "belong_yes", "belong_no")
us <- us %>%
select(-f_scopngbha, -c_scopngbha, -a_scopngbha)
# Civic participation ====
civic_vars <- colnames(us[, grep("f_orga", colnames(us))])
civic_vars <- civic_vars[civic_vars != "f_orga96"] # 'none of these'
us$civic <- NA
us$civic <- rowSums(us[, civic_vars])
civic_vars <- colnames(us[, grep("c_orga", colnames(us))])
civic_vars <- civic_vars[civic_vars != "c_orga96"] # 'none of these'
us$civic[is.na(us$civic)] <- rowSums(us[is.na(us$civic), civic_vars])
us$civic <- if_else(us$civic >= 1, "civic_yes", "civic_no")
us <- us %>%
select(-f_orga1, -f_orga2, -f_orga3, -f_orga4, -f_orga5, -f_orga6,
-f_orga7, -f_orga8, -f_orga9, -f_orga10, -f_orga11, -f_orga12,
-f_orga13, -f_orga14, -f_orga15, -f_orga16,
-c_orga1, -c_orga2, -c_orga3, -c_orga4, -c_orga5, -c_orga6,
-c_orga7, -c_orga8, -c_orga9, -c_orga10, -c_orga11, -c_orga12,
-c_orga13, -c_orga14, -c_orga15, -c_orga16)
rm(civic_vars)
# Social cohesion ====
# People in this neighbourhood get along - paper 13
# People in this neighbourhood DON'T get along - US
# Reverse, so take strongly disagree + disagree
us$social_coh <- NA
us$social_coh <- us$f_nbrcoh4
us$social_coh[is.na(us$social_coh)] <- us$c_nbrcoh4[is.na(us$social_coh)]
us$social_coh <- if_else(us$social_coh >= 4, "scoh_yes", "scoh_no")
us <- us %>%
select(-f_nbrcoh4, -c_nbrcoh4)
# Safe walking alone at night ====
us$safe_alone <- NA
us$safe_alone <- us$f_crdark
us$safe_alone[is.na(us$safe_alone)] <- us$c_crdark[is.na(us$safe_alone)]
# Recode to safe or unsafe
# 1 = very safe
us$safe_alone[us$safe_alone == 2] <- 1 # fairly safe
us$safe_alone[us$safe_alone == 3] <- 2 # a bit unsafe
us$safe_alone[us$safe_alone == 4] <- 2 # very unsafe
us$safe_alone[us$safe_alone == 5] <- 2 # don't go out after dark
us$safe_alone <- factor(us$safe_alone, levels = 1:2,
labels = c("safe_yes",
"safe_no"))
us <- us %>%
select(-f_crdark, -c_crdark)
# Similar income to friends ====
# Paper 13 doesn't specify which levels map to 'hetergeneous' or homogeneous
# friendship. Assume ~ <= 1/2 is hetero
us$sim_inc <- NA
us$sim_inc <- us$f_siminc
us$sim_inc[is.na(us$sim_inc)] <- us$c_siminc[is.na(us$sim_inc)]
us$sim_inc <- if_else(us$sim_inc >= 3, "siminc_het", "siminc_hom")
us <- us %>%
select(-f_siminc, -c_siminc)
# Similiar ethnicity to friends ====
# Paper 13 again doesn't specify which levels map to hetero/homo
# Assume again ~ <= 1/2 hetero; >= 1/2 homo
us$sim_eth <- NA
us$sim_eth <- us$f_simrace
us$sim_eth[is.na(us$sim_eth)] <- us$c_simrace[is.na(us$sim_eth)]
us$sim_eth <- if_else(us$sim_eth >= 3, "simeth_het", "simeth_hom")
us <- us %>%
select(-f_simrace, -c_simrace)
# Political efficacy ====
# Don't have a say in what government does, reversed
us$pol_eff <- NA
us$pol_eff <- us$f_poleff4
us$pol_eff[is.na(us$pol_eff)] <- us$c_poleff4[is.na(us$pol_eff)]
us$pol_eff <- if_else(us$pol_eff >= 4, "poleff_yes", "poleff_no")
us <- us %>%
select(-f_poleff4, -c_poleff4)
# Financial situation ====
us$fin_sit <- NA
us$fin_sit <- us$f_finnow
us$fin_sit[is.na(us$fin_sit)] <- us$e_finnow[is.na(us$fin_sit)]
us$fin_sit[is.na(us$fin_sit)] <- us$d_finnow[is.na(us$fin_sit)]
us$fin_sit[is.na(us$fin_sit)] <- us$c_finnow[is.na(us$fin_sit)]
us$fin_sit[is.na(us$fin_sit)] <- us$b_finnow[is.na(us$fin_sit)]
us$fin_sit[is.na(us$fin_sit)] <- us$a_finnow[is.na(us$fin_sit)]
us$fin_sit <- if_else(us$fin_sit <= 2, "fin_good", "fin_bad")
us <- us %>%
select(-f_finnow, -e_finnow, -d_finnow, -c_finnow, -b_finnow, -a_finnow)
# Social isolation ====
# >=1 close friend
us$soc_isol <- NA
us$soc_isol <- us$f_closenum
us$soc_isol[is.na(us$soc_isol)] <- us$c_closenum[is.na(us$soc_isol)]
us$soc_isol <- if_else(us$soc_isol > 0, "isol_no", "isol_yes")
us <- us %>%
select(-f_closenum, -c_closenum)
# Income ====
# I've tested:
# - total earnings annually (x_prearna)
# - total earnings weekly (x_prearnw)
# - total personal income annually (x_prfitba)
# - total personal income weekly (x_prfitw)
#
# All (even when combined) have low response rates and ~ 77k missing cases
# General Health Questionnaire (GHQ) ====
recode_ghq <- function(varname) {
x <- NA
x <- us[[paste0("f_", varname)]]
x[is.na(x)] <- us[[paste0("f_", varname)]][is.na(x)]
x[is.na(x)] <- us[[paste0("e_", varname)]][is.na(x)]
x[is.na(x)] <- us[[paste0("d_", varname)]][is.na(x)]
x[is.na(x)] <- us[[paste0("c_", varname)]][is.na(x)]
x[is.na(x)] <- us[[paste0("b_", varname)]][is.na(x)]
x[is.na(x)] <- us[[paste0("a_", varname)]][is.na(x)]
x
}
us$ghq_conc <- recode_ghq("scghqa")
us$ghq_sleep <- recode_ghq("scghqb")
us$ghq_useful <- recode_ghq("scghqc")
us$ghq_decision <- recode_ghq("scghqd")
us$ghq_strain <- recode_ghq("scghqe")
us$ghq_diff <- recode_ghq("scghqf")
us$ghq_enjoy <- recode_ghq("scghqg")
us$ghq_face <- recode_ghq("scghqh")
us$ghq_unhappy <- recode_ghq("scghqi")
us$ghq_confid <- recode_ghq("scghqj")
us$ghq_worth <- recode_ghq("scghqk")
us$ghq_happy <- recode_ghq("scghql")
us$ghq_conc <- if_else(us$ghq_conc <= 2,
"ghq_conc_good", "ghq_conc_bad")
us$ghq_sleep <- if_else(us$ghq_sleep <= 2,
"ghq_sleep_good", "ghq_sleep_bad")
us$ghq_useful <- if_else(us$ghq_useful <= 2,
"ghq_useful_good", "ghq_useful_bad")
us$ghq_decision <- if_else(us$ghq_decision <= 2,
"ghq_decision_good", "ghq_decision_bad")
us$ghq_strain <- if_else(us$ghq_strain <= 2,
"ghq_strain_good", "ghq_strain_bad")
us$ghq_diff <- if_else(us$ghq_diff <= 2,
"ghq_diff_good", "ghq_diff_bad")
us$ghq_enjoy <- if_else(us$ghq_enjoy <= 2,
"ghq_enjoy_good", "ghq_enjoy_bad")
us$ghq_face <- if_else(us$ghq_face <= 2,
"ghq_face_good", "ghq_face_bad")
us$ghq_unhappy <- if_else(us$ghq_unhappy <= 2,
"ghq_unhappy_good", "ghq_unhappy_bad")
us$ghq_confid <- if_else(us$ghq_confid <= 2,
"ghq_confid_good", "ghq_confid_bad")
us$ghq_worth <- if_else(us$ghq_worth <= 2,
"ghq_worth_good", "ghq_worth_bad")
us$ghq_happy <- if_else(us$ghq_happy <= 2,
"ghq_happy_good", "ghq_happy_bad")
us <- us %>%
select(-f_scghqa, -e_scghqa, -d_scghqa, -c_scghqa, -b_scghqa, -a_scghqa,
-f_scghqb, -e_scghqb, -d_scghqb, -c_scghqb, -b_scghqb, -a_scghqb,
-f_scghqc, -e_scghqc, -d_scghqc, -c_scghqc, -b_scghqc, -a_scghqc,
-f_scghqd, -e_scghqd, -d_scghqd, -c_scghqd, -b_scghqd, -a_scghqd,
-f_scghqe, -e_scghqe, -d_scghqe, -c_scghqe, -b_scghqe, -a_scghqe,
-f_scghqf, -e_scghqf, -d_scghqf, -c_scghqf, -b_scghqf, -a_scghqf,
-f_scghqg, -e_scghqg, -d_scghqg, -c_scghqg, -b_scghqg, -a_scghqg,
-f_scghqh, -e_scghqh, -d_scghqh, -c_scghqh, -b_scghqh, -a_scghqh,
-f_scghqi, -e_scghqi, -d_scghqi, -c_scghqi, -b_scghqi, -a_scghqi,
-f_scghqj, -e_scghqj, -d_scghqj, -c_scghqj, -b_scghqj, -a_scghqj,
-f_scghqk, -e_scghqk, -d_scghqk, -c_scghqk, -b_scghqk, -a_scghqk,
-f_scghql, -e_scghql, -d_scghql, -c_scghql, -b_scghql, -a_scghql)
rm(recode_ghq)
# Cigarettes smoked ====
us$cig <- NA
us$cig <- us$f_ncigs
us$cig[is.na(us$cig)] <- us$e_ncigs[is.na(us$cig)]
us$cig[is.na(us$cig)] <- us$b_ncigs[is.na(us$cig)]
us$cig <- if_else(us$cig > 0, "cig_yes", "cig_no")
us <- us %>%
select(-f_ncigs, -e_ncigs, -b_ncigs)
# Alcohol ====
# units = (volume (ml) * ABV) / 1000
# 1 pint = 568.261ml (Google)
# Source of alcohol costs: institute alcohol studies
# http://www.ias.org.uk/uploads/pdf/Factsheets/Alcohol%20pricing%20factsheet%20April%202014.pdf
# All dates 2012 data
#
# ABV from NHS.uk/livewell/alcohol/pages/alcohol-units.aspx and IAS
#
# Beer: £2.70 pint 3.6% abv
# Lager: £2.96 pint 5.2% abv
# Cider: £1.82 pint 5.3% abv
# Sherry: £6.55 70cl 17.5% abv
# Wine: £4.04 75cl 12% abv
# Whiskey: £15.35 70cl 40% abv
# Vodka: £13.86 70cl 37.5% abv
units <- data.frame(
beer = c(2.7, 3.6, 568),
lager = c(2.96, 5.2, 568),
cider = c(1.82, 5.3, 568),
sherry = c(6.55, 17.5, 700),
wine = c(4.04, 12.0, 750),
whiskey = c(15.35, 40.0, 700),
vodka = c(13.86, 37.5, 700)
)
units <- as.data.frame(t(units))
colnames(units) <- c("cost", "abv", "ml")
units$units <- (units$ml * units$abv) / 1000
# source: http://www.nhs.uk/Livewell/alcohol/Pages/alcohol-units.aspx
units$unit_cost <- units$units / units$cost
units$unit_cost <- round(units$unit_cost, digits = 2)
unit_cost <- median(units$unit_cost)
us <- us %>%
mutate(f_alc = (f_xpaltob_g3 / (f_hhsize - f_nkids015)) / unit_cost / 4,
e_alc = (e_xpaltob_g3 / (e_hhsize - e_nkids_dv)) / unit_cost / 4,
d_alc = (d_xpaltob_g3 / (d_hhsize - d_nkids015)) / unit_cost / 4,
c_alc = (c_xpaltob_g3 / (c_hhsize - c_nkids015)) / unit_cost / 4,
b_alc = (b_xpaltob_g3 / (b_hhsize - b_nkids015)) / unit_cost / 4,
a_alc = (a_xpaltob_g3 / (a_hhsize - a_numchild)) / unit_cost / 4)
us$alcohol <- NA
us$alcohol <- us$f_alc
us$alcohol[is.na(us$alcohol)] <- us$e_alc[is.na(us$alcohol)]
us$alcohol[is.na(us$alcohol)] <- us$d_alc[is.na(us$alcohol)]
us$alcohol[is.na(us$alcohol)] <- us$c_alc[is.na(us$alcohol)]
us$alcohol[is.na(us$alcohol)] <- us$b_alc[is.na(us$alcohol)]
us$alcohol[is.na(us$alcohol)] <- us$a_alc[is.na(us$alcohol)]
# Low/high risk thresholds
# cmo2016a, p. 4
us$alcohol <- if_else(us$alcohol <= 14, "alcohol_low", "alcohol_high")
us <- us %>%
select(-f_xpaltob_g3, -e_xpaltob_g3, -d_xpaltob_g3,
-c_xpaltob_g3, -b_xpaltob_g3, -a_xpaltob_g3,
-f_alc, -e_alc, -d_alc, -c_alc, -b_alc, -a_alc)
rm(units, unit_cost)
# Best friend distance ====
us <- us %>%
rowwise() %>%
mutate(f_bfdist = min(f_netlv_1, f_netlv_2, f_netlv_3),
c_bfdist = min(c_netlv_1, c_netlv_2, c_netlv_3))
us$bf_dist <- NA
us$bf_dist <- us$f_bfdist
us$bf_dist[is.na(us$bf_dist)] <- us$f_bfdist[is.na(us$bf_dist)]
# Paper: < 5 mins. Use < 1 mile as proxy
us$bf_dist <- if_else(us$bf_dist > 1, "bfdist_far", "bfdist_near")
us <- us %>%
select(-f_bfdist, -c_bfdist,
-f_netlv_1, -f_netlv_2, -f_netlv_3,
-c_netlv_1, -c_netlv_2, -c_netlv_3)
# Best friend often in touch ====
# Paper: frequent contact == at least weekly
us <- us %>%
rowwise() %>%
mutate(f_bfoften = min(f_netph_1, f_netph_2, f_netph_3),
c_bfoften = min(c_netph_1, c_netph_2, c_netph_3))
us$bf_often <- NA
us$bf_often <- us$f_bfoften
us$bf_often[is.na(us$bf_often)] <- us$c_bfoften[is.na(us$bf_often)]
us$bf_often <- if_else(us$bf_often > 1, "bfoften_not", "bfoften_often")
us <- us %>%
select(-f_bfoften, -c_bfoften,
-f_netph_1, -f_netph_2, -f_netph_3,
-c_netph_1, -c_netph_2, -c_netph_3)
# Savings ====
# Budgeting/money management skills
# Too many NAs save_reg
us$save <- NA
us$save <- us$f_save
us$save[is.na(us$save)] <- us$d_save[is.na(us$save)]
us$save[is.na(us$save)] <- us$b_save[is.na(us$save)]
us$save <- if_else(us$save == 1, "save_yes", "save_no")
us <- us %>%
select(-f_save, -d_save, -b_save)
# Number of children ====
us$nkids <- NA
us$nkids <- us$f_nkids015
us$nkids[is.na(us$nkids)] <- us$e_nkids_dv[is.na(us$nkids)]
us$nkids[is.na(us$nkids)] <- us$d_nkids015[is.na(us$nkids)]
us$nkids[is.na(us$nkids)] <- us$c_nkids015[is.na(us$nkids)]
us$nkids[is.na(us$nkids)] <- us$b_nkids015[is.na(us$nkids)]
us$nkids[is.na(us$nkids)] <- us$a_numchild[is.na(us$nkids)]
us <- us %>%
select(-f_nkids015, -e_nkids_dv, -d_nkids015, -c_nkids015, -b_nkids015,
-a_numchild)
# Can't use numeric data with rakeR::extract() easily because it creates a
# new variable for each level of the variable (i.e. one for each unique
# integer)
# Recode and top-code to 5 kids
us$kids <- cut(us$nkids,
breaks = c(0, 1, 2, 3, 4, 5, max(us$nkids, na.rm = TRUE)),
labels = c("kids_0", "kids_1", "kids_2", "kids_3", "kids_4",
"kids_5ormore"),
right = FALSE, include.lowest = TRUE)
