BDDHS data pooling pre-checks

Getting started

Here we show the pre-requisite code sections. Run these at the outset to avoid errors. First we load the required packages.

easypackages::libraries(
  # Data i/o
  "here",                 # relative file path
  "rio",                  # file import-export
  
  # Data manipulation
  "janitor",              # data cleaning fns
  "haven",                # stata, sas, spss data io
  "labelled",             # var labelling
  "readxl",               # excel sheets
  # "scales",               # to change formats and units
  "skimr",                # quick data summary
  "broom",                # view model results
  
  # Data analysis
  "DHS.rates",            # demographic rates for dhs-like surveys
  "GeneralOaxaca",        # BO decomposition for non-linear
  "survey",               # apply survey weights
  
  # Analysis output
  "gt",
  # "modelsummary",          # output summary tables
  "gtsummary",            # output summary tables
  "flextable",            # creating tables from objects
  "officer",              # editing in office docs
  
  # R graph related packages
  "ggstats",
  "RColorBrewer",
  # "scales",
  "patchwork",
  
  # Misc packages
  "tidyverse",            # Data manipulation iron man
  "tictoc"                # Code timing
)

Next we turn off scientific notations.

options(scipen = 999)

Next we set the default gtsummary print engine for tables.

theme_gtsummary_printer(print_engine = "flextable")

Now we set the flextable output defaults.

set_flextable_defaults(
  font.size = 11,
  text.align = "left",
  big.mark = "",
  background.color = "white",
  table.layout = "autofit",
  theme_fun = theme_vanilla
)

Document introduction

Here we document the variable codes and labels of variables across all the Bangladesh Demographic and Health Survey (DHS) datasets. We check the variable labels and codes before running the pooling code in “daprep-v01_bddhs.R”. We pool the following Bangladesh DHS surveys:

# Creating the table of surveys to be used for pooling
bdbr1_tmp_intro |> 
  mutate(n_births = prettyNum(n_births, big.mark = ",")) |> 
  select(c(ctr_name, svy_year, n_births)) |> 
  # Join vars from bdir_tmp_intro
  left_join(
    bdir1_tmp_intro |> 
      mutate(n_women = prettyNum(n_women, big.mark = ",")) |> 
      select(c(year, n_women)),
    by = join_by(svy_year == year)
  ) |> 
  # Join vars from bdhr_tmp_intro
  left_join(
    bdhr1_tmp_intro |> 
      mutate(n_households = prettyNum(n_households, big.mark = ",")) |> 
      select(svy_year, n_households),
    by = join_by(svy_year)
  ) |> 
  # Join vars from bdpr_tmp_intro
  left_join(
    bdpr1_tmp_intro |> 
      mutate(n_persons = prettyNum(n_persons, big.mark = ",")) |> 
      select(svy_year, n_persons),
    by = join_by(svy_year)
  ) |> 
  # convert nested tibble to simple tibble
  unnest(cols = c()) |> 
  mutate(
    ccode = row_number(), 
    .before = ctr_name
  ) |> 
  # convert to flextable object
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 1: Bangladesh DHS datasets and their sample size to be used for pooling

ccode	ctr_name	svy_year	n_births	n_women	n_households	n_persons
1	Bangladesh	1993	32,581	9,493	9,174	51,631
2	Bangladesh	1996	29,344	8,981	8,682	47,432
3	Bangladesh	1999	31,906	10,373	9,854	54,627
4	Bangladesh	2004	33,597	11,300	10,500	55,883
5	Bangladesh	2007	30,527	10,996	10,400	53,413
6	Bangladesh	2011	45,833	17,749	17,141	83,731
7	Bangladesh	2014	43,772	17,863	17,300	81,624
8	Bangladesh	2017	47,828	20,127	19,457	89,819
9	Bangladesh	2022	64,722	30,078	30,018	132,463

We use the following variables for the pooled data analysis:

• Dependent variable
  • infantd = Index child died during infancy period (0-11 months)
• Main Independent variable
  • sibsurv_nmv = Survival status of preceding child (Death scarring)
  • binterval_3c_nmv_opp = Birth interval preceding to index child
• Independent variables [CHILD LEVEL]
  • cyob10y_opp = Birth cohort of index child
  • bord_c = Birth order of index child
  • sex_fm = Gender of index child
  • season = Season during birth
• Independent variables [MOTHER/PARENT LEVEL]
  • myob_opp = Birth cohort of mother
  • macb_c_opp = Mother’s age during birth of index child
  • medu_opp = Mother’s Level of education
  • fedu_opp = Father’s level of education
• Independent variables [HOUSEHOLD LEVEL]
  • religion = Religion
  • nat_lang = Native language of respondent
  • wi_qt_opp = Household wealth quintile
  • hhgen_2c_opp = Generations in household
  • hhstruc_opp = Household structure
  • head_sex_fm = Sex of HH head
• Independent variables [COMMUNITY LEVEL]
  • por = Place of residence of the household
  • ecoreg = Ecological region

Note: (a) Crossed names indicates variable not included.

Data import

We will directly import the nested tibble here. The code for dataset preparation is in the “daprep-v01_bddhs.R” script file.

# Here we import the tibbles to be used for dataset checking
# Import the bdbr nested tibble
bdbr1_pre_tmp0 <- read_rds(file = here("website_data", "bdbr1_nest0.rds"))
# Import the bdhr nested tibble
bdhr1_pre_tmp0 <- read_rds(file = here("website_data", "bdhr1_nest0.rds"))
# Import the bdpr nested tibble
bdpr1_pre_tmp0 <- read_rds(file = here("website_data", "bdpr1_nest0.rds"))

Bangladesh BR dataset use for variable creation

Checking the Women’s weight variable before harmonization

We will check the formatting of the v005 women’s weight variable before creating the pooled survey weight. For this we will use the labelled::look_for().

# First we create the data dictionary of v005 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_v005 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(v005) |> 
        look_for(details = "full") |> 
        # For correctly viewing the range column in data dictionary
        convert_list_columns_to_character() |> 
        select(-c(levels:n_na))
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary 
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_v005)) |> 
  select(-pos) 
# Convert and view the tibble as flextable
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 2: Data dictionary of v005 variable across the bdbr rounds

ctr_name	svy_year	variable	label	col_type	missing	unique_values	range
Bangladesh	1993	v005	sample weight	dbl	0	14	292519 - 1327458
Bangladesh	1996	v005	sample weight	dbl	0	16	191059 - 1453427
Bangladesh	1999	v005	sample weight	dbl	0	12	250565 - 1494863
Bangladesh	2004	v005	sample weight	dbl	0	350	55728 - 2707592
Bangladesh	2007	v005	sample weight	dbl	0	359	135650 - 3592687
Bangladesh	2011	v005	women's individual sample weight (6 decimals)	dbl	0	465	184316 - 3381150
Bangladesh	2014	v005	women's individual sample weight (6 decimals)	dbl	0	454	153594 - 14194492
Bangladesh	2017	v005	women's individual sample weight (6 decimals)	dbl	0	661	158603 - 2465775
Bangladesh	2022	v005	women's individual sample weight (6 decimals)	dbl	0	660	91325 - 3891736

The women’s weight variables are in numeric class and have no missing values. Therefore, we need not reformat them. Hence we directly use it for preparing the pooled survey weight. NOTE that, the women’s weight for the Bangladesh 1993, 1996 and 1999 rounds have few unique values. This could be because there might have been fewer sampling units in the secondary stage.

