In caseybreen/wcensoc:

Summary: Assessing match quality for men in the CenSoc-Numident (V2) file.

## Library packages 
library(tidyverse)
library(data.table)
library(censocdev)
library(cowplot)
library(gt)
library(ipumsr)
library(here)
library(janitor)

Read in data

## read in dmf 
dmf <- fread("/censoc/data/censoc_v2.1/censoc_dmf_v2.1.csv") %>% 
  janitor::clean_names()

## read in 1940 census 
census_1940 <- fread("/ipums-repo2019/1940/TSV/P.tsv", select = c("HISTID", "SERIALP", "AGE", "INCWAGE", "SEX", "EDUC", "RACE", "RELATE", "REGION", "MARST", "SEI")) %>% 
  janitor::clean_names()

## read in 1940 census household
census_1940_h <- fread("/ipums-repo2019/1940/TSV/H.tsv", select = c("SERIAL", "STATEFIP", "OWNERSHP", "URBAN", "REGION")) %>% 
  janitor::clean_names()

## read in data describing probability of dying in left and right truncated window 
dmf_death_prop <- read_csv(here("vignettes/assess_match_quality/data/dmf_death_prop.csv")) %>% 
  mutate(age = 1940 - cohort) ## imprecise but OK

Link data

## combine census person and household vars 
census_1940_hh_vars <- census_1940 %>% 
  inner_join(census_1940_h, by = c("serialp" = "serial"))

## add on dmf  
census_1940_hh_vars <- census_1940_hh_vars %>% 
  left_join(dmf, by = "histid")

## drop women (dmf only has men) 
census_1940_hh_vars <- census_1940_hh_vars %>% 
  filter(sex == 1)

## recode education variable to years of education
census_1940_hh_vars <- census_1940_hh_vars %>% 
  censocdev::recode_education(educ_var = educ)

## read data dictionary initiative (DDI) file with varnames and var labels 
ipums_ddi <- ipumsr::read_ipums_ddi("/ipums-repo2019-1/fullcount.ddi.xml")

## add on ipums value + variable labels  
census_1940_hh_vars <- census_1940_hh_vars %>% 
  janitor::clean_names(case = "all_caps") %>% 
  ipumsr::ipums_collect(ipums_ddi, var_attrs = c("val_labels", "var_label", "var_desc")) %>% 
  janitor::clean_names()

Recode variables

## recode variables 
census_1940_hh_vars <- census_1940_hh_vars %>% 
  mutate(match_conservative = case_when(
    link_abe_exact_conservative == 1 ~ "Matched",
    TRUE ~ "Unmatched"
  ),
  match_standard = case_when(
    link_abe_exact_conservative %in% c(0, 1) ~ "Matched",
    TRUE ~ "Unmatched"
  ), 
  hs = case_when(
    educ >= 60 & educ < 998 ~ 1,
    TRUE ~ 0
  ), 
  rural = case_when(
    urban == 1 ~ 1,
    TRUE ~ 0
  ),
  black = case_when( 
    race == 200 ~ 1,
    TRUE ~ 0
  ), white = case_when(
    race == 100 ~ 1,
    TRUE ~ 0
  ),
  homeown = case_when(
    ownershp == 10 ~ 1, 
    TRUE ~ 0
  ),
  p_hh_head = case_when(
    relate == 101 ~ 1, 
    TRUE ~ 0
  ),
  p_hh_head = case_when(
    relate == 101 ~ 1, 
    TRUE ~ 0
  ),
  educ_level = case_when(
    educ <= 50 ~ "< High School",
    educ %in% 60:90 ~ "High School or some college",
    educ %in% 100 ~ "Bachelors Degree",
    educ %in% 110:116 ~ "Advanced Degree"
  ), 
  sei_recode = case_when(
    sei %in% 1:9 ~ "sei_1_9",
    sei %in% 10:14 ~ "sei_10_14",
    sei %in% 15:25 ~ "sei_15_25",
    sei >= 26 ~      "sei_26+"
  ),
  marital_status = case_when(
      marst %in% 1:2 ~ "married",
      TRUE ~ "not married" ),
  race_recode = case_when(
    race == 100 ~ "White",
    race == 200 ~ "Black",
    TRUE ~ "Other"
  ))