# Household income and poverty ====
# Many measures of poverty based on household income (i.e. children do not
# earn but are not in poverty if earning adults aren't)
# US provides monthly hh income. Convert to annual.
# Poverty defined as <60% median income
# Median income in 2011 was £23,200 (source: ons2013c, p. 5)
us$hh_income <- us$f_fihhmngrs_dv
us$hh_income[is.na(us$hh_income)] <- us$e_fihhmngrs_dv[is.na(us$hh_income)]
us$hh_income[is.na(us$hh_income)] <- us$d_fihhmngrs_dv[is.na(us$hh_income)]
us$hh_income[is.na(us$hh_income)] <- us$c_fihhmngrs_dv[is.na(us$hh_income)]
us$hh_income[is.na(us$hh_income)] <- us$b_fihhmngrs_dv[is.na(us$hh_income)]
us$hh_income[is.na(us$hh_income)] <- us$a_fihhmngrs_dv[is.na(us$hh_income)]
# Set negative values to 0
us$hh_income[us$hh_income < 0] <- 0L
us <- us %>%
select(-f_fihhmngrs_dv, -e_fihhmngrs_dv, -d_fihhmngrs_dv, -c_fihhmngrs_dv,
-b_fihhmngrs_dv, -a_fihhmngrs_dv)
# Number of cohabitants ====
us$cohab <- NA
us$cohab <- us$f_hhsize
us$cohab[is.na(us$cohab)] <- us$e_hhsize[is.na(us$cohab)]
us$cohab[is.na(us$cohab)] <- us$d_hhsize[is.na(us$cohab)]
us$cohab[is.na(us$cohab)] <- us$c_hhsize[is.na(us$cohab)]
us$cohab[is.na(us$cohab)] <- us$b_hhsize[is.na(us$cohab)]
us$cohab[is.na(us$cohab)] <- us$a_hhsize[is.na(us$cohab)]
us$cohab <- us$cohab - 1 # remove the respondent!
us <- us %>%
select(-f_hhsize, -e_hhsize, -d_hhsize, -c_hhsize, -b_hhsize, -a_hhsize)
# Clinical depression ====
us$depress <- NA_integer_
hconds <- colnames(us)[grepl("_hcond", colnames(us))]
for (i in hconds) {
us[[i]] <- as.integer(us[[i]])
us$depress[is.na(us$depress) & us[[i]] == 17 & !is.na(us[[i]])] <- 1
}
rm(i)
us$depress[is.na(us$depress)] <- us$f_hcond17[is.na(us$depress)]
us$depress[is.na(us$depress)] <- us$e_hcond17[is.na(us$depress)]
us$depress[is.na(us$depress)] <- us$d_hcond17[is.na(us$depress)]
us$depress[is.na(us$depress)] <- us$c_hcond17[is.na(us$depress)]
us$depress[is.na(us$depress)] <- us$b_hcondn17[is.na(us$depress)]
us$depress[is.na(us$depress)] <- us$a_hcond17[is.na(us$depress)]
us$depress <- factor(us$depress, levels = 0:1,
labels = c("depress_no", "depress_yes"),
ordered = FALSE)
# IHD and stroke ====
# Air pollution is associated with IHD and stroke (among others)
# This ihd measure therefore includes IND and stroke
# IHD includes angina and myocardial infarction (MI)
us$ihd <- NA_integer_
for (i in hconds) {
us[[i]] <- as.integer(us[[i]])
us$ihd[is.na(us$ihd) & !is.na(us[[i]]) &
us[[i]] == 4 | # coronary heart disease
us[[i]] == 5 | # angina
us[[i]] == 6 | # myocardial infarction
us[[i]] == 7] <- 1 # stroke
}
us$ihd[is.na(us$ihd)] <- 0
us$ihd <- factor(us$ihd, levels = 0:1,
labels = c("ihd_no", "ihd_yes"))
us <- us %>%
select(-f_hcond17, -e_hcond17, -d_hcond17, -c_hcond17, -b_hcondn17,
-a_hcond17)
us <- us %>%
select(-f_hcondno1:-f_hcondno8) %>%
select(-e_hcondno1:-e_hcondno7) %>%
select(-d_hcondno1:-d_hcondno10) %>%
select(-c_hcondno1:-c_hcondno8) %>%
select(-b_hcondno1:-b_hcondno8)
# Problems paying housing costs ====
us$prob_pay <- us$f_xphsdb
us$prob_pay[is.na(us$prob_pay)] <- us$e_xphsdb[is.na(us$prob_pay)]
us$prob_pay[is.na(us$prob_pay)] <- us$d_xphsdb[is.na(us$prob_pay)]
us$prob_pay[is.na(us$prob_pay)] <- us$c_xphsdb[is.na(us$prob_pay)]
us$prob_pay[is.na(us$prob_pay)] <- us$b_xphsdb[is.na(us$prob_pay)]
us$prob_pay[is.na(us$prob_pay)] <- us$a_xphsdb[is.na(us$prob_pay)]
us <- us %>%
select(-f_xphsdb, -e_xphsdb, -d_xphsdb, -c_xphsdb, -b_xphsdb, -a_xphsdb)
us$prob_pay <- factor(us$prob_pay, levels = 1:2,
labels = c("prob_pay_yes", "prob_pay_no"))
# Housing costs ====
# household income is based on monthly, so standardise on monthly
us$f_rent[us$f_rentwc == 1 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 1 & !is.na(us$f_rentwc)] * 52 / 12
us$f_rent[us$f_rentwc == 2 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 2 & !is.na(us$f_rentwc)] * 52 / 2 / 12
us$f_rent[us$f_rentwc == 3 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 3 & !is.na(us$f_rentwc)] * 52 / 3 / 12
us$f_rent[us$f_rentwc == 4 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 4 & !is.na(us$f_rentwc)] * 52 / 4 / 12
us$f_rent[us$f_rentwc == 7 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 7 & !is.na(us$f_rentwc)] / 2
us$f_rent[us$f_rentwc == 8 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 8 & !is.na(us$f_rentwc)] * 8 / 12
us$f_rent[us$f_rentwc == 9 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 9 & !is.na(us$f_rentwc)] * 9 / 12
us$f_rent[us$f_rentwc == 10 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 10 & !is.na(us$f_rentwc)] * 10 / 12
us$f_rent[us$f_rentwc == 13 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 13 & !is.na(us$f_rentwc)] / 3
us$f_rent[us$f_rentwc == 26 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 26 & !is.na(us$f_rentwc)] / 6
us$f_rent[us$f_rentwc == 52 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 52 & !is.na(us$f_rentwc)] / 12
us$f_rent[us$f_rentwc == 90 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 90 & !is.na(us$f_rentwc)] * 52 / 12
us$f_rent[us$f_rentwc == 95 & !is.na(us$f_rentwc)] <- NA # one-off payment
us$f_rent[us$f_rentwc == 96 & !is.na(us$f_rentwc)] <- NA # none of these
us$e_rent[us$e_rentwc == 1 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 1 & !is.na(us$e_rentwc)] * 52 / 12
us$e_rent[us$e_rentwc == 2 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 2 & !is.na(us$e_rentwc)] * 52 / 2 / 12
us$e_rent[us$e_rentwc == 3 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 3 & !is.na(us$e_rentwc)] * 52 / 3 / 12
us$e_rent[us$e_rentwc == 4 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 4 & !is.na(us$e_rentwc)] * 52 / 4 / 12
us$e_rent[us$e_rentwc == 7 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 7 & !is.na(us$e_rentwc)] / 2
us$e_rent[us$e_rentwc == 8 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 8 & !is.na(us$e_rentwc)] * 8 / 12
us$e_rent[us$e_rentwc == 9 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 9 & !is.na(us$e_rentwc)] * 9 / 12
us$e_rent[us$e_rentwc == 10 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 10 & !is.na(us$e_rentwc)] * 10 / 12
us$e_rent[us$e_rentwc == 13 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 13 & !is.na(us$e_rentwc)] / 3
us$e_rent[us$e_rentwc == 26 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 26 & !is.na(us$e_rentwc)] / 6
us$e_rent[us$e_rentwc == 52 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 52 & !is.na(us$e_rentwc)] / 12
us$e_rent[us$e_rentwc == 90 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 90 & !is.na(us$e_rentwc)] * 52 / 12
us$e_rent[us$e_rentwc == 95 & !is.na(us$e_rentwc)] <- NA # one-off payment
us$e_rent[us$e_rentwc == 96 & !is.na(us$e_rentwc)] <- NA # none of these
us$d_rent[us$d_rentwc == 1 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 1 & !is.na(us$d_rentwc)] * 52 / 12
us$d_rent[us$d_rentwc == 2 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 2 & !is.na(us$d_rentwc)] * 52 / 2 / 12
us$d_rent[us$d_rentwc == 3 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 3 & !is.na(us$d_rentwc)] * 52 / 3 / 12
us$d_rent[us$d_rentwc == 4 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 4 & !is.na(us$d_rentwc)] * 52 / 4 / 12
us$d_rent[us$d_rentwc == 7 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 7 & !is.na(us$d_rentwc)] / 2
us$d_rent[us$d_rentwc == 8 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 8 & !is.na(us$d_rentwc)] * 8 / 12
us$d_rent[us$d_rentwc == 9 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 9 & !is.na(us$d_rentwc)] * 9 / 12
us$d_rent[us$d_rentwc == 10 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 10 & !is.na(us$d_rentwc)] * 10 / 12
us$d_rent[us$d_rentwc == 13 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 13 & !is.na(us$d_rentwc)] / 3
us$d_rent[us$d_rentwc == 26 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 26 & !is.na(us$d_rentwc)] / 6
us$d_rent[us$d_rentwc == 52 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 52 & !is.na(us$d_rentwc)] / 12
us$d_rent[us$d_rentwc == 90 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 90 & !is.na(us$d_rentwc)] * 52 / 12
us$d_rent[us$d_rentwc == 95 & !is.na(us$d_rentwc)] <- NA # one-off payment
us$d_rent[us$d_rentwc == 96 & !is.na(us$d_rentwc)] <- NA # none of these
us$c_rent[us$c_rentwc == 1 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 1 & !is.na(us$c_rentwc)] * 52 / 12
us$c_rent[us$c_rentwc == 2 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 2 & !is.na(us$c_rentwc)] * 52 / 2 / 12
us$c_rent[us$c_rentwc == 3 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 3 & !is.na(us$c_rentwc)] * 52 / 3 / 12
us$c_rent[us$c_rentwc == 4 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 4 & !is.na(us$c_rentwc)] * 52 / 4 / 12
us$c_rent[us$c_rentwc == 7 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 7 & !is.na(us$c_rentwc)] / 2
us$c_rent[us$c_rentwc == 8 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 8 & !is.na(us$c_rentwc)] * 8 / 12
us$c_rent[us$c_rentwc == 9 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 9 & !is.na(us$c_rentwc)] * 9 / 12
us$c_rent[us$c_rentwc == 10 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 10 & !is.na(us$c_rentwc)] * 10 / 12
us$c_rent[us$c_rentwc == 13 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 13 & !is.na(us$c_rentwc)] / 3
us$c_rent[us$c_rentwc == 26 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 26 & !is.na(us$c_rentwc)] / 6
us$c_rent[us$c_rentwc == 52 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 52 & !is.na(us$c_rentwc)] / 12
us$c_rent[us$c_rentwc == 90 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 90 & !is.na(us$c_rentwc)] * 52 / 12
us$c_rent[us$c_rentwc == 95 & !is.na(us$c_rentwc)] <- NA # one-off payment
us$c_rent[us$c_rentwc == 96 & !is.na(us$c_rentwc)] <- NA # none of these
us$b_rent[us$b_rentwc == 1 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 1 & !is.na(us$b_rentwc)] * 52 / 12
us$b_rent[us$b_rentwc == 2 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 2 & !is.na(us$b_rentwc)] * 52 / 2 / 12
us$b_rent[us$b_rentwc == 3 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 3 & !is.na(us$b_rentwc)] * 52 / 3 / 12
us$b_rent[us$b_rentwc == 4 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 4 & !is.na(us$b_rentwc)] * 52 / 4 / 12
us$b_rent[us$b_rentwc == 7 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 7 & !is.na(us$b_rentwc)] / 2
us$b_rent[us$b_rentwc == 8 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 8 & !is.na(us$b_rentwc)] * 8 / 12
us$b_rent[us$b_rentwc == 9 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 9 & !is.na(us$b_rentwc)] * 9 / 12
us$b_rent[us$b_rentwc == 10 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 10 & !is.na(us$b_rentwc)] * 10 / 12
us$b_rent[us$b_rentwc == 13 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 13 & !is.na(us$b_rentwc)] / 3
us$b_rent[us$b_rentwc == 26 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 26 & !is.na(us$b_rentwc)] / 6
us$b_rent[us$b_rentwc == 52 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 52 & !is.na(us$b_rentwc)] / 12