Checking the ID variables before harmonization

Here we check the formatting of the variables using which we will prepare the ID variables for the pooled Bangladesh birth history recode (br) dataset. We will use the following constituent variables for creating the ID variables for the pooled dataset:

# We check the var type of ID vars in all bdbr datasets.
# First we create a data dictionary of the bdbr datasets in nested tibble.
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |>
  mutate(lookfor_idvars = map(
    bdbr_data,
    \(df) {
      df |> 
        select(v001, v002, v003, bord, v021, v022, v023, v024) |> 
        lookfor(details = "full") |> 
        select(-c(levels:n_na)) |> 
        # For correctly viewing the range column in data dictionary
        convert_list_columns_to_character()
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and output the pooled data dictionary 
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_idvars)) |> 
  arrange(pos)

# Convert and view the tibble as flextable
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 3: Data dictionary of variables to be used for ID creation across the bdbr rounds

ctr_name	svy_year	pos	variable	label	col_type	missing	unique_values	range
Bangladesh	1993	1	v001	cluster number	dbl	0	301	101 - 573
Bangladesh	1996	1	v001	cluster number	dbl	0	313	101 - 630
Bangladesh	1999	1	v001	cluster number	dbl	0	341	3 - 500
Bangladesh	2004	1	v001	cluster number	dbl	0	361	1 - 550
Bangladesh	2007	1	v001	cluster number	dbl	0	361	1 - 361
Bangladesh	2011	1	v001	cluster number	dbl	0	600	1 - 600
Bangladesh	2014	1	v001	cluster number	dbl	0	600	1 - 600
Bangladesh	2017	1	v001	cluster number	dbl	0	672	1 - 675
Bangladesh	2022	1	v001	cluster number	dbl	0	674	1 - 675
Bangladesh	1993	2	v002	household number	dbl	0	529	1 - 615
Bangladesh	1996	2	v002	household number	dbl	0	537	1 - 638
Bangladesh	1999	2	v002	household number	dbl	0	489	1 - 544
Bangladesh	2004	2	v002	household number	dbl	0	235	1 - 280
Bangladesh	2007	2	v002	household number	dbl	0	194	1 - 252
Bangladesh	2011	2	v002	household number	dbl	0	178	1 - 217
Bangladesh	2014	2	v002	household number	dbl	0	203	1 - 222
Bangladesh	2017	2	v002	household number	dbl	0	250	1 - 299
Bangladesh	2022	2	v002	household number	dbl	0	203	1 - 220
Bangladesh	1993	3	v003	respondent's line number	dbl	0	22	1 - 26
Bangladesh	1996	3	v003	respondent's line number	dbl	0	23	1 - 28
Bangladesh	1999	3	v003	respondent's line number	dbl	0	19	1 - 19
Bangladesh	2004	3	v003	respondent's line number	dbl	0	24	1 - 30
Bangladesh	2007	3	v003	respondent's line number	dbl	0	22	1 - 26
Bangladesh	2011	3	v003	respondent's line number	dbl	0	21	1 - 23
Bangladesh	2014	3	v003	respondent's line number	dbl	0	19	1 - 22
Bangladesh	2017	3	v003	respondent's line number	dbl	0	21	1 - 27
Bangladesh	2022	3	v003	respondent's line number	dbl	0	19	1 - 23
Bangladesh	1993	4	bord	birth order number	dbl	0	15	1 - 15
Bangladesh	1996	4	bord	birth order number	dbl	0	15	1 - 15
Bangladesh	1999	4	bord	birth order number	dbl	0	16	1 - 16
Bangladesh	2004	4	bord	birth order number	dbl	0	15	1 - 15
Bangladesh	2007	4	bord	birth order number	dbl	0	14	1 - 14
Bangladesh	2011	4	bord	birth order number	dbl	0	20	1 - 20
Bangladesh	2014	4	bord	birth order number	dbl	0	15	1 - 15
Bangladesh	2017	4	bord	birth order number	dbl	0	13	1 - 13
Bangladesh	2022	4	bord	birth order number	dbl	0	11	1 - 11
Bangladesh	1993	5	v021	primary sampling unit	dbl	0	301	101 - 573
Bangladesh	1996	5	v021	primary sampling unit	dbl	0	313	101 - 630
Bangladesh	1999	5	v021	primary sampling unit	dbl	0	341	3 - 500
Bangladesh	2004	5	v021	primary sampling unit	dbl	0	361	1 - 550
Bangladesh	2007	5	v021	primary sampling unit	dbl	0	361	1 - 361
Bangladesh	2011	5	v021	primary sampling unit	dbl	0	600	1 - 600
Bangladesh	2014	5	v021	primary sampling unit	dbl	0	600	1 - 600
Bangladesh	2017	5	v021	primary sampling unit	dbl	0	672	1 - 675
Bangladesh	2022	5	v021	primary sampling unit	dbl	0	674	1 - 675
Bangladesh	1993	6	v022	sample stratum number	dbl	0	148	1 - 148
Bangladesh	1996	6	v022	sample stratum number	dbl	0	152	1 - 154
Bangladesh	1999	6	v022	sample stratum number	dbl	0	168	1 - 168
Bangladesh	2004	6	v022	sample stratum number	dbl	0	177	1 - 177
Bangladesh	2007	6	v022	sample stratum number	dbl	0	179	1 - 179
Bangladesh	2011	6	v022	sample strata for sampling errors	dbl	0	20	1 - 20
Bangladesh	2014	6	v022	sample strata for sampling errors	dbl+lbl	0	20	1 - 21
Bangladesh	2017	6	v022	sample strata for sampling errors	dbl+lbl	0	22	1 - 22
Bangladesh	2022	6	v022	sample strata for sampling errors	dbl+lbl	0	16	1 - 16
Bangladesh	1993	7	v023	sample domain	dbl+lbl	0	14	2 - 15
Bangladesh	1996	7	v023	sample domain	dbl+lbl	0	16	1 - 16
Bangladesh	1999	7	v023	sample domain	dbl+lbl	0	12	1 - 12
Bangladesh	2004	7	v023	sample domain	dbl+lbl	0	1	0 - 0
Bangladesh	2007	7	v023	sample domain	dbl	0	22	1 - 22
Bangladesh	2011	7	v023	stratification used in sample design	dbl+lbl	0	20	1 - 20
Bangladesh	2014	7	v023	stratification used in sample design	dbl+lbl	0	21	1 - 21
Bangladesh	2017	7	v023	stratification used in sample design	dbl+lbl	0	22	1 - 22
Bangladesh	2022	7	v023	stratification used in sample design	dbl+lbl	0	16	1 - 16
Bangladesh	1993	8	v024	region	dbl+lbl	0	5	1 - 5
Bangladesh	1996	8	v024	division	dbl+lbl	0	6	1 - 6
Bangladesh	1999	8	v024	division	dbl+lbl	0	6	1 - 6
Bangladesh	2004	8	v024	region	dbl+lbl	0	6	1 - 6
Bangladesh	2007	8	v024	division	dbl+lbl	0	6	1 - 6
Bangladesh	2011	8	v024	region	dbl+lbl	0	7	1 - 7
Bangladesh	2014	8	v024	division	dbl+lbl	0	7	1 - 7
Bangladesh	2017	8	v024	division	dbl+lbl	0	8	1 - 8
Bangladesh	2022	8	v024	division	dbl+lbl	0	8	1 - 8

From the above we can see that v023 and v024 are of labelled class, while the rest are in numeric class. Therefore, we will have to check the numeric and labelled variables in different ways. Note that although survey year is a constituent ID variable we have not checked it. It is imperative that survey year would be a 4-digit variable.

Numeric ID variables check

First, let’s find out the required length of the numeric ID variables by checking the maximum values of the constituent ID variable across the Bangladesh DHS datasets. Here we estimate the summary stats of numeric constituent variables using skim_without_charts().

# Check the summary stats for ID vars using skimr in each bdbr dataset.
# First we estimate the summary stats using skim_without_charts().
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(
    skim_id_num = map(
      bdbr_data,
      function(df) {
        df |> 
          select(v001, v002, v003, bord, v021, v022) |> 
          skim_without_charts() |> 
          as_tibble() |> 
          select(-c(skim_type, n_missing, complete_rate)) |> 
          rename(
            variable = 1,
            mean = 2,
            sd = 3,
            min = 4,
            p25 = 5,
            p50 = 6,
            p75 = 7,
            max = 8
          )
      }
    )
  )
bdbr1_pre_tmp1

Next, we check the summary stats of numeric variables by variable name-wise.