generate table

## recode data  
census_1940_hh_vars_recode  <- census_1940_hh_vars %>% 
  mutate(rural = case_when(
    rural == 1 ~ "Rural",
    TRUE ~ "Urban"
  ),
  homeown = case_when(
    homeown == 1 ~ "Home Owner",
    TRUE ~ "Not Home Owner"
  ),
  region_string = as_factor(region)) %>% 
  filter(age %in% 20:40) %>% 
   mutate(link_abe_exact_standard = case_when(
    link_abe_exact_conservative %in% c(0, 1) ~ 1,
    TRUE ~ 0
  ))

## tabulations for men in 1940 census 
census_characteristics <- census_1940_hh_vars_recode %>% 
  select(histid, race_recode, educ_level, sei_recode, marital_status, region_string, rural, homeown) %>%
  pivot_longer(-histid) %>% 
  group_by(name, value) %>%
  tally() %>%            
  mutate(prop = round(100*prop.table(n), 1)) %>% 
  rename(n_gen = n, prop_gen = prop)

## tabulations for men in CenSoc-DMF standard 
dmf_characteristics_standard <- census_1940_hh_vars_recode %>% 
  filter(link_abe_exact_standard == 1) %>% 
  select(histid, race_recode, educ_level, sei_recode, marital_status, region_string, rural, homeown) %>%
  pivot_longer(-histid) %>% 
  group_by(name, value) %>%
  tally() %>%            
  mutate(prop = round(100*prop.table(n), 1)) %>% 
  rename(n_gen_standard = n, prop_standard = prop)

## tabulations for men in CenSoc-DMF conservative 
dmf_characteristics_conservative <- census_1940_hh_vars_recode %>% 
  filter(link_abe_exact_conservative == 1) %>% 
  select(histid, race_recode, educ_level, sei_recode, marital_status, region_string, rural, homeown) %>%
  pivot_longer(-histid) %>% 
  group_by(name, value) %>%
  tally() %>%            
  mutate(prop = round(100*prop.table(n), 1)) %>% 
  rename(n_gen_conservative = n, prop_conservative = prop)

combined_characteristics <- census_characteristics %>% 
  inner_join(dmf_characteristics_standard, by = c("name", "value")) %>% 
  inner_join(dmf_characteristics_conservative, by = c("name", "value")) %>% 
  mutate(name = as.factor(name), value = as.factor(value)) %>% 
  mutate(name = factor(name, levels = c("educ_level", "race_recode", "marital_status", "homeown", "sei_recode", "rural", "region_string"))) %>% 
  arrange(name, value) 

## create table 
## will have to do some manual clean-up but this gets us 80% of the way
table_s3 <- gt(data = combined_characteristics) %>% 
  tab_spanner(
    label = "General Pop",
    columns = vars(
      n_gen, prop_gen)) %>% 
  tab_spanner(
    label = "Standard",
    columns = vars(
      n_gen_standard, prop_standard)) %>% 
  tab_spanner(
    label = "Conservative",
    columns = vars(
      n_gen_conservative, prop_conservative)) %>% 
  cols_label(
    "n_gen" = "No.",
    "prop_gen" = "%",
    "n_gen_standard" = "No.",
    "prop_standard" = "%",
    "n_gen_conservative"  = "No.",
    "prop_conservative" = "%",
    value = ""
  ) %>% 
  # row_group_order(
  #     groups = c("Gender", "Age", "Education", "Race")
  #   ) %>% 
  tab_style(
    style = list(
      cell_text(weight = "bold")),
    locations = cells_row_groups()
  ) %>% 
  opt_row_striping(row_striping = T) %>% 
  cols_align("left")