us$b_rent[us$b_rentwc == 90 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 90 & !is.na(us$b_rentwc)] * 52 / 12
us$b_rent[us$b_rentwc == 95 & !is.na(us$b_rentwc)] <- NA # one-off payment
us$b_rent[us$b_rentwc == 96 & !is.na(us$b_rentwc)] <- NA # none of these
us$a_rent[us$a_rentwc == 1 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 1 & !is.na(us$a_rentwc)] * 52 / 12
us$a_rent[us$a_rentwc == 2 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 2 & !is.na(us$a_rentwc)] * 52 / 2 / 12
us$a_rent[us$a_rentwc == 3 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 3 & !is.na(us$a_rentwc)] * 52 / 3 / 12
us$a_rent[us$a_rentwc == 4 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 4 & !is.na(us$a_rentwc)] * 52 / 4 / 12
us$a_rent[us$a_rentwc == 7 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 7 & !is.na(us$a_rentwc)] / 2
us$a_rent[us$a_rentwc == 8 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 8 & !is.na(us$a_rentwc)] * 8 / 12
us$a_rent[us$a_rentwc == 9 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 9 & !is.na(us$a_rentwc)] * 9 / 12
us$a_rent[us$a_rentwc == 10 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 10 & !is.na(us$a_rentwc)] * 10 / 12
us$a_rent[us$a_rentwc == 13 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 13 & !is.na(us$a_rentwc)] / 3
us$a_rent[us$a_rentwc == 26 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 26 & !is.na(us$a_rentwc)] / 6
us$a_rent[us$a_rentwc == 52 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 52 & !is.na(us$a_rentwc)] / 12
us$a_rent[us$a_rentwc == 90 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 90 & !is.na(us$a_rentwc)] * 52 / 12
us$a_rent[us$a_rentwc == 95 & !is.na(us$a_rentwc)] <- NA # one-off payment
us$a_rent[us$a_rentwc == 96 & !is.na(us$a_rentwc)] <- NA # none of these
us <- us %>%
select(-f_rentwc, -e_rentwc, -d_rentwc, -c_rentwc, -b_rentwc, -a_rentwc)
us$hh_cost <- us$f_xpmg
us$hh_cost[is.na(us$hh_cost)] <- us$f_rent[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$e_xpmg[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$e_rent[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$d_xpmg[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$d_rent[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$c_xpmg[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$c_rent[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$b_xpmg[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$b_rent[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$a_xpmg[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$a_rent[is.na(us$hh_cost)]
# top code to £20,000
us$hh_cost[us$hh_cost > 20000] <- 20000
us <- us %>%
select(-f_xpmg, -e_xpmg, -d_xpmg, -c_xpmg, -b_xpmg, -a_xpmg) %>%
select(-f_rent, -e_rent, -d_rent, -c_rent, -b_rent, -a_rent)
# Benefits ====
us$uc_replace <- 0
us$uc_replace[us$f_benbase1 == 1] <- 1 # income support
us$uc_replace[us$f_benbase2 == 1] <- 1 # JSA
us$uc_replace[us$f_othben5 == 1] <- 1 # working tax credit
us$uc_replace[us$f_othben8 == 1] <- 1 # housing benefit
us$uc_replace[us$f_benctc == 1] <- 1 # housing benefit (2 == no)
us$uc_replace[us$f_bendis2 == 1] <- 1 # ESA
us$uc_replace[is.na(us$uc_replace) & us$e_benunemp1 == 1] <- 1 # JSA
us$uc_replace[is.na(us$uc_replace) & us$e_btype2 == 1] <- 1 # income sup
us$uc_replace[is.na(us$uc_replace) & us$e_btype3 == 1] <- 1 # ESA
us$uc_replace[is.na(us$uc_replace) & us$e_btype6 == 1] <- 1 # tax cred
us$uc_replace[is.na(us$uc_replace) & us$d_benunemp1 == 1] <- 1 # JSA
us$uc_replace[is.na(us$uc_replace) & us$d_btype2 == 1] <- 1 # income sup
us$uc_replace[is.na(us$uc_replace) & us$d_btype3 == 1] <- 1 # ESA
us$uc_replace[is.na(us$uc_replace) & us$d_btype6 == 1] <- 1 # tax cred
us$uc_replace[is.na(us$uc_replace) & us$c_benunemp1 == 1] <- 1 # JSA
us$uc_replace[is.na(us$uc_replace) & us$c_btype2 == 1] <- 1 # income sup
us$uc_replace[is.na(us$uc_replace) & us$c_btype3 == 1] <- 1 # ESA
us$uc_replace[is.na(us$uc_replace) & us$c_btype6 == 1] <- 1 # tax cred
us$uc_replace[is.na(us$uc_replace) & us$b_benunemp1 == 1] <- 1 # JSA
us$uc_replace[is.na(us$uc_replace) & us$b_btype2 == 1] <- 1 # income sup
us$uc_replace[is.na(us$uc_replace) & us$b_btype3 == 1] <- 1 # ESA
us$uc_replace[is.na(us$uc_replace) & us$b_btype6 == 1] <- 1 # tax cred
us$uc_replace[is.na(us$uc_replace) & us$a_benunemp1 == 1] <- 1 # JSA
us$uc_replace[is.na(us$uc_replace) & us$a_btype2 == 1] <- 1 # income sup
us$uc_replace[is.na(us$uc_replace) & us$a_btype3 == 1] <- 1 # ESA
us$uc_replace[is.na(us$uc_replace) & us$a_btype6 == 1] <- 1 # tax cred
us <- us %>%
select(-f_benbase1, -f_benbase2, -f_othben5, -f_othben8, -f_benctc,
-f_bendis2,
-e_benunemp1, -e_btype2, -e_btype3, -e_btype6,
-d_benunemp1, -d_btype2, -d_btype3, -d_btype6,
-c_benunemp1, -c_btype2, -c_btype3, -c_btype6,
-b_benunemp1, -b_btype2, -b_btype3, -b_btype6,
-a_benunemp1, -a_btype2, -a_btype3, -a_btype6)
# Save data frame ====
save(us, file = "data/us.RData", compress = "xz")
rm(us)
}
philmikejones/thesis documentation built on May 31, 2019, 2:50 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
##
## Warning!
## Requires ~8GB RAM to run the script
##
if (!dir.exists("inst/extdata/restricted_us")) {
if (Sys.info()["sysname"] != "Linux") {
warning("It looks like you're not using a Linux system\n
I cannot automatically unzip and place the
Understanding Society data files in the correct place.\n
Manually copy the *.tab files in UKDA-6614-tab/tab
to inst/extdata/restricted_us/\n
You can also move the documentation to
inst/doc/restricted_us [optional]")
} else {
dir.create("inst/extdata/restricted_us/", showWarnings = FALSE)
# Looks like filename is unique; use 6614TAB_ to obtain the filename
us_zip <- list.files("inst/extdata/", pattern = "6614TAB_",
full.names = TRUE)
unzip(us_zip, exdir = "inst/extdata/restricted_us/")
# Move documentation
dir.create("inst/doc/restricted_us",
recursive = TRUE, showWarnings = FALSE)
system2("mv", args = c("inst/extdata/restricted_us/UKDA-6614-tab/mrdoc",
"inst/doc/restricted_us"))
# Move data files
system2("mv",
args = c("inst/extdata/restricted_us/UKDA-6614-tab/tab/*.tab",
"inst/extdata/restricted_us/"))
system2("rm", c("-rf", "inst/extdata/restricted_us/UKDA-6614-tab/tab/"))
# Move README and file information
system2("mv", c("inst/extdata/restricted_us/UKDA-6614-tab/*.*",
"inst/doc/restricted_us/"))
system2("rm", c("-rf", "inst/extdata/restricted_us/UKDA-6614-tab/"))
}
rm(us_zip)
}
if (!file.exists("data/us.RData")) {
# Prepare wave f ====
usf <- readr::read_tsv("inst/extdata/restricted_us/f_indresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")
# -11 only available for IEMB
# -10 not available for IEMB
# -9 missing
# -8 Inapplicable
# -7 Proxy
# -2 Refused
# -1 Don't know
) %>%
select(pidp, f_hidp, # id and household id
f_hhorig, # sample origin (i.e. BHPS)
f_dvage, # derived age
f_sex, # sex corrected
f_racel, # ethnic group
f_hiqual_dv, # highest qualification
f_jbstat, # current economic activity
f_health, # long-standing illness or disability
f_scsf1, # self-reported general health
f_jbnssec_dv, # NS-SEC/social class
f_jbnssec8_dv, # NS-SEC 8
f_jbnssec5_dv, # NS-SEC 5
f_jbnssec3_dv, # NS-SEC 3
f_mlstat, # marital status
f_scopngbhe, # willing to improve neighbourhood
f_scopngbhg, # similar to others in neighbourhood
f_nbrcoh3, # people in neighbourhood can be trusted
f_scopngbha, # belong to neighbourhood
f_orga1, # active in organisation
f_orga2, # active in organisation
f_orga3, # active in organisation
f_orga4, # active in organisation
f_orga5, # active in organisation
f_orga6, # active in organisation
f_orga7, # active in organisation
f_orga8, # active in organisation
f_orga9, # active in organisation
f_orga10, # active in organisation
f_orga11, # active in organisation
f_orga12, # active in organisation
f_orga13, # active in organisation
f_orga14, # active in organisation
f_orga15, # active in organisation
f_orga16, # active in organisation
f_nbrcoh4, # people in neighbourhood don't get along
f_crdark, # feel safe walking alone at night
f_siminc, # Prop friends with similar income
f_simrace, # Friends similar ethnicity
f_poleff4, # don't have say in what g'ment does
f_finnow, # subjective financial status
f_closenum, # number of close friends
f_scghqa, # GHQ: concentration
f_scghqb, # GHQ: loss of sleep
f_scghqc, # GHQ: playing a useful role
f_scghqd, # GHQ: capable of making decisions
f_scghqe, # GHQ: constantly under strain
f_scghqf, # GHQ: problem overcoming difficulties
f_scghqg, # GHQ: enjoy day-to-day activities
f_scghqh, # GHQ: ability to face problems
f_scghqi, # GHQ: unhappy or depressed
f_scghqj, # GHQ: losing confidence
f_scghqk, # GHQ: believe worthless
f_scghql, # GHQ: general happiness
f_ncigs, # Cigs smoked
f_netlv_1, # distance best friend lives
f_netlv_2, # distance best friend lives
f_netlv_3, # distance best friend lives
f_netph_1, # best friend often in touch
f_netph_2, # best friend often in touch
f_netph_3, # best friend often in touch
f_save, # saves money
# received core benefit:
f_benbase1, # 1 == income support
f_benbase2, # 2 == job-seekers allowance
# other benefit:
f_othben5, # 5 == working tax credit;
f_othben8, # 8 == housing benefit
f_benctc, # received child tax credit
f_bendis2, # disability benefits (ESA == 2)
f_hubuys, # who does food shopping? (for cohabiting couples)
f_hcondno1,
f_hcondno2,
f_hcondno3,
f_hcondno4,
f_hcondno5,
f_hcondno6,
f_hcondno7,
f_hcondno8,
f_hcond17 # clinical depression
# f_btype6, # tax credits
# f_disdif, # Type of impairment or disability
# f_hcondno1, # Diagnosed health condition(s)
# f_hcondno2, f_hcondno3, f_hcondno4, f_hcondno5, f_hcondno6,
# f_hcondno7,
# f_hcondns1, # still has condition
# f_hcondns2, f_hcondns3, f_hcondns4, f_hcondns5, f_hcondns6,
# f_hcondns7,
# f_scsf2a, # health limits moderate activities
# f_scsf2b, # health limits several flights of stairs
# f_scsf3a, # health limits work (last 4 wks)
# f_scsf3b, # health limits kind of work (last 4 wks)
# f_scsf4a, # mental health accomp. less (last 4 wks)
# f_scsf4b, # mental health worked less carefully (last 4 wks)
# f_scsf5, # pain interfered w/ work (last 4 wks)
# f_scsf6a, # felt calm & peaceful (last 4 wks)
# f_scsf6b, # Had lots of energy (last 4 wks)
# f_scsf6c, # Felt downhearted/depressed (last 4 wks)
# f_scsf7, # health (p or m) interfered w/ social life (4 wks)
# f_sclfsat1, # Satisfaction with health
# f_sclfsato, # Satisfaction with life overall
)
# Prepare wave e ====
use <- readr::read_tsv("inst/extdata/restricted_us/e_indresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(pidp, e_hidp, # id and household id
e_hhorig, # sample origin (i.e. BHPS)
e_dvage, # derived age
e_sex, # sex corrected
e_racel, # ethnic group
e_hiqual_dv, # highest qualification
e_jbstat, # current economic activity
e_health, # long-standing illness or disability
e_scsf1, # self-reported general health
e_jbnssec_dv, # NS-SEC/social class
e_jbnssec8_dv, # NS-SEC 8
e_jbnssec5_dv, # NS-SEC 5
e_jbnssec3_dv, # NS-SEC 3
e_mlstat, # marital status
e_finnow, # subjective financial status
e_scghqa, # GHQ: concentration