# Now we unnest the nested tibble so that we can compare the variable length 
# across the bdbr datasets.
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(skim_id_num)) |> 
  arrange(variable, svy_year) |> 
  # change the decimal places of selected variables
  mutate(
    mean = sprintf("%.1f", mean),
    sd = sprintf("%.1f", sd),
    p75 = sprintf("%.0f", p75)
  )
# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 4: Summary statistics of the numeric ID variables

ctr_name	svy_year	variable	mean	sd	min	p25	p50	p75	max
Bangladesh	1993	bord	3.2	2.2	1	1	3	4	15
Bangladesh	1996	bord	3.1	2.1	1	1	3	4	15
Bangladesh	1999	bord	2.9	2.0	1	1	2	4	16
Bangladesh	2004	bord	2.8	1.9	1	1	2	4	15
Bangladesh	2007	bord	2.7	1.8	1	1	2	4	14
Bangladesh	2011	bord	2.5	1.6	1	1	2	3	20
Bangladesh	2014	bord	2.4	1.5	1	1	2	3	15
Bangladesh	2017	bord	2.2	1.4	1	1	2	3	13
Bangladesh	2022	bord	2.0	1.2	1	1	2	3	11
Bangladesh	1993	v001	356.6	137.1	101	247	361	501	573
Bangladesh	1996	v001	392.2	151.6	101	266	384	533	630
Bangladesh	1999	v001	264.0	141.3	3	149	270	392	500
Bangladesh	2004	v001	212.1	160.2	1	81	175	325	550
Bangladesh	2007	v001	181.8	107.6	1	87	177	279	361
Bangladesh	2011	v001	302.7	178.0	1	143	299	461	600
Bangladesh	2014	v001	301.1	180.8	1	136	299	466	600
Bangladesh	2017	v001	338.1	202.7	1	152	338	522	675
Bangladesh	2022	v001	338.9	200.1	1	158	338	517	675
Bangladesh	1993	v002	145.5	115.7	1	53	117	214	615
Bangladesh	1996	v002	149.4	118.3	1	54	119	221	638
Bangladesh	1999	v002	154.3	110.3	1	61	135	229	544
Bangladesh	2004	v002	54.4	39.6	1	23	48	78	280
Bangladesh	2007	v002	53.7	36.4	1	24	49	78	252
Bangladesh	2011	v002	58.8	35.9	1	28	57	87	217
Bangladesh	2014	v002	58.3	38.0	1	27	54	84	222
Bangladesh	2017	v002	64.4	42.1	1	30	60	93	299
Bangladesh	2022	v002	59.7	38.2	1	28	56	86	220
Bangladesh	1993	v003	2.4	1.6	1	2	2	2	26
Bangladesh	1996	v003	2.4	1.6	1	2	2	2	28
Bangladesh	1999	v003	2.5	1.7	1	2	2	2	19
Bangladesh	2004	v003	2.4	1.7	1	2	2	2	30
Bangladesh	2007	v003	2.4	1.6	1	2	2	2	26
Bangladesh	2011	v003	2.4	1.5	1	2	2	2	23
Bangladesh	2014	v003	2.3	1.4	1	2	2	2	22
Bangladesh	2017	v003	2.3	1.5	1	2	2	2	27
Bangladesh	2022	v003	2.3	1.3	1	2	2	2	23
Bangladesh	1993	v021	356.6	137.1	101	247	361	501	573
Bangladesh	1996	v021	392.2	151.6	101	266	384	533	630
Bangladesh	1999	v021	264.0	141.3	3	149	270	392	500
Bangladesh	2004	v021	212.1	160.2	1	81	175	325	550
Bangladesh	2007	v021	181.8	107.6	1	87	177	279	361
Bangladesh	2011	v021	302.7	178.0	1	143	299	461	600
Bangladesh	2014	v021	301.1	180.8	1	136	299	466	600
Bangladesh	2017	v021	338.1	202.7	1	152	338	522	675
Bangladesh	2022	v021	338.9	200.1	1	158	338	517	675
Bangladesh	1993	v022	76.1	43.8	1	37	78	113	148
Bangladesh	1996	v022	80.1	45.3	1	38	81	120	154
Bangladesh	1999	v022	87.6	49.1	1	44	90	131	168
Bangladesh	2004	v022	86.3	51.7	1	40	87	131	177
Bangladesh	2007	v022	90.6	53.6	1	43	88	139	179
Bangladesh	2011	v022	14.1	4.9	1	11	15	18	20
Bangladesh	2014	v022	15.1	5.2	1	12	16	19	21
Bangladesh	2017	v022	12.2	6.2	1	6	12	18	22
Bangladesh	2022	v022	8.5	4.6	1	4	8	12	16

Now we find out the required length of the numeric ID variables to be set, so that we can correctly concatenate them to create the ID variables. The required length of the numeric ID variables are given in max_digits column. Note that survey year is also a constituent ID variable of 4-digits.

# Processing the above nested tibble further
bdbr1_pre_tmp3 <- bdbr1_pre_tmp2 |> 
  group_by(variable) |> 
  # find the minimum and maximum values across surveys 
  summarize(
    min_val = min(min),
    max_val = max(max)
  ) |> 
  mutate(
    # calculate the num of digits in the maximum values
    max_digits = nchar(as.character(max_val)),
    # convert char var to factor
    variable = fct(
      variable, 
      levels = c("v001", "v002", "v003", "bord", "v021", "v022")
    )
  ) |> 
  # sort the rows by factor levels 
  arrange(variable) |> 
  # add variable labels and relocate it after variable name.
  bind_cols(vlabel = c("cluster number", "household number", 
                       "respondent's line number", "birth order", 
                       "primary sampling unit", "sample strata for se")) |> 
  relocate(vlabel, .after = 1)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp3 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 5: The maximum length of numeric variables to be set across the bdbr rounds for concatenating the ID variables

variable	vlabel	min_val	max_val	max_digits
v001	cluster number	1	675	3
v002	household number	1	638	3
v003	respondent's line number	1	30	2
bord	birth order	1	20	2
v021	primary sampling unit	1	675	3
v022	sample strata for se	1	179	3

Labelled ID variables check

First we check the labels in sub-national region variable coded as v024 across the bdbr datasets. Let’s create a nested tibble of v024’s value labels.

# Create the data dictionary for v024 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_v024 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(v024) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

Now we view the value labels of v024 in the table below.

# Now we unnest the tibble and refine the pooled data dictionary 
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_v024)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |> 
  # Show the variable name in a col
  mutate(var_name = "v024", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 6: Data dictionary of v024 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	v024	1	[1] barishal	[1] barisal	[1] barisal	[1] barisal	[1] barisal	[1] barisal	[1] barisal	[1] barisal	[1] barishal
Bangladesh	v024	2	[2] chittagong	[2] chittagong	[2] chittagong	[2] chittagong	[2] chittagong	[2] chittagong	[2] chittagong	[2] chittagong	[2] chattogram
Bangladesh	v024	3	[3] dhaka	[3] dhaka	[3] dhaka	[3] dhaka	[3] dhaka	[3] dhaka	[3] dhaka	[3] dhaka	[3] dhaka
Bangladesh	v024	4	[4] khulna	[4] khulna	[4] khulna	[4] khulna	[4] khulna	[4] khulna	[4] khulna	[4] khulna	[4] khulna
Bangladesh	v024	5	[5] rajshani	[5] rajashahi	[5] rajashahi	[5] rajshahi	[5] rajshahi	[5] rajshahi	[5] rajshahi	[5] mymensingh	[5] mymensingh
Bangladesh	v024	6		[6] sylhet	[6] sylhet	[6] sylhet	[6] sylhet	[6] rangpur	[6] rangpur	[6] rajshahi	[6] rajshahi
Bangladesh	v024	7						[7] sylhet	[7] sylhet	[7] rangpur	[7] rangpur
Bangladesh	v024	8								[8] sylhet	[8] sylhet

NOTE: The sub-national region variable, v024 denote the same variable concept across all bdbr rounds. The number of value labels increases across the survey years denoting creation of new sub-national regions. In 1999 there were 5 regions, which increased to 6 for bdbr 1996, 1999, 2004 and 2007, then it increased 7 for bdbr 2011 and 2014, finally there were 8 regions in 2017 and 2022.
VERD: In this analysis, we do not use the region var in the ID var.