## save table 
table_s3 %>%
  gtsave("dmf_characteristics_table.tex", path = here("vignettes/assess_match_quality/figs/"))

data frame with just black americans

census_1940_hh_vars_recode_ba  <- census_1940_hh_vars %>% 
  mutate(rural = case_when(
    rural == 1 ~ "Rural",
    TRUE ~ "Urban"
  ),
  homeown = case_when(
    homeown == 1 ~ "Home Owner",
    TRUE ~ "Not Home Owner"
  ),
  region_string = as_factor(region)) %>% 
  filter(age %in% 20:40, race_recode == "Black") %>% 
  mutate(link_abe_exact_standard = case_when(
    link_abe_exact_conservative %in% c(0, 1) ~ 1,
    TRUE ~ 0
  ))

table for black americans: dmf

## tabulations for men in 1940 census 
census_characteristics_ba <- census_1940_hh_vars_recode_ba %>% 
  select(histid, educ_level, sei_recode, marital_status, region_string, rural, homeown) %>%
  pivot_longer(-histid) %>% 
  group_by(name, value) %>%
  tally() %>%            
  mutate(prop = round(100*prop.table(n), 1)) %>% 
  rename(n_gen = n, prop_gen = prop)

## tabulations for men in CenSoc-DMF standard 
dmf_characteristics_standard_ba <- census_1940_hh_vars_recode_ba %>% 
  filter(link_abe_exact_standard == 1) %>% 
  select(histid, educ_level, sei_recode, marital_status, region_string, rural, homeown) %>%
  pivot_longer(-histid) %>% 
  group_by(name, value) %>%
  tally() %>%            
  mutate(prop = round(100*prop.table(n), 1)) %>% 
  rename(n_gen_standard = n, prop_standard = prop)

## tabulations for men in CenSoc-DMF conservative 
dmf_characteristics_conservative_ba <- census_1940_hh_vars_recode_ba %>% 
  filter(link_abe_exact_conservative == 1) %>% 
  select(histid, educ_level, sei_recode, marital_status, region_string, rural, homeown) %>%
  pivot_longer(-histid) %>% 
  group_by(name, value) %>%
  tally() %>%            
  mutate(prop = round(100*prop.table(n), 1)) %>% 
  rename(n_gen_conservative = n, prop_conservative = prop)

combined_characteristics_ba <- census_characteristics_ba %>% 
  inner_join(dmf_characteristics_standard_ba, by = c("name", "value")) %>% 
  inner_join(dmf_characteristics_conservative_ba, by = c("name", "value")) %>% 
  mutate(name = as.factor(name), value = as.factor(value)) %>% 
  mutate(name = factor(name, levels = c("educ_level", "marital_status", "homeown", "sei_recode", "rural", "region_string"))) %>% 
  arrange(name, value) 

## create table 
## will have to do some manual clean-up but this gets us 80% of the way
table_s3_ba <- gt(data = combined_characteristics_ba) %>% 
  tab_spanner(
    label = "General Pop",
    columns = vars(
      n_gen, prop_gen)) %>% 
  tab_spanner(
    label = "Standard",
    columns = vars(
      n_gen_standard, prop_standard)) %>% 
  tab_spanner(
    label = "Conservative",
    columns = vars(
      n_gen_conservative, prop_conservative)) %>% 
  cols_label(
    "n_gen" = "No.",
    "prop_gen" = "%",
    "n_gen_standard" = "No.",
    "prop_standard" = "%",
    "n_gen_conservative"  = "No.",
    "prop_conservative" = "%",
    value = ""
  ) %>% 
  # row_group_order(
  #     groups = c("Gender", "Age", "Education", "Race")
  #   ) %>% 
  tab_style(
    style = list(
      cell_text(weight = "bold")),
    locations = cells_row_groups()
  ) %>% 
  opt_row_striping(row_striping = T) %>% 
  cols_align("left")

## save table 
table_s3_ba %>%
  gtsave("dmf_characteristics_blackamericans_table.tex", path = here("vignettes/assess_match_quality/figs/"))