e_scghqb, # GHQ: loss of sleep
e_scghqc, # GHQ: playing a useful role
e_scghqd, # GHQ: capable of making decisions
e_scghqe, # GHQ: constantly under strain
e_scghqf, # GHQ: problem overcoming difficulties
e_scghqg, # GHQ: enjoy day-to-day activities
e_scghqh, # GHQ: ability to face problems
e_scghqi, # GHQ: unhappy or depressed
e_scghqj, # GHQ: losing confidence
e_scghqk, # GHQ: believe worthless
e_scghql, # GHQ: general happiness
e_ncigs, # cigs smoked per day
e_benunemp1, # job-seekers allowance
e_btype2, # income support
e_btype3, # sickness benefit, incl. ESA
# I think there's an error in the coding of Universal Credit
# it's labelled as e_btype10 but is in the wrong order
e_btype6, # tax credits
e_hcondno1,
e_hcondno2,
e_hcondno3,
e_hcondno4,
e_hcondno5,
e_hcondno6,
e_hcondno7,
e_hcond17 # clinical depression
)
# Prepare wave d ====
usd <- readr::read_tsv("inst/extdata/restricted_us/d_indresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(pidp, d_hidp,
d_hhorig,
d_dvage,
d_sex,
d_racel,
d_hiqual_dv,
d_jbstat, # economic activity
d_health,
d_scsf1,
d_jbnssec_dv, # NS-SEC/social class
d_jbnssec8_dv, # NS-SEC 8
d_jbnssec5_dv, # NS-SEC 5
d_jbnssec3_dv, # NS-SEC 3
d_mlstat, # marital status
d_finnow, # subjective financial situation
d_scghqa, # GHQ: concentration
d_scghqb, # GHQ: loss of sleep
d_scghqc, # GHQ: playing a useful role
d_scghqd, # GHQ: capable of making decisions
d_scghqe, # GHQ: constantly under strain
d_scghqf, # GHQ: problem overcoming difficulties
d_scghqg, # GHQ: enjoy day-to-day activities
d_scghqh, # GHQ: ability to face problems
d_scghqi, # GHQ: unhappy or depressed
d_scghqj, # GHQ: losing confidence
d_scghqk, # GHQ: believe worthless
d_scghql, # GHQ: general happiness
d_save, # saves money
d_benunemp1, # job-seekers allowance
d_btype2, # income support
d_btype3, # sickness benefit, incl. ESA
# I think there's an error in the coding of Universal Credit
# it's labelled as e_btype10 but is in the wrong order
d_btype6, # tax credits
d_hubuys, # who does food shopping? (for cohabiting couples)
d_hcondno1,
d_hcondno2,
d_hcondno3,
d_hcondno4,
d_hcondno5,
d_hcondno6,
d_hcondno7,
d_hcondno8,
d_hcondno9,
d_hcondno10,
d_hcond17 # clinical depression
)
# Prepare wave c ====
usc <- readr::read_tsv("inst/extdata/restricted_us/c_indresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(pidp, c_hidp,
c_hhorig,
c_dvage,
c_sex,
c_racel,
c_hiqual_dv,
c_jbstat,
c_health,
c_sf1,
c_jbnssec_dv, # NS-SEC/social class
c_jbnssec8_dv, # NS-SEC 8
c_jbnssec5_dv, # NS-SEC 5
c_jbnssec3_dv, # NS-SEC 3
c_mlstat, # marital status
c_scopngbhe, # willing improve neighbourhood
c_scopngbhg, # similar to others in neighbourhood
c_scopngbha, # belong neighbourhood
c_nbrcoh3, # people in neighbourhood can be trusted
c_crdark, # safe walking alone at night
c_orga1, # active in organisation
c_orga2, # active in organisation
c_orga3, # active in organisation
c_orga4, # active in organisation
c_orga5, # active in organisation
c_orga6, # active in organisation
c_orga7, # active in organisation
c_orga8, # active in organisation
c_orga9, # active in organisation
c_orga10, # active in organisation
c_orga11, # active in organisation
c_orga12, # active in organisation
c_orga13, # active in organisation
c_orga14, # active in organisation
c_orga15, # active in organisation
c_orga16, # active in organisation
c_nbrcoh4, # people in neighbourhood don't get along
c_poleff4, # don't have say in what g'ment does
c_siminc, # Prop. friends with similar income
c_simrace, # Prop. friends similar ethnicity
c_finnow, # subjective financial situation
c_benunemp1, # job-seekers allowance
c_btype2, # income support
c_btype3, # sickness benefit (inc. ESA)
# I think there's an error in the coding of Universal Credit
# it's labelled as e_btype10 but is in the wrong order
c_btype6, # tax credits
c_closenum, # Number of close friends
c_scghqa, # GHQ: concentration
c_scghqb, # GHQ: loss of sleep
c_scghqc, # GHQ: playing a useful role
c_scghqd, # GHQ: capable of making decisions
c_scghqe, # GHQ: constantly under strain
c_scghqf, # GHQ: problem overcoming difficulties
c_scghqg, # GHQ: enjoy day-to-day activities
c_scghqh, # GHQ: ability to face problems
c_scghqi, # GHQ: unhappy or depressed
c_scghqj, # GHQ: losing confidence
c_scghqk, # GHQ: believe worthless
c_scghql, # GHQ: general happiness
c_netlv_1, # distance best friend lives
c_netlv_2, # distance best friend lives
c_netlv_3, # distance best friend lives
c_netph_1, # best friend often in touch
c_netph_2, # best friend often in touch
c_netph_3, # best friend often in touch
c_hcondno1,
c_hcondno2,
c_hcondno3,
c_hcondno4,
c_hcondno5,
c_hcondno6,
c_hcondno7,
c_hcondno8,
c_hcond17 # clinical depression
)
# Prepare wave b ====
usb <- readr::read_tsv("inst/extdata/restricted_us/b_indresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(pidp, b_hidp,
b_hhorig,
b_dvage,
b_sex,
b_racel,
b_hiqual_dv,
b_jbstat,
b_health,
b_sf1,
b_jbnssec_dv, # NS-SEC/social class
b_jbnssec8_dv, # NS-SEC 8
b_jbnssec5_dv, # NS-SEC 5
b_jbnssec3_dv, # NS-SEC 3
b_mlstat, # marital status
b_finnow, # subjective financial situation
b_scghqa, # GHQ: concentration
b_scghqb, # GHQ: loss of sleep
b_scghqc, # GHQ: playing a useful role
b_scghqd, # GHQ: capable of making decisions
b_scghqe, # GHQ: constantly under strain
b_scghqf, # GHQ: problem overcoming difficulties
b_scghqg, # GHQ: enjoy day-to-day activities
b_scghqh, # GHQ: ability to face problems
b_scghqi, # GHQ: unhappy or depressed
b_scghqj, # GHQ: losing confidence
b_scghqk, # GHQ: believe worthless
b_scghql, # GHQ: general happiness
b_ncigs, # cigs smoked per day
b_save, # saves money
b_benunemp1, # job-seekers allowance
b_btype2, # income support
b_btype3, # sickness benefit (inc. ESA)
# I think there's an error in the coding of Universal Credit
# it's labelled as e_btype10 but is in the wrong order
b_btype6, # tax credits
b_hubuys, # who does food shopping? (for cohabiting couples)
b_hcondno1,
b_hcondno2,
b_hcondno3,
b_hcondno4,
b_hcondno5,
b_hcondno6,
b_hcondno7,
b_hcondno8,
b_hcondn17 # clinical depression
)
# Prepare wave a ====
usa <- readr::read_tsv("inst/extdata/restricted_us/a_indresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(pidp, a_hidp,
a_hhorig,
a_dvage,
a_sex,
a_racel,
a_hiqual_dv,
a_jbstat,
a_health,
a_sf1,
a_jbnssec_dv, # NS-SEC/social class
a_jbnssec8_dv, # NS-SEC 8
a_jbnssec5_dv, # NS-SEC 5
a_jbnssec3_dv, # NS-SEC 3
a_mlstat, # marital status
a_scopngbhe, # willing to improve neighbourhood
a_scopngbhg, # similar to others in neighbourhood
a_scopngbha, # belong to neighbourhood
a_sctrust, # trust in others (neighbourhood)
a_finnow, # subjective financial situation
a_benunemp1, # job-seekers allowance
a_btype2, # income support
a_btype3, # sickness benefit (inc. ESA)
# I think there's an error in the coding of Universal Credit
# it's labelled as e_btype10 but is in the wrong order
a_btype6, # tax credits
a_scghqa, # GHQ: concentration
a_scghqb, # GHQ: loss of sleep
a_scghqc, # GHQ: playing a useful role
a_scghqd, # GHQ: capable of making decisions
a_scghqe, # GHQ: constantly under strain
a_scghqf, # GHQ: problem overcoming difficulties
a_scghqg, # GHQ: enjoy day-to-day activities
a_scghqh, # GHQ: ability to face problems
a_scghqi, # GHQ: unhappy or depressed
a_scghqj, # GHQ: losing confidence
a_scghqk, # GHQ: believe worthless
a_scghql, # GHQ: general happiness
a_hcond17 # clinical depression
)
# Merge waves ====
us <- usf %>%
full_join(use, by = "pidp") %>%
full_join(usd, by = "pidp") %>%
full_join(usc, by = "pidp") %>%
full_join(usb, by = "pidp") %>%
full_join(usa, by = "pidp")
rm(usa, usb, usc, usd, use, usf)
stopifnot(
us$pidp == unique(us$pidp),
!(any(us == -9, na.rm = TRUE))
)
# Prepare household responses ====
hhf <- readr::read_tsv("inst/extdata/restricted_us/f_hhresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(f_hidp,
f_hsownd, f_tenure_dv,
f_ncars,
f_heatch,
f_gor_dv,
f_hsbeds, # no. bedrooms
f_hsrooms, # no. other rooms
f_hhsize, # no. people currently in household
f_nkids015, # no. children 0-15 in hhold
f_fihhmngrs_dv, # gross hh income month before interview
f_rent, # last rental payment
f_rentwc, # weeks covered by last rent payment f_rent
f_xpmg, # last mortgage payment
f_xphsdb, # problems paying housing costs
f_xpaltob_g3 # £ spent on alcohol
)
hhe <- readr::read_tsv("inst/extdata/restricted_us/e_hhresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(e_hidp,
e_hsownd, e_tenure_dv,
e_ncars,
e_heatch,
e_gor_dv,
e_hsbeds, # no. bedrooms
e_hsrooms, # no. other rooms
e_hhsize, # no. people currently in household
e_nkids_dv, # no. children in hhold
e_fihhmngrs_dv, # gross hh income month before interview
e_rent, # last rent payment £
e_rentwc, # weeks covered by e_rent
e_xpmg, # last mortgage payment
e_xphsdb, # behind with rent/mortgage
e_xpaltob_g3 # £ spent on alcohol
)
hhd <- readr::read_tsv("inst/extdata/restricted_us/d_hhresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(d_hidp,
d_hsownd, d_tenure_dv,
d_ncars,
d_heatch,
d_gor_dv,
d_hsbeds,
d_hsrooms,
d_hhsize,
d_nkids015,
d_fihhmngrs_dv, # gross hh income month before interview
d_rent, # last rent payment £
d_rentwc, # weeks covered by d_rent
d_xphsdb, # problems paying for housing costs
d_xpmg, # last mortgage payment
d_xpaltob_g3 # £ spent on alcohol
)
hhc <- readr::read_tsv("inst/extdata/restricted_us/c_hhresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(c_hidp,
c_hsownd, c_tenure_dv,
c_ncars,
c_heatch,
c_gor_dv,
c_hsbeds,
c_hsrooms,
c_hhsize,
c_nkids015, # no. children (<16) in hhold
c_fihhmngrs_dv, # gross hh income month before interview
c_rent, # last rent payment £
c_rentwc, # period covered c_rent
c_xpmg, # last mortgage payment
c_xphsdb, # behind with rent/mortgage
c_xpaltob_g3 # £ spent on alcohol
)
hhb <- readr::read_tsv("inst/extdata/restricted_us/b_hhresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(b_hidp,
b_hsownd, b_tenure_dv,
b_ncars,
b_heatch,
b_gor_dv,
b_hsbeds,
b_hsrooms,
b_hhsize,
b_nkids015, # no. children in hhold (<16)
b_fihhmngrs_dv, # gross hh income month before interview
b_xphsdb, # problems paying housing costs
b_xpmg, # last mortgage payment
b_rent, # last rent payment £
b_rentwc, # weeks covered b_rent
b_xpaltob_g3 # £ spent on alcohol
)
hha <- readr::read_tsv("inst/extdata/restricted_us/a_hhresp.tab",
na = c("-11", "-10", "-9", "-8", "-7", "-2", "-1")) %>%
select(a_hidp,
a_hsownd, a_tenure_dv,
a_ncars,
a_heatch,
a_gor_dv,
a_hsbeds,
a_hsrooms,
a_hhsize,
a_numchild, # no. children in hhold
a_fihhmngrs_dv, # gross hh income month before interview
a_rent, # last rent payment £
a_rentwc, # weeks covered last rent payment
a_xpmg, # last mortgage payment
a_xphsdb, # problems paying housing costs
a_xpaltob_g3 # £ spent on alcohol
)
stopifnot(
!(any(hhe == -9, na.rm = TRUE)),
!(any(hhd == -9, na.rm = TRUE)),
!(any(hhc == -9, na.rm = TRUE)),
!(any(hhb == -9, na.rm = TRUE)),
!(any(hha == -9, na.rm = TRUE))
)
# Merge household responses ====
us <- us %>%
left_join(hhf, by = "f_hidp") %>%
left_join(hhe, by = "e_hidp") %>%
left_join(hhd, by = "d_hidp") %>%
left_join(hhc, by = "c_hidp") %>%