---

Secondly, we check the labels in v023 variable that denotes the stratifications used for sampling design. First we create a nested tibble of v023’s value labels.

# Create the data dictionary for v023 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_v023 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(v023) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

Now we view the value labels of v023 in the table below.

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_v023)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "v023", .before = 2) 

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 7: Data dictionary of v023 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	v023	0	[0] national			[0] national
Bangladesh	v023	1	[1] country specific	[1] barisal - municip.	[1] barisal : urban	[1] country specific		[1] barisal city corp.	[1] barisal city corp.	[1] barisal - city corporation	[1] barishal urban
Bangladesh	v023	2		[2] barisal - rural	[2] barisal : rural			[2] chittagong city corp.	[2] chittagong city corp.	[2] barisal - other urban	[2] barishal rural
Bangladesh	v023	3		[3] chittagong - sma	[3] chittagong : urban			[3] dhaka city corp.	[3] dhaka city corp.	[3] barisal - rural	[3] chattogram urban
Bangladesh	v023	4		[4] chittagong - municip	[4] chittagong : rural			[4] khulna city corp.	[4] khulna city corp.	[4] chittagong - city corporation	[4] chattogram rural
Bangladesh	v023	5		[5] chittagong - rural	[5] dhaka : urban			[5] rajshahi city corp.	[5] rajshahi city corp.	[5] chittagong - other urban	[5] dhaka urban
Bangladesh	v023	6		[6] dhaka - sma	[6] dhaka : rural			[6] sylhet city corp.	[6] rangpur city corp.	[6] chittagong - rural	[6] dhaka rural
Bangladesh	v023	7		[7] dhaka - municip.	[7] khulna : urban			[7] barisal other urban	[7] sylhet city corp.	[7] dhaka - city corporation	[7] khulna urban
Bangladesh	v023	8		[8] dhaka - rural	[8] khulna : rural			[8] chittagong other urban	[8] barisal other urban	[8] dhaka - other urban	[8] khulna rural
Bangladesh	v023	9		[9] khulna - sma	[9] rajashahi : urban			[9] dhaka other urban	[9] chittagong other urban	[9] dhaka - rural	[9] mymensingh urban
Bangladesh	v023	10		[10] khulna - municip.	[10] rajashahi : rural			[10] khulna other urban	[10] dhaka other urban	[10] khulna - city corporation	[10] mymensingh rural
Bangladesh	v023	11		[11] khulna - rural	[11] sylhet : urban			[11] rajshahi other urban	[11] khulna other urban	[11] khulna - other urban	[11] rajshahi urban
Bangladesh	v023	12		[12] rajshahi - sma	[12] sylhet : rural			[12] rangpur other urban	[12] rajshahi other urban	[12] khulna - rural	[12] rajshahi rural
Bangladesh	v023	13		[13] rajshahi - municip.				[13] sylhet other urban	[13] rangpur other urban	[13] mymensingh - other urban	[13] rangpur urban
Bangladesh	v023	14		[14] rajshahi - rural				[14] barisal rural	[14] sylhet other urban	[14] mymensingh - rural	[14] rangpur rural
Bangladesh	v023	15		[15] sylhet - municip.				[15] chittagong rural	[15] barisal rural	[15] rajshahi - city corporation	[15] sylhet urban
Bangladesh	v023	16		[16] sylhet - rural				[16] dhaka rural	[16] chittagong rural	[16] rajshahi - other urban	[16] sylhet rural
Bangladesh	v023	17						[17] khulna rural	[17] dhaka rural	[17] rajshahi - rural
Bangladesh	v023	18						[18] rajshahi rural	[18] khulna rural	[18] rangpur - other urban
Bangladesh	v023	19						[19] rangpur rural	[19] rajshahi rural	[19] rangpur - rural
Bangladesh	v023	20						[20] sylhet rural	[20] rangpur rural	[20] sylhet - city corporation
Bangladesh	v023	21							[21] sylhet rural	[21] sylhet - other urban
Bangladesh	v023	22								[22] sylhet - rural
Bangladesh	v023

NOTE: The labels of v023 are different across the survey rounds.
VERD: Therefore we cannot use v023 in the ID variable preparation.

Checking the Birth History variables before harmonization

Undoubtedly the birth history variables are important for this study objective. Therefore, we need to scrutinize all the birth history variables before using them to prepare harmonized variables for the pooled dataset.

# We check the birth history vars in all bdbr datasets.
# First we create a data dictionary in nested tibble.
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |>
  mutate(lookfor_bhvars = map(
    bdbr_data,
    \(df) {
      df |> 
        select(bidx, matches("^b[0-9]+")) |> 
        lookfor(details = "full") |> 
        select(-c(levels:n_na)) |> 
        # For correctly viewing the range column in data dictionary
        convert_list_columns_to_character()
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_bhvars)) |> 
  arrange(pos)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 8: Data dictionary of birth history variables across the bdbr rounds

ctr_name	svy_year	pos	variable	label	col_type	missing	unique_values	range
Bangladesh	1993	1	bidx	birth column number	dbl	0	15	1 - 15
Bangladesh	1996	1	bidx	birth column number	dbl	0	15	1 - 15
Bangladesh	1999	1	bidx	birth column number	dbl	0	16	1 - 16
Bangladesh	2004	1	bidx	birth column number	dbl	0	15	1 - 15
Bangladesh	2007	1	bidx	birth column number	dbl	0	14	1 - 14
Bangladesh	2011	1	bidx	birth column number	dbl	0	20	1 - 20
Bangladesh	2014	1	bidx	birth column number	dbl	0	15	1 - 15
Bangladesh	2017	1	bidx	birth column number	dbl	0	13	1 - 13
Bangladesh	2022	1	bidx	birth column number	dbl	0	11	1 - 11
Bangladesh	1993	2	b0	child is twin	dbl+lbl	0	4	0 - 3
Bangladesh	1996	2	b0	child is twin	dbl+lbl	0	3	0 - 2
Bangladesh	1999	2	b0	child is twin	dbl+lbl	0	4	0 - 3
Bangladesh	2004	2	b0	child is twin	dbl+lbl	0	4	0 - 3
Bangladesh	2007	2	b0	child is twin	dbl+lbl	0	4	0 - 3
Bangladesh	2011	2	b0	child is twin	dbl+lbl	0	4	0 - 3
Bangladesh	2014	2	b0	child is twin	dbl+lbl	0	4	0 - 3
Bangladesh	2017	2	b0	child is twin	dbl+lbl	0	5	0 - 4
Bangladesh	2022	2	b0	child is twin	dbl+lbl	0	4	0 - 3
Bangladesh	1993	3	b1	month of birth	dbl	0	12	1 - 12
Bangladesh	1996	3	b1	month of birth	dbl	0	12	1 - 12
Bangladesh	1999	3	b1	month of birth	dbl	0	12	1 - 12
Bangladesh	2004	3	b1	month of birth	dbl	0	12	1 - 12
Bangladesh	2007	3	b1	month of birth	dbl	0	12	1 - 12
Bangladesh	2011	3	b1	month of birth	dbl	0	12	1 - 12
Bangladesh	2014	3	b1	month of birth	dbl	0	12	1 - 12
Bangladesh	2017	3	b1	month of birth	dbl	0	12	1 - 12
Bangladesh	2022	3	b1	month of birth	dbl	0	12	1 - 12
Bangladesh	1993	4	b2	year of birth	dbl	0	38	57 - 94
Bangladesh	1996	4	b2	year of birth	dbl	0	38	60 - 97
Bangladesh	1999	4	b2	year of birth	dbl+lbl	0	38	0 - 99
Bangladesh	2004	4	b2	year of birth	dbl	0	39	1966 - 2004
Bangladesh	2007	4	b2	year of birth	dbl	0	38	1970 - 2007
Bangladesh	2011	4	b2	year of birth	dbl	0	38	1974 - 2011
Bangladesh	2014	4	b2	year of birth	dbl	0	37	1978 - 2014
Bangladesh	2017	4	b2	year of birth	dbl	0	38	1981 - 2018
Bangladesh	2022	4	b2	year of birth	dbl	0	39	1984 - 2022
Bangladesh	1993	5	b3	date of birth (cmc)	dbl	0	430	694 - 1131
Bangladesh	1996	5	b3	date of birth (cmc)	dbl	0	429	729 - 1166
Bangladesh	1999	5	b3	date of birth (cmc)	dbl	0	434	763 - 1203
Bangladesh	2004	5	b3	date of birth (cmc)	dbl	0	437	797 - 1253
Bangladesh	2007	5	b3	date of birth (cmc)	dbl	0	436	850 - 1291
Bangladesh	2011	5	b3	date of birth (cmc)	dbl	0	443	893 - 1344
Bangladesh	2014	5	b3	date of birth (cmc)	dbl	0	438	938 - 1378
Bangladesh	2017	5	b3	date of birth (cmc)	dbl	0	436	984 - 1419
Bangladesh	2022	5	b3	date of birth (cmc)	dbl	0	447	1010 - 1476
Bangladesh	1993	6	b4	sex of child	dbl+lbl	0	2	1 - 2
Bangladesh	1996	6	b4	sex of child	dbl+lbl	0	2	1 - 2
Bangladesh	1999	6	b4	sex of child	dbl+lbl	0	2	1 - 2
Bangladesh	2004	6	b4	sex of child	dbl+lbl	0	2	1 - 2
Bangladesh	2007	6	b4	sex of child	dbl+lbl	0	2	1 - 2
Bangladesh	2011	6	b4	sex of child	dbl+lbl	0	2	1 - 2
Bangladesh	2014	6	b4	sex of child	dbl+lbl	0	2	1 - 2
Bangladesh	2017	6	b4	sex of child	dbl+lbl	0	2	1 - 2
Bangladesh	2022	6	b4	sex of child	dbl+lbl	0	2	1 - 2
Bangladesh	1993	7	b5	child is alive	dbl+lbl	0	2	0 - 1
Bangladesh	1996	7	b5	child is alive	dbl+lbl	0	2	0 - 1
Bangladesh	1999	7	b5	child is alive	dbl+lbl	0	2	0 - 1
Bangladesh	2004	7	b5	child is alive	dbl+lbl	0	2	0 - 1
Bangladesh	2007	7	b5	child is alive	dbl+lbl	0	2	0 - 1
Bangladesh	2011	7	b5	child is alive	dbl+lbl	0	2	0 - 1
Bangladesh	2014	7	b5	child is alive	dbl+lbl	0	2	0 - 1
Bangladesh	2017	7	b5	child is alive	dbl+lbl	0	2	0 - 1
Bangladesh	2022	7	b5	child is alive	dbl+lbl	0	2	0 - 1
Bangladesh	1993	8	b6	age at death	dbl+lbl	26546	93	100 - 399
Bangladesh	1996	8	b6	age at death	dbl+lbl	24383	91	100 - 399
Bangladesh	1999	8	b6	age at death	dbl+lbl	26984	89	100 - 333
Bangladesh	2004	8	b6	age at death	dbl+lbl	28912	87	100 - 331
Bangladesh	2007	8	b6	age at death	dbl+lbl	26730	89	100 - 999
Bangladesh	2011	8	b6	age at death	dbl+lbl	41057	87	100 - 999
Bangladesh	2014	8	b6	age at death	dbl+lbl	39772	86	100 - 330
Bangladesh	2017	8	b6	age at death	dbl+lbl	43840	86	100 - 334
Bangladesh	2022	8	b6	age at death	dbl+lbl	60613	85	100 - 332
Bangladesh	1993	9	b7	age at death (months-imputed)	dbl	26538	56	0 - 348
Bangladesh	1996	9	b7	age at death (months-imputed)	dbl	24380	56	0 - 372
Bangladesh	1999	9	b7	age at death (months-imputed)	dbl	26972	55	0 - 396
Bangladesh	2004	9	b7	age at death (months-imputed)	dbl	28911	55	0 - 372
Bangladesh	2007	9	b7	age at death (months-imputed)	dbl	26730	58	0 - 384
Bangladesh	2011	9	b7	age at death (months, imputed)	dbl	41057	56	0 - 384
Bangladesh	2014	9	b7	age at death (months, imputed)	dbl	39770	55	0 - 360
Bangladesh	2017	9	b7	age at death (months, imputed)	dbl	43840	56	0 - 408
Bangladesh	2022	9	b7	age at death (months, imputed)	dbl	60613	55	0 - 384
Bangladesh	1993	10	b8	current age of child	dbl	6043	38	0 - 36
Bangladesh	1996	10	b8	current age of child	dbl	4964	38	0 - 36
Bangladesh	1999	10	b8	current age of child	dbl	4934	38	0 - 36
Bangladesh	2004	10	b8	current age of child	dbl	4686	38	0 - 36
Bangladesh	2007	10	b8	current age of child	dbl	3797	38	0 - 36
Bangladesh	2011	10	b8	current age of child	dbl	4776	39	0 - 37
Bangladesh	2014	10	b8	current age of child	dbl	4002	38	0 - 36
Bangladesh	2017	10	b8	current age of child	dbl	3988	37	0 - 35
Bangladesh	2022	10	b8	current age of child	dbl	4109	40	0 - 38
Bangladesh	1993	11	b9	who child lives with	dbl+lbl	6043	3	0 - 4
Bangladesh	1996	11	b9	who child lives with	dbl+lbl	4964	3	0 - 4
Bangladesh	1999	11	b9	who child lives with	dbl+lbl	4934	3	0 - 4
Bangladesh	2004	11	b9	child lives with whom	dbl+lbl	4686	3	0 - 4
Bangladesh	2007	11	b9	child lives with whom	dbl+lbl	3797	3	0 - 4
Bangladesh	2011	11	b9	child lives with whom	dbl+lbl	4776	3	0 - 4
Bangladesh	2014	11	b9	child lives with whom	dbl+lbl	4002	3	0 - 4
Bangladesh	2017	11	b9	child lives with whom	dbl+lbl	3988	3	0 - 4
Bangladesh	2022	11	b9	child lives with whom	dbl+lbl	4109	3	0 - 4
Bangladesh	1993	12	b10	completeness of information	dbl+lbl	0	8	1 - 8
Bangladesh	1996	12	b10	completeness of information	dbl+lbl	0	6	1 - 8
Bangladesh	1999	12	b10	completeness of information	dbl+lbl	0	7	1 - 8
Bangladesh	2004	12	b10	completeness of information	dbl+lbl	0	3	1 - 5
Bangladesh	2007	12	b10	completeness of information	dbl+lbl	0	3	1 - 5
Bangladesh	2011	12	b10	completeness of information	dbl+lbl	0	6	1 - 8
Bangladesh	2014	12	b10	completeness of information	dbl+lbl	0	6	1 - 8
Bangladesh	2017	12	b10	completeness of information	dbl+lbl	0	4	0 - 5
Bangladesh	2022	12	b10	completeness of information	dbl+lbl	0	5	0 - 6
Bangladesh	1993	13	b11	preceding birth interval	dbl	8577	168	7 - 270
Bangladesh	1996	13	b11	preceding birth interval	dbl	8109	164	6 - 300
Bangladesh	1999	13	b11	preceding birth interval	dbl	9405	177	9 - 246
Bangladesh	2004	13	b11	preceding birth interval	dbl	10189	179	8 - 231
Bangladesh	2007	13	b11	preceding birth interval	dbl	9893	196	8 - 318
Bangladesh	2011	13	b11	preceding birth interval (months)	dbl	16107	207	-3 - 246
Bangladesh	2014	13	b11	preceding birth interval (months)	dbl	16181	205	9 - 267
Bangladesh	2017	13	b11	preceding birth interval (months)	dbl	18241	215	7 - 291
Bangladesh	2022	13	b11	preceding birth interval (months)	dbl	27116	235	0 - 343
Bangladesh	1993	14	b12	succeeding birth interval	dbl	8614	168	7 - 270
Bangladesh	1996	14	b12	succeeding birth interval	dbl	8156	164	6 - 300
Bangladesh	1999	14	b12	succeeding birth interval	dbl	9442	177	9 - 246
Bangladesh	2004	14	b12	succeeding birth interval	dbl	10246	179	8 - 231
Bangladesh	2007	14	b12	succeeding birth interval	dbl	9947	196	8 - 318
Bangladesh	2011	14	b12	succeeding birth interval (months)	dbl	16183	207	-3 - 246
Bangladesh	2014	14	b12	succeeding birth interval (months)	dbl	16225	205	9 - 267
Bangladesh	2017	14	b12	succeeding birth interval (months)	dbl	18329	215	7 - 291
Bangladesh	2022	14	b12	succeeding birth interval (months)	dbl	27214	235	0 - 343
Bangladesh	1993	15	b13	flag for age at death	dbl+lbl	26538	5	0 - 8
Bangladesh	1996	15	b13	flag for age at death	dbl+lbl	24380	6	0 - 8
Bangladesh	1999	15	b13	flag for age at death	dbl+lbl	26972	7	0 - 8
Bangladesh	2004	15	b13	flag for age at death	dbl+lbl	28911	6	0 - 8
Bangladesh	2007	15	b13	flag for age at death	dbl+lbl	26730	6	0 - 8
Bangladesh	2011	15	b13	flag for age at death	dbl+lbl	41057	5	0 - 8
Bangladesh	2014	15	b13	flag for age at death	dbl+lbl	39770	4	0 - 8
Bangladesh	2017	15	b13	flag for age at death	dbl+lbl	43840	3	0 - 6
Bangladesh	2022	15	b13	flag for age at death	dbl+lbl	60613	2	0 - 0
Bangladesh	1993	16	b14	birth interval >= 4 years -na	dbl+lbl	32581	1
Bangladesh	1996	16	b14	birth interval >= 4 years	dbl+lbl	8088	3	0 - 1
Bangladesh	1999	16	b14	birth interval >= 4 years -na	dbl	31906	1
Bangladesh	2004	16	b15	live birth between births	dbl+lbl	10173	3	0 - 1
Bangladesh	2007	16	b15	live birth between births	dbl+lbl	9849	4	0 - 9
Bangladesh	2011	16	b15	live birth between births	dbl+lbl	16014	4	0 - 9
Bangladesh	2014	16	b15	live birth between births	dbl+lbl	16087	3	0 - 1
Bangladesh	2017	16	b15	live birth between births	dbl+lbl	18142	3	0 - 1
Bangladesh	2022	16	b15	live birth between births	dbl+lbl	0	2	0 - 1
Bangladesh	1993	17	b15	live birth between births -na	dbl+lbl	32581	1
Bangladesh	1996	17	b15	live birth between births	dbl+lbl	25432	3	0 - 1
Bangladesh	1999	17	b15	live birth between births	dbl+lbl	9366	2	0 - 0
Bangladesh	2004	17	b16	child's line number in household	dbl+lbl	4686	29	0 - 31
Bangladesh	2007	17	b16	child's line number in household	dbl+lbl	3797	29	0 - 28
Bangladesh	2011	17	b16	child's line number in household	dbl+lbl	4776	31	0 - 29
Bangladesh	2014	17	b16	child's line number in household	dbl+lbl	4002	24	0 - 24
Bangladesh	2017	17	b16	child's line number in household	dbl+lbl	3988	26	0 - 29
Bangladesh	2022	17	b16	child's line number in household	dbl+lbl	4109	25	0 - 25
Bangladesh	2017	18	b17	day of birth	dbl	0	31	1 - 31
Bangladesh	2022	18	b17	day of birth	dbl	0	31	1 - 31
Bangladesh	2017	19	b18	century day code of birth (cdc)	dbl	0	11428	29926 - 43167
Bangladesh	2022	19	b18	century day code of birth (cdc)	dbl	0	11617	30713 - 44899
Bangladesh	2017	20	b19	current age of child in months (months since birth for dead children)	dbl	0	432	0 - 431
Bangladesh	2022	20	b19	current age of child in months (months since birth for dead children)	dbl	0	445	0 - 462
Bangladesh	2017	21	b20	na - duration of pregnancy	dbl	47828	1
Bangladesh	2022	21	b20	duration of pregnancy in months	dbl	0	6	5 - 10
Bangladesh	2022	22	b21	duration of pregnancy	dbl	0	22	124 - 210