left_join(hhb, by = "b_hidp") %>%
left_join(hha, by = "a_hidp")
rm(hhf, hhe, hhd, hhc, hhb, hha)
# # Nurse health assessments ====
#
# nhac <- readr::read_tsv("inst/extdata/restricted_us_nha/c_indresp_ns.tab")
#
# nhac <- select(nhac, -c_waist3) # All NA
#
# nhac[nhac == -9] <- NA # missing
# nhac[nhac == -8] <- NA # Inapplicable
# nhac[nhac == -7] <- NA # Proxy
# nhac[nhac == -2] <- NA # Refused
# nhac[nhac == -1] <- NA # Don't know
#
# stopifnot(
# !(any(nhac == -9, na.rm = TRUE))
# )
#
# nhab <- readr::read_tsv("inst/extdata/restricted_us_nha/b_indresp_ns.tab",
# col_types = cols(
# b_wtpc = col_double()
# )
# )
#
# nhab[nhab == -9] <- NA # missing
# nhab[nhab == -8] <- NA # Inapplicable
# nhab[nhab == -7] <- NA # Proxy
# nhab[nhab == -2] <- NA # Refused
# nhab[nhab == -1] <- NA # Don't know
#
# stopifnot(
# !(any(nhab == -9, na.rm = TRUE))
# )
#
#
# # Merge nurse health assessments ====
#
# us <- left_join(us, nhac, by = "pidp")
# us <- left_join(us, nhab, by = "pidp")
#
# rm(nhac, nhab)
# Test ====
stopifnot( # make sure variables from all waves are present
any(grepl("a_", colnames(us))),
any(grepl("b_", colnames(us))),
any(grepl("c_", colnames(us))),
any(grepl("d_", colnames(us))),
any(grepl("e_", colnames(us))),
any(grepl("f_", colnames(us))),
any(grepl("hh", colnames(us)))
)
# Code region ====
us$region <- NA
us$region <- us$f_gor_dv
us$region[is.na(us$region)] <- us$e_gor_dv[is.na(us$region)]
us$region[is.na(us$region)] <- us$d_gor_dv[is.na(us$region)]
us$region[is.na(us$region)] <- us$c_gor_dv[is.na(us$region)]
us$region[is.na(us$region)] <- us$b_gor_dv[is.na(us$region)]
us$region[is.na(us$region)] <- us$a_gor_dv[is.na(us$region)]
us$region <- factor(us$region, levels = 1:12,
labels = c("NE", "NW", "YH", "EM", "WM", "EE",
"LN", "SE", "SW", "WA", "SC", "NI"),
ordered = FALSE)
us <- us %>%
select(-f_gor_dv, -e_gor_dv, -d_gor_dv, -c_gor_dv, -b_gor_dv, -a_gor_dv)
# Code respondent origin ====
# Was going to remove BHPS sample members because of survey fatigue
# us$origin <- NA
# us$origin <- us$f_hhorig
# us$origin[is.na(us$origin)] <- us$e_hhorig[is.na(us$origin)]
# us$origin[is.na(us$origin)] <- us$d_hhorig[is.na(us$origin)]
# us$origin[is.na(us$origin)] <- us$c_hhorig[is.na(us$origin)]
# us$origin[is.na(us$origin)] <- us$b_hhorig[is.na(us$origin)]
# us$origin[is.na(us$origin)] <- us$a_hhorig[is.na(us$origin)]
# us <- us %>%
# filter(origin != 3) %>%
# filter(origin != 4) %>%
# filter(origin != 5) %>%
# filter(origin != 6)
us <- us %>%
select(-f_hhorig, -e_hhorig, -d_hhorig, -c_hhorig, -b_hhorig, -a_hhorig)
# Recode age ====
# cut us$c_dvage to match census
us$age_full <- NA
us$age_full <- us$f_dvage
us$age_full[is.na(us$age_full)] <- us$e_dvage[is.na(us$age_full)] + 1
us$age_full[is.na(us$age_full)] <- us$d_dvage[is.na(us$age_full)] + 2
us$age_full[is.na(us$age_full)] <- us$c_dvage[is.na(us$age_full)] + 3
us$age_full[is.na(us$age_full)] <- us$b_dvage[is.na(us$age_full)] + 4
us$age_full[is.na(us$age_full)] <- us$a_dvage[is.na(us$age_full)] + 5
us$age_full[is.na(us$age_full)] <- round(mean(us$age_full, na.rm = TRUE))
us$age <- cut(us$age_full,
breaks = c(min(us$age_full),
17, 19, 24, 29, 44, 59, 64, 74, 84, 89,
max(us$age_full)),
include.lowest = TRUE,
labels = c("age_16_17", "age_18_19", "age_20_24",
"age_25_29", "age_30_44", "age_45_59",
"age_60_64", "age_65_74", "age_75_84",
"age_85_89", "age_90_plus"))
stopifnot(
all(!is.na(us$age))
)
us <- us %>%
select(-f_dvage, -e_dvage, -d_dvage, -c_dvage, -b_dvage, -a_dvage)
# Recode sex ====
us$sex <- NA
us$sex <- us$f_sex
us$sex[is.na(us$sex)] <- us$e_sex[is.na(us$sex)]
us$sex[is.na(us$sex)] <- us$d_sex[is.na(us$sex)]
us$sex[is.na(us$sex)] <- us$c_sex[is.na(us$sex)]
us$sex[is.na(us$sex)] <- us$b_sex[is.na(us$sex)]
us$sex[is.na(us$sex)] <- us$a_sex[is.na(us$sex)]
us$sex <- factor(us$sex,
levels = 1:2,
labels = c("sex_male", "sex_female"))
us$sex <- forcats::fct_relevel(us$sex, "sex_female", "sex_male")
us <- us %>%
select(-f_sex, -e_sex, -d_sex, -c_sex, -b_sex, -a_sex)
# Recode ethnicity ====
us$eth_full <- NA
us$eth_full <- us$f_racel
us$eth_full[is.na(us$eth_full)] <- us$e_racel[is.na(us$eth_full)]
us$eth_full[is.na(us$eth_full)] <- us$d_racel[is.na(us$eth_full)]
us$eth_full[is.na(us$eth_full)] <- us$c_racel[is.na(us$eth_full)]
us$eth_full[is.na(us$eth_full)] <- us$b_racel[is.na(us$eth_full)]
us$eth_full[is.na(us$eth_full)] <- us$a_racel[is.na(us$eth_full)]
us$eth_full <- factor(us$eth_full, levels = c(1:17, 97),
labels = c("White British",
"Irish",
"Gypsy or Irish Traveller",
"other White",
"White and Black Caribbean",
"White and Black African",
"White and Asian",
"other mixed",
"Indian",
"Pakistani",
"Bangladeshi",
"Chinese",
"other Asian",
"Caribbean",
"African",
"other black",
"Arab",
"other ethnic group"),
ordered = FALSE)
us <- us %>%
select(-f_racel, -e_racel, -d_racel, -c_racel, -b_racel, -a_racel)
# Code ethnic minority group
us$eth_minority <- NA
us$eth_minority[us$eth_full == "White British"] <- FALSE
us$eth_minority[us$eth_full != "White British"] <- TRUE
# Code reduced ethnic groups
# Reduced ethnic group classes - to match census aggregate files
us$eth <- NA
us$eth[us$eth_full == "White British"] <- 1
us$eth[us$eth_full == "Irish"] <- 2
us$eth[us$eth_full == "Gypsy or Irish Traveller"] <- 3
us$eth[us$eth_full == "other White"] <- 3
us$eth[us$eth_full == "White and Black Caribbean"] <- 4
us$eth[us$eth_full == "White and Black African"] <- 4
us$eth[us$eth_full == "White and Asian"] <- 4
us$eth[us$eth_full == "other mixed"] <- 4
us$eth[us$eth_full == "Indian"] <- 5
us$eth[us$eth_full == "Pakistani"] <- 5
us$eth[us$eth_full == "Bangladeshi"] <- 5
us$eth[us$eth_full == "Chinese"] <- 5
us$eth[us$eth_full == "other Asian"] <- 5
us$eth[us$eth_full == "Caribbean"] <- 6
us$eth[us$eth_full == "African"] <- 6
us$eth[us$eth_full == "other black"] <- 6
us$eth[us$eth_full == "Arab"] <- 7
us$eth[us$eth_full == "other ethnic group"] <- 7
us$eth <- factor(us$eth,
levels = 1:7,
labels = c("eth_british",
"eth_irish",
"eth_other_white",
"eth_mixed_multiple_ethnic",
"eth_asian_asian_british",
"eth_black_african_caribbean_british",
"eth_other_ethnicity"),
ordered = FALSE)
# Recode education ====
# Create most recent known hiqual
us$tmp <- NA
us$tmp <- us$f_hiqual_dv
us$tmp[is.na(us$tmp)] <- us$e_hiqual_dv[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$d_hiqual_dv[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$c_hiqual_dv[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$b_hiqual_dv[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$a_hiqual_dv[is.na(us$tmp)]
us$qual <- NA
us$qual[us$tmp == 9] <- 0 # No qualification
# GCSE A*-C Level 2; GCSE D-G Level 1. US not split, so all GCSEs L2
us$qual[us$tmp == 4] <- 2 # L1 or L2: GCSE
us$qual[us$tmp == 3] <- 3 # L3: AS/A Level
us$qual[us$tmp == 2] <- 4 # L4: Degree or above
us$qual[us$tmp == 1] <- 4 # L4: Degree or above
us$qual[us$tmp == 5] <- 5 # other qualification
us$qual <- factor(us$qual,
levels = c(0, 2:5),
labels = c("qual_0",
"qual_2",
"qual_3",
"qual_4_plus",
"qual_other"),
ordered = FALSE)
us <- us %>%
select(-tmp,
-f_hiqual_dv, -e_hiqual_dv, -d_hiqual_dv, -c_hiqual_dv, -b_hiqual_dv,
-a_hiqual_dv)
# Recode car or van availability ====
us$car <- NA
us$car <- us$f_ncars
us$car[is.na(us$car)] <- us$e_ncars[is.na(us$car)]
us$car[is.na(us$car)] <- us$d_ncars[is.na(us$car)]
us$car[is.na(us$car)] <- us$c_ncars[is.na(us$car)]
us$car[is.na(us$car)] <- us$b_ncars[is.na(us$car)]
us$car[is.na(us$car)] <- us$a_ncars[is.na(us$car)]
us$car <- replace(us$car, us$car > 2 & !is.na(us$car), 2)
us$car <- factor(us$car, levels = 0:2,
labels = c("car_0", "car_1", "car_2_plus"),
ordered = FALSE)
us <- us %>%
select(-f_ncars, -e_ncars, -d_ncars, -c_ncars, -b_ncars, -a_ncars)
# Recode heat ====
us$heat <- NA
us$heat <- us$f_heatch
us$heat[is.na(us$heat)] <- us$e_heatch[is.na(us$heat)]
us$heat[is.na(us$heat)] <- us$d_heatch[is.na(us$heat)]
us$heat[is.na(us$heat)] <- us$c_heatch[is.na(us$heat)]
us$heat[is.na(us$heat)] <- us$b_heatch[is.na(us$heat)]
us$heat[is.na(us$heat)] <- us$a_heatch[is.na(us$heat)]
us$heat[us$heat == 2] <- 0
us$heat <- factor(us$heat, levels = 0:1,
labels = c("heat_no", "heat_yes"),
ordered = FALSE)
us <- us %>%
select(-f_heatch, -e_heatch, -d_heatch, -c_heatch, -b_heatch, -a_heatch)
# Recode tenure ====
us$ten_full <- NA
us$ten_full <- us$f_hsownd
us$ten_full[is.na(us$ten_full)] <- us$e_hsownd[is.na(us$ten_full)]
us$ten_full[is.na(us$ten_full)] <- us$d_hsownd[is.na(us$ten_full)]
us$ten_full[is.na(us$ten_full)] <- us$c_hsownd[is.na(us$ten_full)]
us$ten_full[is.na(us$ten_full)] <- us$b_hsownd[is.na(us$ten_full)]
us$ten_full[is.na(us$ten_full)] <- us$a_hsownd[is.na(us$ten_full)]
us$ten_full <- factor(us$ten_full, levels = c(1:5, 97),
labels = c("owned_outright",
"owned_mortgage_loan",
"shared_ownership",
"rented",
"living_rent_free",
"other_tenure"),
ordered = FALSE)
# Reduced categories of tenure to match census aggregates
us$ten <- NA
us$ten[us$ten_full == "owned_outright"] <- 1
us$ten[us$ten_full == "owned_mortgage_loan"] <- 2
us$ten[us$ten_full == "shared_ownership"] <- 2
us$ten[us$ten_full == "rented"] <- 3
us$ten[us$ten_full == "living_rent_free"] <- 3
us$ten[us$ten_full == "other_tenure"] <- NA
us$ten <- factor(us$ten,
levels = 1:3,
labels = c("ten_owned_outright",
"ten_owned_mortgage_shared",
"ten_rented"),
ordered = FALSE)
us <- us %>%
select(-f_tenure_dv, -e_tenure_dv, -d_tenure_dv, -c_tenure_dv,
-b_tenure_dv, -a_tenure_dv) %>%
select(-f_hsownd, -e_hsownd, -d_hsownd, -c_hsownd, -b_hsownd, -a_hsownd)
# Recode economic activity ====
us$econ_act <- NA
us$econ_act <- us$f_jbstat
us$econ_act[is.na(us$econ_act)] <- us$e_jbstat[is.na(us$econ_act)]
us$econ_act[is.na(us$econ_act)] <- us$d_jbstat[is.na(us$econ_act)]
us$econ_act[is.na(us$econ_act)] <- us$c_jbstat[is.na(us$econ_act)]
us$econ_act[is.na(us$econ_act)] <- us$b_jbstat[is.na(us$econ_act)]
us$econ_act[is.na(us$econ_act)] <- us$a_jbstat[is.na(us$econ_act)]
# levels checked as correct 2018-01-12
us$econ_act <- factor(us$econ_act, levels = c(1:11, 97),
labels = c("self_emp",
"emp",
"unemp",
"retired",
"maternity",
"home_fam",
"stu_ft",
"lt_sick",
"gment_train",
"unpaid_fam",
"apprent",
"other"))
# Reaggregate some variables to match census
us$econ_act <- us$econ_act %>%
forcats::fct_collapse(home_fam = c("home_fam", "maternity"),
other = c("other", "gment_train", "unpaid_fam",
"apprent"))
us$econ_act <- us$econ_act %>%
forcats::fct_recode(
eca_emp = "emp",
eca_homefam = "home_fam",
eca_ltsick = "lt_sick",
eca_other = "other",
eca_retired = "retired",
eca_selfemp = "self_emp",
eca_student = "stu_ft",
eca_unemp = "unemp"
) %>%
forcats::fct_relevel( # order alphabetically for simulation
"eca_emp", "eca_homefam", "eca_ltsick", "eca_other", "eca_retired",
"eca_selfemp", "eca_student", "eca_unemp"
)
us <- us %>%
select(-f_jbstat, -e_jbstat, -d_jbstat, -c_jbstat, -b_jbstat, -a_jbstat)
# Overcrowding (> 1 persons per room) ====
us$rooms <- NA
# add bedrooms + other accommodation rooms
us$rooms <- us$f_hsbeds + us$f_hsrooms