From the above table we get an overall snapshot of the birth history variables. We see that the variables b1-b13 are common in all the 9 bdbr datasets. Notably bdbr 2001 and 2006 have some extra variables that are not available in other rounds. Next, we look at the other labelled variables which are common across bdbr in more details. We would like to see if the value labels of the common birth history variables are similar across the bdbr datasets.

b0 - child is twin

We check the value labels of b0 variable that denotes whether the child is twin. First we create a nested tibble of b0’s value labels.

# Create the data dictionary for b0 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_b0 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(b0) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_b0)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "b0", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 9: Data dictionary of b0 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	b0	0	[0] single birth	[0] single birth	[0] single birth	[0] single birth	[0] single birth	[0] single birth	[0] single birth	[0] single birth	[0] single birth
Bangladesh	b0	1	[1] 1st of multiple	[1] 1st of multiple	[1] 1st of multiple	[1] 1st of multiple	[1] 1st of multiple	[1] 1st of multiple	[1] 1st of multiple	[1] 1st of multiple	[1] 1st of multiple
Bangladesh	b0	2	[2] 2nd of multiple	[2] 2nd of multiple	[2] 2nd of multiple	[2] 2nd of multiple	[2] 2nd of multiple	[2] 2nd of multiple	[2] 2nd of multiple	[2] 2nd of multiple	[2] 2nd of multiple
Bangladesh	b0	3	[3] 3rd of multiple	[3] 3rd of multiple	[3] 3rd of multiple	[3] 3rd of multiple	[3] 3rd of multiple	[3] 3rd of multiple	[3] 3rd of multiple	[3] 3rd of multiple	[3] 3rd of multiple
Bangladesh	b0	4	[4] 4th of multiple	[4] 4th of multiple	[4] 4th of multiple	[4] 4th of multiple	[4] 4th of multiple	[4] 4th of multiple	[4] 4th of multiple	[4] 4th of multiple	[4] 4th of multiple
Bangladesh	b0	5	[5] 5th of multiple	[5] 5th of multiple	[5] 5th of multiple	[5] 5th of multiple	[5] 5th of multiple	[5] 5th of multiple	[5] 5th of multiple	[5] 5th of multiple	[5] 5th of multiple

We can see the value labels of b0 in the above table. We see that the value labels are same across all the bdbr datasets.

b2 - year of birth

We see that the b2 variable has value labels only for bdbr 1999. Therefore, we check the value labels of the variable during this round.

# Create the data dictionary for b2 in bdbr 1999
bdbr1_pre_tmp0$bdbr_data$bdbr_1999 |> 
  select(b2) |> 
  look_for(details = "full") |> 
  lookfor_to_long_format() |> 
  convert_list_columns_to_character() |> 
  select(-c(pos, levels, class:n_na)) |> 
  qflextable() |> 
  autofit()

Table 10: Data dictionary of b2 in bdbr 1999

variable	label	col_type	missing	value_labels	unique_values	range
b2	year of birth	dbl+lbl	0	[0] year 2000	38	0 - 99

Note that, the birth year 0 which corresponds to the year 2000 is denoted with the label 2000. That is the only label in the variable.

# Check the distribution of b2 in bdbr 1999
bdbr1_pre_tmp0$bdbr_data$bdbr_1999 |> 
  tabyl(b2) |> 
  adorn_totals() |> 
  adorn_pct_formatting() |> 
  qflextable() |> 
  autofit()

Table 11: Distribution of b2 in bdbr 1999

b2	n	percent
0	85	0.3%
63	2	0.0%
64	16	0.1%
65	48	0.2%
66	69	0.2%
67	125	0.4%
68	147	0.5%
69	245	0.8%
70	295	0.9%
71	426	1.3%
72	371	1.2%
73	485	1.5%
74	584	1.8%
75	608	1.9%
76	628	2.0%
77	744	2.3%
78	707	2.2%
79	893	2.8%
80	846	2.7%
81	999	3.1%
82	975	3.1%
83	1119	3.5%
84	1150	3.6%
85	1197	3.8%
86	1210	3.8%
87	1384	4.3%
88	1360	4.3%
89	1489	4.7%
90	1340	4.2%
91	1370	4.3%
92	1488	4.7%
93	1411	4.4%
94	1255	3.9%
95	1365	4.3%
96	1319	4.1%
97	1383	4.3%
98	1440	4.5%
99	1328	4.2%
Total	31906	100.0%

For more clarity we checked the univariate distribution of birth year above in Table 11. It is clear that the survey years during 1900’s have been represented by denoting just the last two digits and the year 2000 is represented by 0.

b4 - sex of child

We check the value labels of b4 variable which gives the sex of the child. First we create a nested tibble of b4’s value labels.

# Create the data dictionary for b4 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_b4 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(b4) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_b4)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "b4", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 12: Data dictionary of b4 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	b4	1	[1] male	[1] male	[1] male	[1] male	[1] male	[1] male	[1] male	[1] male	[1] male
Bangladesh	b4	2	[2] female	[2] female	[2] female	[2] female	[2] female	[2] female	[2] female	[2] female	[2] female

We can see the value labels of b4 in the above table. The value labels are same across all the bdbr datasets.

b5 - child is alive

We check the value labels of b5 variable which gives the survival status of the child. First we create a nested tibble of b5’s value labels.

# Create the data dictionary for b5 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_b5 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(b5) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_b5)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "b5", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 13: Data dictionary of b5 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	b5	0	[0] no	[0] no	[0] no	[0] no	[0] no	[0] no	[0] no	[0] no	[0] no
Bangladesh	b5	1	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes

The above table shows that the value labels of survival status of child are same across all the bdbr datasets.

b6 - age at death

We check the value labels of b6 variable which shows the age at death of children. Note that this variable has many missing values across all bdbr rounds as not all children experienced mortality throughout their lifetime. First we create a nested tibble of b6’s value labels.

# Create the data dictionary for b5 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_b6 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(b6) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_b6)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "b6", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 14: Data dictionary of b6 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	b6	100						[100] died on day of birth	[100] died on day of birth	[100] died on day of birth	[100] died on day of birth
Bangladesh	b6	101						[101] days: 1	[101] days: 1	[101] days: 1	[101] days: 1
Bangladesh	b6	199					[199] days, missing	[199] days: number missing	[199] days: number missing	[199] days: number missing	[199] days: number missing
Bangladesh	b6	201						[201] months: 1	[201] months: 1	[201] months: 1	[201] months: 1
Bangladesh	b6	299					[299] months, missing	[299] months: number missing	[299] months: number missing	[299] months: number missing	[299] months: number missing
Bangladesh	b6	301						[301] years: 1	[301] years: 1	[301] years: 1	[301] years: 1
Bangladesh	b6	399					[399] years, missing	[399] years: number missing	[399] years: number missing	[399] years: number missing	[399] years: number missing
Bangladesh	b6	997	[997] inconsistent	[997] inconsistent	[997] inconsistent	[997] inconsistent	[997] inconsistent	[997] inconsistent	[997] inconsistent	[997] inconsistent	[997] inconsistent
Bangladesh	b6	998	[998] don't know	[998] don't know	[998] don't know	[998] don't know	[998] don't know	[998] don't know	[998] don't know	[998] don't know	[998] don't know

The above table shows that the value labels of age at death of child are in two groups. First, they are same for bdbr 1996, 2001 and 2006 and and then for bdbr 2011, 2016 and 2022.

b9 - child lives with whom

We check the value labels of b9 variable which gives info on who the child lives with. First we create a nested tibble of b9’s value labels.

# Create the data dictionary for b9 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_b9 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(b9) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_b9)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "b9", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 15: Data dictionary of b9 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	b9	0	[0] respondent	[0] respondent	[0] respondent	[0] respondent	[0] respondent	[0] respondent	[0] respondent	[0] respondent	[0] respondent
Bangladesh	b9	1	[1] father	[1] father	[1] father	[1] father	[1] father	[1] father	[1] father	[1] father	[1] father
Bangladesh	b9	2	[2] other relative	[2] other relative	[2] other relative	[2] other relative	[2] other relative	[2] other relative	[2] other relative	[2] other relative	[2] other relative
Bangladesh	b9	3	[3] someone else	[3] someone else	[3] someone else	[3] someone else	[3] someone else	[3] someone else	[3] someone else	[3] someone else	[3] someone else
Bangladesh	b9	4	[4] lives elsewhere	[4] lives elsewhere	[4] lives elsewhere	[4] lives elsewhere	[4] lives elsewhere	[4] lives elsewhere	[4] lives elsewhere	[4] lives elsewhere	[4] lives elsewhere

We can see in the above table that the value labels of b9 are same across all the bdbr datasets.

b10 - completeness of information

We check the value labels of b10 variable which gives the completeness of birth history information. First we create a nested tibble of b10’s value labels.

# Create the data dictionary for b10 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_b10 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(b10) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_b10)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "b10", .before = 2) 

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |>
  align(align = "left", part = "all") |> 
  autofit()

Table 16: Data dictionary of b10 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	b10	0								[0] month, year and day	[0] month, year and day
Bangladesh	b10	1	[1] month and year	[1] month and year	[1] month and year	[1] month and year	[1] month and year	[1] month and year - information complete	[1] month and year - information complete	[1] month and year - information complete	[1] month and year - information complete
Bangladesh	b10	2	[2] month and age -y imp	[2] month and age -y imp	[2] month and age -y imp	[2] month and age -y imp	[2] month and age -y imp	[2] month and age - year imputed	[2] month and age - year imputed	[2] month and age - year imputed	[2] month and age - year imputed
Bangladesh	b10	3	[3] year and age - m imp	[3] year and age - m imp	[3] year and age - m imp	[3] year and age - m imp	[3] year and age - m imp	[3] year and age - month imputed	[3] year and age - month imputed	[3] year and age - month imputed	[3] year and age - month imputed
Bangladesh	b10	4	[4] y & age - y ignored	[4] y & age - y ignored	[4] y & age - y ignored	[4] y & age - y ignored	[4] y & age - y ignored	[4] year and age - year ignored	[4] year and age - year ignored	[4] year and age - year ignored	[4] year and age - year ignored
Bangladesh	b10	5	[5] year - a, m imp	[5] year - a, m imp	[5] year - a, m imp	[5] year - a, m imp	[5] year - a, m imp	[5] year - age/month imputed	[5] year - age/month imputed	[5] year - age/month imputed	[5] year - age/month imputed
Bangladesh	b10	6	[6] age - y, m imp	[6] age - y, m imp	[6] age - y, m imp	[6] age - y, m imp	[6] age - y, m imp	[6] age - year/month imputed	[6] age - year/month imputed	[6] age - year/month imputed	[6] age - year/month imputed
Bangladesh	b10	7	[7] month - a, y imp	[7] month - a, y imp	[7] month - a, y imp	[7] month - a, y imp	[7] month - a, y imp	[7] month - age/year imputed	[7] month - age/year imputed	[7] month - age/year imputed	[7] month - age/year imputed
Bangladesh	b10	8	[8] none - all imp	[8] none - all imp	[8] none - all imp	[8] none - all imp	[8] none - all imp	[8] none - all imputed	[8] none - all imputed	[8] none - all imputed	[8] none - all imputed

We can see in the above table that the value labels of b10 are same across bdbr 1996, 2001, 2006 and 2011 datasets. Then they are same for bdbr 2016 and 2022.