# make sure bedrooms and other rooms are from the same wave, i.e. the
# same property
# e.g. e_hsrooms + d_hsbeds might be two different properties
us$rooms[is.na(us$rooms)] <-
us$e_hsbeds[is.na(us$rooms)] + us$e_hsrooms[is.na(us$rooms)]
us$rooms[is.na(us$rooms)] <-
us$d_hsbeds[is.na(us$rooms)] + us$d_hsrooms[is.na(us$rooms)]
us$rooms[is.na(us$rooms)] <-
us$c_hsbeds[is.na(us$rooms)] + us$c_hsrooms[is.na(us$rooms)]
us$rooms[is.na(us$rooms)] <-
us$b_hsbeds[is.na(us$rooms)] + us$b_hsrooms[is.na(us$rooms)]
us$rooms[is.na(us$rooms)] <-
us$a_hsbeds[is.na(us$rooms)] + us$a_hsrooms[is.na(us$rooms)]
us$ppr <- NA
us$ppr <- us$f_hhsize
us$ppr[is.na(us$ppr)] <- us$e_hhsize[is.na(us$ppr)]
us$ppr[is.na(us$ppr)] <- us$d_hhsize[is.na(us$ppr)]
us$ppr[is.na(us$ppr)] <- us$c_hhsize[is.na(us$ppr)]
us$ppr[is.na(us$ppr)] <- us$b_hhsize[is.na(us$ppr)]
us$ppr[is.na(us$ppr)] <- us$a_hhsize[is.na(us$ppr)]
us$ppr <- us$ppr / us$rooms
us$ocrowd <- NA
us$ocrowd[us$ppr > 1] <- 1
us$ocrowd[us$ppr <= 1] <- 0
us$ocrowd <- factor(us$ocrowd, levels = c(0, 1),
labels = c("ocrowd_no",
"ocrowd_yes"),
ordered = FALSE)
stopifnot( # check ocrowd coded properly
all.equal(nrow(filter(us, ppr > 1, ocrowd == "ocrowd_no")), 0),
all.equal(nrow(filter(us, ppr <= 1, ocrowd == "ocrowd_yes")), 0)
)
# hhsize needed for alcohol consumption
us <- us %>%
select(-rooms, -ppr) %>%
select(-f_hsbeds, -e_hsbeds, -d_hsbeds, -c_hsbeds, -b_hsbeds, -a_hsbeds) %>%
select(-f_hsrooms, -e_hsrooms, -d_hsrooms, -c_hsrooms, -b_hsrooms,
-a_hsrooms)
# Limiting long-term illness or disability ====
us$llid <- NA
us$llid <- us$f_health
us$llid[is.na(us$llid)] <- us$e_health[is.na(us$llid)]
us$llid[is.na(us$llid)] <- us$d_health[is.na(us$llid)]
us$llid[is.na(us$llid)] <- us$c_health[is.na(us$llid)]
us$llid[is.na(us$llid)] <- us$b_health[is.na(us$llid)]
us$llid[is.na(us$llid)] <- us$a_health[is.na(us$llid)]
us$llid <- replace(us$llid, us$llid == 2, 0)
us$llid <- factor(us$llid, levels = 0:1,
labels = c("llid_no", "llid_yes"))
us <- us %>%
select(-f_health, -e_health, -d_health, -c_health, -b_health, -a_health)
# Recode self-reported general health ====
us$tmp <- NA
us$tmp <- us$f_scsf1
us$tmp[is.na(us$tmp)] <- us$e_scsf1[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$d_scsf1[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$c_sf1[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$b_sf1[is.na(us$tmp)]
us$tmp[is.na(us$tmp)] <- us$a_sf1[is.na(us$tmp)]
us$srgh <- NA
us$srgh[us$tmp == 1] <- 5 # excellent
us$srgh[us$tmp == 2] <- 4 # very good
us$srgh[us$tmp == 3] <- 3 # good
us$srgh[us$tmp == 4] <- 2 # fair
us$srgh[us$tmp == 5] <- 1 # poor
us$srgh <- factor(us$srgh, levels = 5:1,
labels = c("srgh_excellent",
"srgh_verygood",
"srgh_good",
"srgh_fair",
"srgh_poor"),
ordered = FALSE)
us <- us %>%
select(-f_scsf1, -e_scsf1, -d_scsf1, -c_sf1, -b_sf1, -a_sf1,
-tmp)
# Recode social class ====
us$class8 <- NA
us$class8 <- us$f_jbnssec8_dv
us$class8[is.na(us$class8)] <- us$e_jbnssec8_dv[is.na(us$class8)]
us$class8[is.na(us$class8)] <- us$d_jbnssec8_dv[is.na(us$class8)]
us$class8[is.na(us$class8)] <- us$c_jbnssec8_dv[is.na(us$class8)]
us$class8[is.na(us$class8)] <- us$b_jbnssec8_dv[is.na(us$class8)]
us$class8[is.na(us$class8)] <- us$a_jbnssec8_dv[is.na(us$class8)]
# Full NS-SEC, NS-SEC5 and NS-SEC3 have same number of NAs as NSSEC8
# us$class <- NA
# us$class <- us$f_jbnssec_dv
# us$class[is.na(us$class)] <- us$e_jbnssec_dv[is.na(us$class)]
# us$class[is.na(us$class)] <- us$d_jbnssec_dv[is.na(us$class)]
# us$class[is.na(us$class)] <- us$c_jbnssec_dv[is.na(us$class)]
# us$class[is.na(us$class)] <- us$b_jbnssec_dv[is.na(us$class)]
# us$class[is.na(us$class)] <- us$a_jbnssec_dv[is.na(us$class)]
#
# us$class5 <- NA
# us$class5 <- us$f_jbnssec5_dv
# us$class5[is.na(us$class5)] <- us$e_jbnssec5_dv[is.na(us$class5)]
# us$class5[is.na(us$class5)] <- us$d_jbnssec5_dv[is.na(us$class5)]
# us$class5[is.na(us$class5)] <- us$c_jbnssec5_dv[is.na(us$class5)]
# us$class5[is.na(us$class5)] <- us$b_jbnssec5_dv[is.na(us$class5)]
# us$class5[is.na(us$class5)] <- us$a_jbnssec5_dv[is.na(us$class5)]
#
# us$class3 <- NA
# us$class3 <- us$f_jbnssec3_dv
# us$class3[is.na(us$class3)] <- us$e_jbnssec3_dv[is.na(us$class3)]
# us$class3[is.na(us$class3)] <- us$d_jbnssec3_dv[is.na(us$class3)]
# us$class3[is.na(us$class3)] <- us$c_jbnssec3_dv[is.na(us$class3)]
# us$class3[is.na(us$class3)] <- us$b_jbnssec3_dv[is.na(us$class3)]
# us$class3[is.na(us$class3)] <- us$a_jbnssec3_dv[is.na(us$class3)]
us$class8 <- factor(us$class8, levels = 1:9,
labels = c("Large employers & higher management",
"Higher professional",
"Lower management & professional",
"Intermediate",
"Small employers & own account",
"Lower supervisory & technical",
"Semi-routine",
"Routine",
"Never worked & LT unemployed"),
ordered = FALSE)
us <- us %>%
select(-f_jbnssec_dv, -f_jbnssec8_dv, -f_jbnssec5_dv, -f_jbnssec3_dv) %>%
select(-e_jbnssec_dv, -e_jbnssec8_dv, -e_jbnssec5_dv, -e_jbnssec3_dv) %>%
select(-d_jbnssec_dv, -d_jbnssec8_dv, -d_jbnssec5_dv, -d_jbnssec3_dv) %>%
select(-c_jbnssec_dv, -c_jbnssec8_dv, -c_jbnssec5_dv, -c_jbnssec3_dv) %>%
select(-b_jbnssec_dv, -b_jbnssec8_dv, -b_jbnssec5_dv, -b_jbnssec3_dv) %>%
select(-a_jbnssec_dv, -a_jbnssec8_dv, -a_jbnssec5_dv, -a_jbnssec3_dv)
# # Tax credits ====
# us$taxcred <- NA
# us$taxcred <- us$f_btype6
# us$taxcred[is.na(us$taxcred)] <- us$e_btype6[is.na(us$taxcred)]
# us$taxcred[is.na(us$taxcred)] <- us$d_btype6[is.na(us$taxcred)]
# us$taxcred[is.na(us$taxcred)] <- us$c_btype6[is.na(us$taxcred)]
# us$taxcred[is.na(us$taxcred)] <- us$b_btype6[is.na(us$taxcred)]
# us$taxcred[is.na(us$taxcred)] <- us$a_btype6[is.na(us$taxcred)]
# Marital status ====
us$marital <- NA
us$marital <- us$f_mlstat
us$marital[is.na(us$marital)] <- us$e_mlstat[is.na(us$marital)]
us$marital[is.na(us$marital)] <- us$d_mlstat[is.na(us$marital)]
us$marital[is.na(us$marital)] <- us$c_mlstat[is.na(us$marital)]
us$marital[is.na(us$marital)] <- us$b_mlstat[is.na(us$marital)]
us$marital[is.na(us$marital)] <- us$a_mlstat[is.na(us$marital)]
# checked these levels are correct 2018-01-18
us$marital <- factor(us$marital, levels = 1:9,
labels = c("single",
"married",
"civil_part",
"separated",
"divorced",
"widowed",
"sep_civil_part",
"former_civil_part",
"surviving_civil_part")) %>%
forcats::fct_collapse(
separated = c("separated", "sep_civil_part"),
divorced = c("divorced", "former_civil_part"),
widowed = c("widowed", "surviving_civil_part")) %>%
forcats::fct_recode(
mar_civil_part = "civil_part",
mar_divorced = "divorced",
mar_married = "married",
mar_separated = "separated",
mar_single = "single",
mar_widowed = "widowed"
) %>%
forcats::fct_relevel( # make alphabetical for simulation
"mar_civil_part", "mar_divorced", "mar_married", "mar_separated",
"mar_single", "mar_widowed"
)
us <- us %>%
select(-f_mlstat, -e_mlstat, -d_mlstat, -c_mlstat, -b_mlstat, -a_mlstat)
# Couple ====
# This is couple, married or not
# In waves f_, d_ and b_ asked about domestic chores
# if there's any answer (or they're married) they're part of a couple
# I have to assume the remaining NAs are single
us$couple <- NA
us$couple[us$f_hubuys > 0 & !is.na(us$f_hubuys)] <- 1
us$couple[is.na(us$couple) & us$marital == "mar_married"] <- 1
us$couple[is.na(us$couple) & us$d_hubuys > 0 & !is.na(us$d_hubuys)] <- 1
us$couple[is.na(us$couple) & us$b_hubuys > 0 & !is.na(us$b_hubuys)] <- 1
us <- us %>%
select(-f_hubuys, -d_hubuys, -b_hubuys)
# Neighbourhood cohesion ====
# Paper 13 neighbourhood cohesion:
# extent to which neighbours pull together to improve neighbourhood
# No direct analogy but proxy:
# willing to improve neighbourhood, and
# similar to others in neighbourhood
# Willing to improve neighbourhood
us$imp_nhood <- NA
us$imp_nhood <- us$f_scopngbhe
us$imp_nhood[is.na(us$imp_nhood)] <- us$c_scopngbhe[is.na(us$imp_nhood)]
us$imp_nhood[is.na(us$imp_nhood)] <- us$a_scopngbhe[is.na(us$imp_nhood)]
# Similar to others in neighbourhood
us$sim_nhood <- NA
us$sim_nhood <- us$f_scopngbhg
us$sim_nhood[is.na(us$sim_nhood)] <- us$c_scopngbhg[is.na(us$sim_nhood)]
us$sim_nhood[is.na(us$sim_nhood)] <- us$a_scopngbhg[is.na(us$sim_nhood)]
# dplyr::case_when() doesn't yet work inside mutate()
# <= 2 == strongly agree + agree
# >= 3 == Neither + disagree + strongly disagree
us <- us %>%
mutate(nhood_coh = if_else(imp_nhood <= 2 & sim_nhood <= 2,
"ncoh_yes", "ncoh_no"))
stopifnot(
# more agree + strongly agree than cohesive neighbourhoods
nrow(us[us$imp_nhood <= 2 & !is.na(us$imp_nhood), ]) >
nrow(us[us$nhood_coh == "ncoh_yes" & !is.na(us$nhood_coh), ]),
nrow(us[us$sim_nhood <= 2 & !is.na(us$sim_nhood), ]) >
nrow(us[us$nhood_coh == "ncoh_yes" & !is.na(us$nhood_coh), ])
)
us <- us %>%
select(-imp_nhood, -sim_nhood,
-f_scopngbhe, -c_scopngbhe, -a_scopngbhe,
-f_scopngbhg, -c_scopngbhg, -a_scopngbhg)
# Neighbourhood trust ====
# Waves f and c are different to how wave a is coded
# Wave f and c: strongly agree + agree == trust
# wave a: most people can be trusted == trust
us$nhood_trust <- NA
us$nhood_trust <- if_else(us$f_nbrcoh3 <= 2, "trust_yes", "trust_no")
us$nhood_trust[is.na(us$nhood_trust)] <-
if_else(us$c_nbrcoh3[is.na(us$nhood_trust)] <= 2,
"trust_yes",
"trust_no")
us$nhood_trust[is.na(us$nhood_trust)] <-
if_else(us$a_sctrust[is.na(us$nhood_trust)] == 1,
"trust_yes",
"trust_no")
us <- us %>%
select(-f_nbrcoh3, -c_nbrcoh3, -a_sctrust)
# Neighbourhood belonging ====
us$nhood_belong <- NA
us$nhood_belong <- us$f_scopngbha
us$nhood_belong[is.na(us$nhood_belong)] <-
us$c_scopngbha[is.na(us$nhood_belong)]
us$nhood_belong[is.na(us$nhood_belong)] <-
us$a_scopngbha[is.na(us$nhood_belong)]
us$nhood_belong <- if_else(us$nhood_belong <= 2, "belong_yes", "belong_no")
us <- us %>%
select(-f_scopngbha, -c_scopngbha, -a_scopngbha)
# Civic participation ====
civic_vars <- colnames(us[, grep("f_orga", colnames(us))])
civic_vars <- civic_vars[civic_vars != "f_orga96"] # 'none of these'
us$civic <- NA
us$civic <- rowSums(us[, civic_vars])
civic_vars <- colnames(us[, grep("c_orga", colnames(us))])
civic_vars <- civic_vars[civic_vars != "c_orga96"] # 'none of these'
us$civic[is.na(us$civic)] <- rowSums(us[is.na(us$civic), civic_vars])
us$civic <- if_else(us$civic >= 1, "civic_yes", "civic_no")
us <- us %>%
select(-f_orga1, -f_orga2, -f_orga3, -f_orga4, -f_orga5, -f_orga6,
-f_orga7, -f_orga8, -f_orga9, -f_orga10, -f_orga11, -f_orga12,
-f_orga13, -f_orga14, -f_orga15, -f_orga16,
-c_orga1, -c_orga2, -c_orga3, -c_orga4, -c_orga5, -c_orga6,
-c_orga7, -c_orga8, -c_orga9, -c_orga10, -c_orga11, -c_orga12,
-c_orga13, -c_orga14, -c_orga15, -c_orga16)
rm(civic_vars)
# Social cohesion ====
# People in this neighbourhood get along - paper 13
# People in this neighbourhood DON'T get along - US
# Reverse, so take strongly disagree + disagree
us$social_coh <- NA
us$social_coh <- us$f_nbrcoh4
us$social_coh[is.na(us$social_coh)] <- us$c_nbrcoh4[is.na(us$social_coh)]
us$social_coh <- if_else(us$social_coh >= 4, "scoh_yes", "scoh_no")
us <- us %>%
select(-f_nbrcoh4, -c_nbrcoh4)
# Safe walking alone at night ====
us$safe_alone <- NA
us$safe_alone <- us$f_crdark
us$safe_alone[is.na(us$safe_alone)] <- us$c_crdark[is.na(us$safe_alone)]
# Recode to safe or unsafe
# 1 = very safe
us$safe_alone[us$safe_alone == 2] <- 1 # fairly safe
us$safe_alone[us$safe_alone == 3] <- 2 # a bit unsafe
us$safe_alone[us$safe_alone == 4] <- 2 # very unsafe
us$safe_alone[us$safe_alone == 5] <- 2 # don't go out after dark
us$safe_alone <- factor(us$safe_alone, levels = 1:2,
labels = c("safe_yes",
"safe_no"))
us <- us %>%
select(-f_crdark, -c_crdark)
# Similar income to friends ====
# Paper 13 doesn't specify which levels map to 'hetergeneous' or homogeneous
# friendship. Assume ~ <= 1/2 is hetero
us$sim_inc <- NA
us$sim_inc <- us$f_siminc
us$sim_inc[is.na(us$sim_inc)] <- us$c_siminc[is.na(us$sim_inc)]
us$sim_inc <- if_else(us$sim_inc >= 3, "siminc_het", "siminc_hom")
us <- us %>%
select(-f_siminc, -c_siminc)
# Similiar ethnicity to friends ====
# Paper 13 again doesn't specify which levels map to hetero/homo
# Assume again ~ <= 1/2 hetero; >= 1/2 homo
us$sim_eth <- NA
us$sim_eth <- us$f_simrace
us$sim_eth[is.na(us$sim_eth)] <- us$c_simrace[is.na(us$sim_eth)]
us$sim_eth <- if_else(us$sim_eth >= 3, "simeth_het", "simeth_hom")
us <- us %>%
select(-f_simrace, -c_simrace)
# Political efficacy ====
# Don't have a say in what government does, reversed
us$pol_eff <- NA
us$pol_eff <- us$f_poleff4
us$pol_eff[is.na(us$pol_eff)] <- us$c_poleff4[is.na(us$pol_eff)]
us$pol_eff <- if_else(us$pol_eff >= 4, "poleff_yes", "poleff_no")
us <- us %>%
select(-f_poleff4, -c_poleff4)
# Financial situation ====
us$fin_sit <- NA
us$fin_sit <- us$f_finnow
us$fin_sit[is.na(us$fin_sit)] <- us$e_finnow[is.na(us$fin_sit)]
us$fin_sit[is.na(us$fin_sit)] <- us$d_finnow[is.na(us$fin_sit)]
us$fin_sit[is.na(us$fin_sit)] <- us$c_finnow[is.na(us$fin_sit)]
us$fin_sit[is.na(us$fin_sit)] <- us$b_finnow[is.na(us$fin_sit)]
us$fin_sit[is.na(us$fin_sit)] <- us$a_finnow[is.na(us$fin_sit)]
us$fin_sit <- if_else(us$fin_sit <= 2, "fin_good", "fin_bad")
us <- us %>%
select(-f_finnow, -e_finnow, -d_finnow, -c_finnow, -b_finnow, -a_finnow)
# Social isolation ====
# >=1 close friend
us$soc_isol <- NA
us$soc_isol <- us$f_closenum
us$soc_isol[is.na(us$soc_isol)] <- us$c_closenum[is.na(us$soc_isol)]
us$soc_isol <- if_else(us$soc_isol > 0, "isol_no", "isol_yes")
us <- us %>%
select(-f_closenum, -c_closenum)
# Income ====
# I've tested:
# - total earnings annually (x_prearna)
# - total earnings weekly (x_prearnw)
# - total personal income annually (x_prfitba)
# - total personal income weekly (x_prfitw)
#
# All (even when combined) have low response rates and ~ 77k missing cases
# General Health Questionnaire (GHQ) ====
recode_ghq <- function(varname) {
x <- NA
x <- us[[paste0("f_", varname)]]
x[is.na(x)] <- us[[paste0("f_", varname)]][is.na(x)]
x[is.na(x)] <- us[[paste0("e_", varname)]][is.na(x)]
x[is.na(x)] <- us[[paste0("d_", varname)]][is.na(x)]
x[is.na(x)] <- us[[paste0("c_", varname)]][is.na(x)]
x[is.na(x)] <- us[[paste0("b_", varname)]][is.na(x)]
x[is.na(x)] <- us[[paste0("a_", varname)]][is.na(x)]
x
}
us$ghq_conc <- recode_ghq("scghqa")
us$ghq_sleep <- recode_ghq("scghqb")
us$ghq_useful <- recode_ghq("scghqc")
us$ghq_decision <- recode_ghq("scghqd")
us$ghq_strain <- recode_ghq("scghqe")
us$ghq_diff <- recode_ghq("scghqf")
us$ghq_enjoy <- recode_ghq("scghqg")
us$ghq_face <- recode_ghq("scghqh")
us$ghq_unhappy <- recode_ghq("scghqi")
us$ghq_confid <- recode_ghq("scghqj")
us$ghq_worth <- recode_ghq("scghqk")
us$ghq_happy <- recode_ghq("scghql")
us$ghq_conc <- if_else(us$ghq_conc <= 2,
"ghq_conc_good", "ghq_conc_bad")
us$ghq_sleep <- if_else(us$ghq_sleep <= 2,
"ghq_sleep_good", "ghq_sleep_bad")
us$ghq_useful <- if_else(us$ghq_useful <= 2,
"ghq_useful_good", "ghq_useful_bad")
us$ghq_decision <- if_else(us$ghq_decision <= 2,
"ghq_decision_good", "ghq_decision_bad")
us$ghq_strain <- if_else(us$ghq_strain <= 2,
"ghq_strain_good", "ghq_strain_bad")
us$ghq_diff <- if_else(us$ghq_diff <= 2,
"ghq_diff_good", "ghq_diff_bad")
us$ghq_enjoy <- if_else(us$ghq_enjoy <= 2,
"ghq_enjoy_good", "ghq_enjoy_bad")
us$ghq_face <- if_else(us$ghq_face <= 2,
"ghq_face_good", "ghq_face_bad")
us$ghq_unhappy <- if_else(us$ghq_unhappy <= 2,
"ghq_unhappy_good", "ghq_unhappy_bad")
us$ghq_confid <- if_else(us$ghq_confid <= 2,
"ghq_confid_good", "ghq_confid_bad")
us$ghq_worth <- if_else(us$ghq_worth <= 2,
"ghq_worth_good", "ghq_worth_bad")
us$ghq_happy <- if_else(us$ghq_happy <= 2,
"ghq_happy_good", "ghq_happy_bad")
us <- us %>%
select(-f_scghqa, -e_scghqa, -d_scghqa, -c_scghqa, -b_scghqa, -a_scghqa,
-f_scghqb, -e_scghqb, -d_scghqb, -c_scghqb, -b_scghqb, -a_scghqb,
-f_scghqc, -e_scghqc, -d_scghqc, -c_scghqc, -b_scghqc, -a_scghqc,
-f_scghqd, -e_scghqd, -d_scghqd, -c_scghqd, -b_scghqd, -a_scghqd,
-f_scghqe, -e_scghqe, -d_scghqe, -c_scghqe, -b_scghqe, -a_scghqe,
-f_scghqf, -e_scghqf, -d_scghqf, -c_scghqf, -b_scghqf, -a_scghqf,
-f_scghqg, -e_scghqg, -d_scghqg, -c_scghqg, -b_scghqg, -a_scghqg,
-f_scghqh, -e_scghqh, -d_scghqh, -c_scghqh, -b_scghqh, -a_scghqh,
-f_scghqi, -e_scghqi, -d_scghqi, -c_scghqi, -b_scghqi, -a_scghqi,
-f_scghqj, -e_scghqj, -d_scghqj, -c_scghqj, -b_scghqj, -a_scghqj,
-f_scghqk, -e_scghqk, -d_scghqk, -c_scghqk, -b_scghqk, -a_scghqk,
-f_scghql, -e_scghql, -d_scghql, -c_scghql, -b_scghql, -a_scghql)
rm(recode_ghq)
# Cigarettes smoked ====
us$cig <- NA
us$cig <- us$f_ncigs
us$cig[is.na(us$cig)] <- us$e_ncigs[is.na(us$cig)]
us$cig[is.na(us$cig)] <- us$b_ncigs[is.na(us$cig)]
us$cig <- if_else(us$cig > 0, "cig_yes", "cig_no")
us <- us %>%
select(-f_ncigs, -e_ncigs, -b_ncigs)
# Alcohol ====
# units = (volume (ml) * ABV) / 1000
# 1 pint = 568.261ml (Google)
# Source of alcohol costs: institute alcohol studies
# http://www.ias.org.uk/uploads/pdf/Factsheets/Alcohol%20pricing%20factsheet%20April%202014.pdf
# All dates 2012 data
#
# ABV from NHS.uk/livewell/alcohol/pages/alcohol-units.aspx and IAS
#
# Beer: £2.70 pint 3.6% abv
# Lager: £2.96 pint 5.2% abv
# Cider: £1.82 pint 5.3% abv
# Sherry: £6.55 70cl 17.5% abv
# Wine: £4.04 75cl 12% abv
# Whiskey: £15.35 70cl 40% abv
# Vodka: £13.86 70cl 37.5% abv
units <- data.frame(
beer = c(2.7, 3.6, 568),
lager = c(2.96, 5.2, 568),
cider = c(1.82, 5.3, 568),
sherry = c(6.55, 17.5, 700),
wine = c(4.04, 12.0, 750),
whiskey = c(15.35, 40.0, 700),
vodka = c(13.86, 37.5, 700)
)
units <- as.data.frame(t(units))
colnames(units) <- c("cost", "abv", "ml")
units$units <- (units$ml * units$abv) / 1000
# source: http://www.nhs.uk/Livewell/alcohol/Pages/alcohol-units.aspx
units$unit_cost <- units$units / units$cost
units$unit_cost <- round(units$unit_cost, digits = 2)
unit_cost <- median(units$unit_cost)
us <- us %>%
mutate(f_alc = (f_xpaltob_g3 / (f_hhsize - f_nkids015)) / unit_cost / 4,
e_alc = (e_xpaltob_g3 / (e_hhsize - e_nkids_dv)) / unit_cost / 4,
d_alc = (d_xpaltob_g3 / (d_hhsize - d_nkids015)) / unit_cost / 4,
c_alc = (c_xpaltob_g3 / (c_hhsize - c_nkids015)) / unit_cost / 4,
b_alc = (b_xpaltob_g3 / (b_hhsize - b_nkids015)) / unit_cost / 4,
a_alc = (a_xpaltob_g3 / (a_hhsize - a_numchild)) / unit_cost / 4)
us$alcohol <- NA
us$alcohol <- us$f_alc
us$alcohol[is.na(us$alcohol)] <- us$e_alc[is.na(us$alcohol)]
us$alcohol[is.na(us$alcohol)] <- us$d_alc[is.na(us$alcohol)]
us$alcohol[is.na(us$alcohol)] <- us$c_alc[is.na(us$alcohol)]
us$alcohol[is.na(us$alcohol)] <- us$b_alc[is.na(us$alcohol)]
us$alcohol[is.na(us$alcohol)] <- us$a_alc[is.na(us$alcohol)]
# Low/high risk thresholds
# cmo2016a, p. 4
us$alcohol <- if_else(us$alcohol <= 14, "alcohol_low", "alcohol_high")
us <- us %>%
select(-f_xpaltob_g3, -e_xpaltob_g3, -d_xpaltob_g3,
-c_xpaltob_g3, -b_xpaltob_g3, -a_xpaltob_g3,
-f_alc, -e_alc, -d_alc, -c_alc, -b_alc, -a_alc)
rm(units, unit_cost)
# Best friend distance ====
us <- us %>%
rowwise() %>%
mutate(f_bfdist = min(f_netlv_1, f_netlv_2, f_netlv_3),
c_bfdist = min(c_netlv_1, c_netlv_2, c_netlv_3))
us$bf_dist <- NA
us$bf_dist <- us$f_bfdist
us$bf_dist[is.na(us$bf_dist)] <- us$f_bfdist[is.na(us$bf_dist)]
# Paper: < 5 mins. Use < 1 mile as proxy
us$bf_dist <- if_else(us$bf_dist > 1, "bfdist_far", "bfdist_near")
us <- us %>%
select(-f_bfdist, -c_bfdist,
-f_netlv_1, -f_netlv_2, -f_netlv_3,
-c_netlv_1, -c_netlv_2, -c_netlv_3)
# Best friend often in touch ====
# Paper: frequent contact == at least weekly
us <- us %>%
rowwise() %>%
mutate(f_bfoften = min(f_netph_1, f_netph_2, f_netph_3),
c_bfoften = min(c_netph_1, c_netph_2, c_netph_3))
us$bf_often <- NA
us$bf_often <- us$f_bfoften
us$bf_often[is.na(us$bf_often)] <- us$c_bfoften[is.na(us$bf_often)]
us$bf_often <- if_else(us$bf_often > 1, "bfoften_not", "bfoften_often")
us <- us %>%
select(-f_bfoften, -c_bfoften,
-f_netph_1, -f_netph_2, -f_netph_3,
-c_netph_1, -c_netph_2, -c_netph_3)
# Savings ====
# Budgeting/money management skills
# Too many NAs save_reg
us$save <- NA
us$save <- us$f_save
us$save[is.na(us$save)] <- us$d_save[is.na(us$save)]
us$save[is.na(us$save)] <- us$b_save[is.na(us$save)]
us$save <- if_else(us$save == 1, "save_yes", "save_no")
us <- us %>%
select(-f_save, -d_save, -b_save)
# Number of children ====
us$nkids <- NA
us$nkids <- us$f_nkids015
us$nkids[is.na(us$nkids)] <- us$e_nkids_dv[is.na(us$nkids)]
us$nkids[is.na(us$nkids)] <- us$d_nkids015[is.na(us$nkids)]
us$nkids[is.na(us$nkids)] <- us$c_nkids015[is.na(us$nkids)]
us$nkids[is.na(us$nkids)] <- us$b_nkids015[is.na(us$nkids)]
us$nkids[is.na(us$nkids)] <- us$a_numchild[is.na(us$nkids)]
us <- us %>%
select(-f_nkids015, -e_nkids_dv, -d_nkids015, -c_nkids015, -b_nkids015,
-a_numchild)
# Can't use numeric data with rakeR::extract() easily because it creates a
# new variable for each level of the variable (i.e. one for each unique
# integer)
# Recode and top-code to 5 kids
us$kids <- cut(us$nkids,
breaks = c(0, 1, 2, 3, 4, 5, max(us$nkids, na.rm = TRUE)),
labels = c("kids_0", "kids_1", "kids_2", "kids_3", "kids_4",
"kids_5ormore"),
right = FALSE, include.lowest = TRUE)
# Household income and poverty ====
# Many measures of poverty based on household income (i.e. children do not
# earn but are not in poverty if earning adults aren't)
# US provides monthly hh income. Convert to annual.
# Poverty defined as <60% median income
# Median income in 2011 was £23,200 (source: ons2013c, p. 5)
us$hh_income <- us$f_fihhmngrs_dv
us$hh_income[is.na(us$hh_income)] <- us$e_fihhmngrs_dv[is.na(us$hh_income)]
us$hh_income[is.na(us$hh_income)] <- us$d_fihhmngrs_dv[is.na(us$hh_income)]
us$hh_income[is.na(us$hh_income)] <- us$c_fihhmngrs_dv[is.na(us$hh_income)]
us$hh_income[is.na(us$hh_income)] <- us$b_fihhmngrs_dv[is.na(us$hh_income)]
us$hh_income[is.na(us$hh_income)] <- us$a_fihhmngrs_dv[is.na(us$hh_income)]
# Set negative values to 0
us$hh_income[us$hh_income < 0] <- 0L
us <- us %>%
select(-f_fihhmngrs_dv, -e_fihhmngrs_dv, -d_fihhmngrs_dv, -c_fihhmngrs_dv,
-b_fihhmngrs_dv, -a_fihhmngrs_dv)
# Number of cohabitants ====
us$cohab <- NA
us$cohab <- us$f_hhsize
us$cohab[is.na(us$cohab)] <- us$e_hhsize[is.na(us$cohab)]
us$cohab[is.na(us$cohab)] <- us$d_hhsize[is.na(us$cohab)]
us$cohab[is.na(us$cohab)] <- us$c_hhsize[is.na(us$cohab)]
us$cohab[is.na(us$cohab)] <- us$b_hhsize[is.na(us$cohab)]
us$cohab[is.na(us$cohab)] <- us$a_hhsize[is.na(us$cohab)]
us$cohab <- us$cohab - 1 # remove the respondent!
us <- us %>%
select(-f_hhsize, -e_hhsize, -d_hhsize, -c_hhsize, -b_hhsize, -a_hhsize)
# Clinical depression ====
us$depress <- NA_integer_
hconds <- colnames(us)[grepl("_hcond", colnames(us))]
for (i in hconds) {
us[[i]] <- as.integer(us[[i]])
us$depress[is.na(us$depress) & us[[i]] == 17 & !is.na(us[[i]])] <- 1
}
rm(i)
us$depress[is.na(us$depress)] <- us$f_hcond17[is.na(us$depress)]
us$depress[is.na(us$depress)] <- us$e_hcond17[is.na(us$depress)]
us$depress[is.na(us$depress)] <- us$d_hcond17[is.na(us$depress)]
us$depress[is.na(us$depress)] <- us$c_hcond17[is.na(us$depress)]
us$depress[is.na(us$depress)] <- us$b_hcondn17[is.na(us$depress)]
us$depress[is.na(us$depress)] <- us$a_hcond17[is.na(us$depress)]
us$depress <- factor(us$depress, levels = 0:1,
labels = c("depress_no", "depress_yes"),
ordered = FALSE)
# IHD and stroke ====
# Air pollution is associated with IHD and stroke (among others)
# This ihd measure therefore includes IND and stroke
# IHD includes angina and myocardial infarction (MI)
us$ihd <- NA_integer_
for (i in hconds) {
us[[i]] <- as.integer(us[[i]])
us$ihd[is.na(us$ihd) & !is.na(us[[i]]) &
us[[i]] == 4 | # coronary heart disease
us[[i]] == 5 | # angina
us[[i]] == 6 | # myocardial infarction
us[[i]] == 7] <- 1 # stroke
}
us$ihd[is.na(us$ihd)] <- 0
us$ihd <- factor(us$ihd, levels = 0:1,
labels = c("ihd_no", "ihd_yes"))
us <- us %>%
select(-f_hcond17, -e_hcond17, -d_hcond17, -c_hcond17, -b_hcondn17,
-a_hcond17)
us <- us %>%
select(-f_hcondno1:-f_hcondno8) %>%
select(-e_hcondno1:-e_hcondno7) %>%
select(-d_hcondno1:-d_hcondno10) %>%
select(-c_hcondno1:-c_hcondno8) %>%
select(-b_hcondno1:-b_hcondno8)
# Problems paying housing costs ====
us$prob_pay <- us$f_xphsdb
us$prob_pay[is.na(us$prob_pay)] <- us$e_xphsdb[is.na(us$prob_pay)]
us$prob_pay[is.na(us$prob_pay)] <- us$d_xphsdb[is.na(us$prob_pay)]
us$prob_pay[is.na(us$prob_pay)] <- us$c_xphsdb[is.na(us$prob_pay)]
us$prob_pay[is.na(us$prob_pay)] <- us$b_xphsdb[is.na(us$prob_pay)]
us$prob_pay[is.na(us$prob_pay)] <- us$a_xphsdb[is.na(us$prob_pay)]
us <- us %>%
select(-f_xphsdb, -e_xphsdb, -d_xphsdb, -c_xphsdb, -b_xphsdb, -a_xphsdb)
us$prob_pay <- factor(us$prob_pay, levels = 1:2,
labels = c("prob_pay_yes", "prob_pay_no"))
# Housing costs ====
# household income is based on monthly, so standardise on monthly
us$f_rent[us$f_rentwc == 1 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 1 & !is.na(us$f_rentwc)] * 52 / 12
us$f_rent[us$f_rentwc == 2 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 2 & !is.na(us$f_rentwc)] * 52 / 2 / 12
us$f_rent[us$f_rentwc == 3 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 3 & !is.na(us$f_rentwc)] * 52 / 3 / 12
us$f_rent[us$f_rentwc == 4 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 4 & !is.na(us$f_rentwc)] * 52 / 4 / 12
us$f_rent[us$f_rentwc == 7 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 7 & !is.na(us$f_rentwc)] / 2
us$f_rent[us$f_rentwc == 8 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 8 & !is.na(us$f_rentwc)] * 8 / 12
us$f_rent[us$f_rentwc == 9 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 9 & !is.na(us$f_rentwc)] * 9 / 12
us$f_rent[us$f_rentwc == 10 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 10 & !is.na(us$f_rentwc)] * 10 / 12
us$f_rent[us$f_rentwc == 13 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 13 & !is.na(us$f_rentwc)] / 3
us$f_rent[us$f_rentwc == 26 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 26 & !is.na(us$f_rentwc)] / 6
us$f_rent[us$f_rentwc == 52 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 52 & !is.na(us$f_rentwc)] / 12
us$f_rent[us$f_rentwc == 90 & !is.na(us$f_rentwc)] <-
us$f_rent[us$f_rentwc == 90 & !is.na(us$f_rentwc)] * 52 / 12
us$f_rent[us$f_rentwc == 95 & !is.na(us$f_rentwc)] <- NA # one-off payment
us$f_rent[us$f_rentwc == 96 & !is.na(us$f_rentwc)] <- NA # none of these
us$e_rent[us$e_rentwc == 1 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 1 & !is.na(us$e_rentwc)] * 52 / 12
us$e_rent[us$e_rentwc == 2 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 2 & !is.na(us$e_rentwc)] * 52 / 2 / 12
us$e_rent[us$e_rentwc == 3 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 3 & !is.na(us$e_rentwc)] * 52 / 3 / 12
us$e_rent[us$e_rentwc == 4 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 4 & !is.na(us$e_rentwc)] * 52 / 4 / 12
us$e_rent[us$e_rentwc == 7 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 7 & !is.na(us$e_rentwc)] / 2
us$e_rent[us$e_rentwc == 8 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 8 & !is.na(us$e_rentwc)] * 8 / 12
us$e_rent[us$e_rentwc == 9 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 9 & !is.na(us$e_rentwc)] * 9 / 12
us$e_rent[us$e_rentwc == 10 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 10 & !is.na(us$e_rentwc)] * 10 / 12
us$e_rent[us$e_rentwc == 13 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 13 & !is.na(us$e_rentwc)] / 3
us$e_rent[us$e_rentwc == 26 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 26 & !is.na(us$e_rentwc)] / 6
us$e_rent[us$e_rentwc == 52 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 52 & !is.na(us$e_rentwc)] / 12
us$e_rent[us$e_rentwc == 90 & !is.na(us$e_rentwc)] <-
us$e_rent[us$e_rentwc == 90 & !is.na(us$e_rentwc)] * 52 / 12
us$e_rent[us$e_rentwc == 95 & !is.na(us$e_rentwc)] <- NA # one-off payment
us$e_rent[us$e_rentwc == 96 & !is.na(us$e_rentwc)] <- NA # none of these
us$d_rent[us$d_rentwc == 1 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 1 & !is.na(us$d_rentwc)] * 52 / 12
us$d_rent[us$d_rentwc == 2 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 2 & !is.na(us$d_rentwc)] * 52 / 2 / 12
us$d_rent[us$d_rentwc == 3 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 3 & !is.na(us$d_rentwc)] * 52 / 3 / 12
us$d_rent[us$d_rentwc == 4 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 4 & !is.na(us$d_rentwc)] * 52 / 4 / 12
us$d_rent[us$d_rentwc == 7 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 7 & !is.na(us$d_rentwc)] / 2
us$d_rent[us$d_rentwc == 8 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 8 & !is.na(us$d_rentwc)] * 8 / 12
us$d_rent[us$d_rentwc == 9 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 9 & !is.na(us$d_rentwc)] * 9 / 12
us$d_rent[us$d_rentwc == 10 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 10 & !is.na(us$d_rentwc)] * 10 / 12
us$d_rent[us$d_rentwc == 13 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 13 & !is.na(us$d_rentwc)] / 3
us$d_rent[us$d_rentwc == 26 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 26 & !is.na(us$d_rentwc)] / 6
us$d_rent[us$d_rentwc == 52 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 52 & !is.na(us$d_rentwc)] / 12
us$d_rent[us$d_rentwc == 90 & !is.na(us$d_rentwc)] <-
us$d_rent[us$d_rentwc == 90 & !is.na(us$d_rentwc)] * 52 / 12
us$d_rent[us$d_rentwc == 95 & !is.na(us$d_rentwc)] <- NA # one-off payment
us$d_rent[us$d_rentwc == 96 & !is.na(us$d_rentwc)] <- NA # none of these
us$c_rent[us$c_rentwc == 1 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 1 & !is.na(us$c_rentwc)] * 52 / 12
us$c_rent[us$c_rentwc == 2 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 2 & !is.na(us$c_rentwc)] * 52 / 2 / 12
us$c_rent[us$c_rentwc == 3 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 3 & !is.na(us$c_rentwc)] * 52 / 3 / 12
us$c_rent[us$c_rentwc == 4 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 4 & !is.na(us$c_rentwc)] * 52 / 4 / 12
us$c_rent[us$c_rentwc == 7 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 7 & !is.na(us$c_rentwc)] / 2
us$c_rent[us$c_rentwc == 8 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 8 & !is.na(us$c_rentwc)] * 8 / 12
us$c_rent[us$c_rentwc == 9 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 9 & !is.na(us$c_rentwc)] * 9 / 12
us$c_rent[us$c_rentwc == 10 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 10 & !is.na(us$c_rentwc)] * 10 / 12
us$c_rent[us$c_rentwc == 13 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 13 & !is.na(us$c_rentwc)] / 3
us$c_rent[us$c_rentwc == 26 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 26 & !is.na(us$c_rentwc)] / 6
us$c_rent[us$c_rentwc == 52 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 52 & !is.na(us$c_rentwc)] / 12
us$c_rent[us$c_rentwc == 90 & !is.na(us$c_rentwc)] <-
us$c_rent[us$c_rentwc == 90 & !is.na(us$c_rentwc)] * 52 / 12
us$c_rent[us$c_rentwc == 95 & !is.na(us$c_rentwc)] <- NA # one-off payment
us$c_rent[us$c_rentwc == 96 & !is.na(us$c_rentwc)] <- NA # none of these
us$b_rent[us$b_rentwc == 1 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 1 & !is.na(us$b_rentwc)] * 52 / 12
us$b_rent[us$b_rentwc == 2 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 2 & !is.na(us$b_rentwc)] * 52 / 2 / 12
us$b_rent[us$b_rentwc == 3 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 3 & !is.na(us$b_rentwc)] * 52 / 3 / 12
us$b_rent[us$b_rentwc == 4 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 4 & !is.na(us$b_rentwc)] * 52 / 4 / 12
us$b_rent[us$b_rentwc == 7 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 7 & !is.na(us$b_rentwc)] / 2
us$b_rent[us$b_rentwc == 8 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 8 & !is.na(us$b_rentwc)] * 8 / 12
us$b_rent[us$b_rentwc == 9 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 9 & !is.na(us$b_rentwc)] * 9 / 12
us$b_rent[us$b_rentwc == 10 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 10 & !is.na(us$b_rentwc)] * 10 / 12
us$b_rent[us$b_rentwc == 13 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 13 & !is.na(us$b_rentwc)] / 3
us$b_rent[us$b_rentwc == 26 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 26 & !is.na(us$b_rentwc)] / 6
us$b_rent[us$b_rentwc == 52 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 52 & !is.na(us$b_rentwc)] / 12
us$b_rent[us$b_rentwc == 90 & !is.na(us$b_rentwc)] <-
us$b_rent[us$b_rentwc == 90 & !is.na(us$b_rentwc)] * 52 / 12
us$b_rent[us$b_rentwc == 95 & !is.na(us$b_rentwc)] <- NA # one-off payment
us$b_rent[us$b_rentwc == 96 & !is.na(us$b_rentwc)] <- NA # none of these
us$a_rent[us$a_rentwc == 1 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 1 & !is.na(us$a_rentwc)] * 52 / 12
us$a_rent[us$a_rentwc == 2 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 2 & !is.na(us$a_rentwc)] * 52 / 2 / 12
us$a_rent[us$a_rentwc == 3 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 3 & !is.na(us$a_rentwc)] * 52 / 3 / 12
us$a_rent[us$a_rentwc == 4 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 4 & !is.na(us$a_rentwc)] * 52 / 4 / 12
us$a_rent[us$a_rentwc == 7 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 7 & !is.na(us$a_rentwc)] / 2
us$a_rent[us$a_rentwc == 8 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 8 & !is.na(us$a_rentwc)] * 8 / 12
us$a_rent[us$a_rentwc == 9 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 9 & !is.na(us$a_rentwc)] * 9 / 12
us$a_rent[us$a_rentwc == 10 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 10 & !is.na(us$a_rentwc)] * 10 / 12
us$a_rent[us$a_rentwc == 13 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 13 & !is.na(us$a_rentwc)] / 3
us$a_rent[us$a_rentwc == 26 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 26 & !is.na(us$a_rentwc)] / 6
us$a_rent[us$a_rentwc == 52 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 52 & !is.na(us$a_rentwc)] / 12
us$a_rent[us$a_rentwc == 90 & !is.na(us$a_rentwc)] <-
us$a_rent[us$a_rentwc == 90 & !is.na(us$a_rentwc)] * 52 / 12
us$a_rent[us$a_rentwc == 95 & !is.na(us$a_rentwc)] <- NA # one-off payment
us$a_rent[us$a_rentwc == 96 & !is.na(us$a_rentwc)] <- NA # none of these
us <- us %>%
select(-f_rentwc, -e_rentwc, -d_rentwc, -c_rentwc, -b_rentwc, -a_rentwc)
us$hh_cost <- us$f_xpmg
us$hh_cost[is.na(us$hh_cost)] <- us$f_rent[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$e_xpmg[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$e_rent[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$d_xpmg[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$d_rent[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$c_xpmg[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$c_rent[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$b_xpmg[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$b_rent[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$a_xpmg[is.na(us$hh_cost)]
us$hh_cost[is.na(us$hh_cost)] <- us$a_rent[is.na(us$hh_cost)]
# top code to £20,000
us$hh_cost[us$hh_cost > 20000] <- 20000
us <- us %>%
select(-f_xpmg, -e_xpmg, -d_xpmg, -c_xpmg, -b_xpmg, -a_xpmg) %>%
select(-f_rent, -e_rent, -d_rent, -c_rent, -b_rent, -a_rent)
# Benefits ====
us$uc_replace <- 0
us$uc_replace[us$f_benbase1 == 1] <- 1 # income support
us$uc_replace[us$f_benbase2 == 1] <- 1 # JSA
us$uc_replace[us$f_othben5 == 1] <- 1 # working tax credit
us$uc_replace[us$f_othben8 == 1] <- 1 # housing benefit
us$uc_replace[us$f_benctc == 1] <- 1 # housing benefit (2 == no)
us$uc_replace[us$f_bendis2 == 1] <- 1 # ESA
us$uc_replace[is.na(us$uc_replace) & us$e_benunemp1 == 1] <- 1 # JSA
us$uc_replace[is.na(us$uc_replace) & us$e_btype2 == 1] <- 1 # income sup
us$uc_replace[is.na(us$uc_replace) & us$e_btype3 == 1] <- 1 # ESA
us$uc_replace[is.na(us$uc_replace) & us$e_btype6 == 1] <- 1 # tax cred
us$uc_replace[is.na(us$uc_replace) & us$d_benunemp1 == 1] <- 1 # JSA
us$uc_replace[is.na(us$uc_replace) & us$d_btype2 == 1] <- 1 # income sup
us$uc_replace[is.na(us$uc_replace) & us$d_btype3 == 1] <- 1 # ESA
us$uc_replace[is.na(us$uc_replace) & us$d_btype6 == 1] <- 1 # tax cred
us$uc_replace[is.na(us$uc_replace) & us$c_benunemp1 == 1] <- 1 # JSA
us$uc_replace[is.na(us$uc_replace) & us$c_btype2 == 1] <- 1 # income sup
us$uc_replace[is.na(us$uc_replace) & us$c_btype3 == 1] <- 1 # ESA
us$uc_replace[is.na(us$uc_replace) & us$c_btype6 == 1] <- 1 # tax cred
us$uc_replace[is.na(us$uc_replace) & us$b_benunemp1 == 1] <- 1 # JSA
us$uc_replace[is.na(us$uc_replace) & us$b_btype2 == 1] <- 1 # income sup
us$uc_replace[is.na(us$uc_replace) & us$b_btype3 == 1] <- 1 # ESA
us$uc_replace[is.na(us$uc_replace) & us$b_btype6 == 1] <- 1 # tax cred
us$uc_replace[is.na(us$uc_replace) & us$a_benunemp1 == 1] <- 1 # JSA
us$uc_replace[is.na(us$uc_replace) & us$a_btype2 == 1] <- 1 # income sup
us$uc_replace[is.na(us$uc_replace) & us$a_btype3 == 1] <- 1 # ESA
us$uc_replace[is.na(us$uc_replace) & us$a_btype6 == 1] <- 1 # tax cred
us <- us %>%
select(-f_benbase1, -f_benbase2, -f_othben5, -f_othben8, -f_benctc,
-f_bendis2,
-e_benunemp1, -e_btype2, -e_btype3, -e_btype6,
-d_benunemp1, -d_btype2, -d_btype3, -d_btype6,
-c_benunemp1, -c_btype2, -c_btype3, -c_btype6,
-b_benunemp1, -b_btype2, -b_btype3, -b_btype6,
-a_benunemp1, -a_btype2, -a_btype3, -a_btype6)
# Save data frame ====
save(us, file = "data/us.RData", compress = "xz")
rm(us)
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.