Checking the Common independent variables before harmonization

Next we start documenting the common independent variables. First we will check the data dictionary of the common independent variables. Then we will check them variable wise.

# We check the common independent vars in all bdbr datasets.
# First we create the data dictionary in nested tibble.
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |>
  mutate(lookfor_comindvars = map(
    bdbr_data,
    \(df) {
      df |> 
        # select the common independent variables
        select(v106, v011, v501, v701, v025, v151, v152) |> 
        lookfor(details = "full") |> 
        select(-c(levels:n_na)) |> 
        # For correctly viewing the range column in data dictionary
        convert_list_columns_to_character()
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary 
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_comindvars)) |> 
  arrange(pos)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 17: Data dictionary of common independent variables across the bdbr rounds

ctr_name	svy_year	pos	variable	label	col_type	missing	unique_values	range
Bangladesh	1993	1	v106	highest educational level	dbl+lbl	0	4	0 - 3
Bangladesh	1996	1	v106	highest educational level	dbl+lbl	0	4	0 - 3
Bangladesh	1999	1	v106	highest educational level	dbl+lbl	0	4	0 - 3
Bangladesh	2004	1	v106	highest educational level	dbl+lbl	0	4	0 - 3
Bangladesh	2007	1	v106	highest educational level	dbl+lbl	0	5	0 - 9
Bangladesh	2011	1	v106	highest educational level	dbl+lbl	0	4	0 - 3
Bangladesh	2014	1	v106	highest educational level	dbl+lbl	0	4	0 - 3
Bangladesh	2017	1	v106	highest educational level	dbl+lbl	0	4	0 - 3
Bangladesh	2022	1	v106	highest educational level	dbl+lbl	0	4	0 - 3
Bangladesh	1993	2	v011	date of birth (cmc)	dbl	0	421	529 - 949
Bangladesh	1996	2	v011	date of birth (cmc)	dbl	0	417	565 - 987
Bangladesh	1999	2	v011	date of birth (cmc)	dbl	0	423	600 - 1022
Bangladesh	2004	2	v011	date of birth (cmc)	dbl	0	424	650 - 1073
Bangladesh	2007	2	v011	date of birth (cmc)	dbl	0	423	688 - 1111
Bangladesh	2011	2	v011	date of birth (cmc)	dbl	0	424	740 - 1163
Bangladesh	2014	2	v011	date of birth (cmc)	dbl	0	422	776 - 1197
Bangladesh	2017	2	v011	date of birth (cmc)	dbl	0	422	815 - 1237
Bangladesh	2022	2	v011	date of birth (cmc)	dbl	0	423	871 - 1293
Bangladesh	1993	3	v501	current marital status	dbl+lbl	0	3	1 - 4
Bangladesh	1996	3	v501	current marital status	dbl+lbl	0	3	1 - 4
Bangladesh	1999	3	v501	current marital status	dbl+lbl	0	4	1 - 5
Bangladesh	2004	3	v501	current marital status	dbl+lbl	0	4	1 - 5
Bangladesh	2007	3	v501	current marital status	dbl+lbl	0	4	1 - 5
Bangladesh	2011	3	v501	current marital status	dbl+lbl	0	4	1 - 5
Bangladesh	2014	3	v501	current marital status	dbl+lbl	0	4	1 - 5
Bangladesh	2017	3	v501	current marital status	dbl+lbl	0	4	1 - 5
Bangladesh	2022	3	v501	current marital status	dbl+lbl	0	4	1 - 5
Bangladesh	1993	4	v701	partner's education level	dbl+lbl	17	6	0 - 8
Bangladesh	1996	4	v701	partner's education level	dbl+lbl	35	6	0 - 8
Bangladesh	1999	4	v701	partner's education level	dbl+lbl	47	6	0 - 8
Bangladesh	2004	4	v701	partner's education level	dbl+lbl	35	5	0 - 3
Bangladesh	2007	4	v701	partner's education level	dbl+lbl	10	6	0 - 9
Bangladesh	2011	4	v701	husband/partner's education level	dbl+lbl	0	5	0 - 9
Bangladesh	2014	4	v701	husband/partner's education level	dbl+lbl	3	5	0 - 3
Bangladesh	2017	4	v701	husband/partner's education level	dbl+lbl	2935	6	0 - 8
Bangladesh	2022	4	v701	husband/partner's education level	dbl+lbl	23811	6	0 - 8
Bangladesh	1993	5	v025	type of place of residence	dbl+lbl	0	2	1 - 2
Bangladesh	1996	5	v025	type of place of residence	dbl+lbl	0	2	1 - 2
Bangladesh	1999	5	v025	type of place of residence	dbl+lbl	0	2	1 - 2
Bangladesh	2004	5	v025	type of place of residence	dbl+lbl	0	2	1 - 2
Bangladesh	2007	5	v025	type of place of residence	dbl+lbl	0	2	1 - 2
Bangladesh	2011	5	v025	type of place of residence	dbl+lbl	0	2	1 - 2
Bangladesh	2014	5	v025	type of place of residence	dbl+lbl	0	2	1 - 2
Bangladesh	2017	5	v025	type of place of residence	dbl+lbl	0	2	1 - 2
Bangladesh	2022	5	v025	type of place of residence	dbl+lbl	0	2	1 - 2
Bangladesh	1993	6	v151	sex of household head	dbl+lbl	0	2	1 - 2
Bangladesh	1996	6	v151	sex of household head	dbl+lbl	0	2	1 - 2
Bangladesh	1999	6	v151	sex of household head	dbl+lbl	0	2	1 - 2
Bangladesh	2004	6	v151	sex of household head	dbl+lbl	0	2	1 - 2
Bangladesh	2007	6	v151	sex of household head	dbl+lbl	0	2	1 - 2
Bangladesh	2011	6	v151	sex of household head	dbl+lbl	0	2	1 - 2
Bangladesh	2014	6	v151	sex of household head	dbl+lbl	0	2	1 - 2
Bangladesh	2017	6	v151	sex of household head	dbl+lbl	0	2	1 - 2
Bangladesh	2022	6	v151	sex of household head	dbl+lbl	0	2	1 - 2
Bangladesh	1993	7	v152	age of household head	dbl+lbl	0	83	11 - 98
Bangladesh	1996	7	v152	age of household head	dbl	1	79	9 - 98
Bangladesh	1999	7	v152	age of household head	dbl	12	79	11 - 95
Bangladesh	2004	7	v152	age of household head	dbl+lbl	0	82	10 - 98
Bangladesh	2007	7	v152	age of household head	dbl+lbl	0	79	14 - 99
Bangladesh	2011	7	v152	age of household head	dbl+lbl	0	80	13 - 95
Bangladesh	2014	7	v152	age of household head	dbl+lbl	0	75	12 - 95
Bangladesh	2017	7	v152	age of household head	dbl+lbl	0	78	15 - 95
Bangladesh	2022	7	v152	age of household head	dbl+lbl	0	77	15 - 95

From the above table we get an overall snapshot of the common independent variables. We see that majority of the have different number of value labels across the bdbr datasets. Only v025 and v151 have the same number of value labels across bdbr rounds. Next, we look at the labelled class variables among these common variables in more details. We would like to see if the value labels and codes of the common independent variables are similar across the bdbr datasets.

v106 - Mother’s education level

We check the value labels of v106 variable that denotes the highest education level of mother. First we create a nested tibble of v106’s value labels.

# Create the data dictionary for v106 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_v106 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(v106) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_v106)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "v106", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 18: Data dictionary of v106 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	v106	0	[0] no education	[0] no education	[0] no education	[0] no education	[0] no education	[0] no education	[0] no education	[0] no education	[0] no education
Bangladesh	v106	1	[1] primary	[1] primary	[1] primary	[1] primary	[1] primary	[1] primary	[1] primary	[1] primary	[1] primary
Bangladesh	v106	2	[2] secondary	[2] secondary	[2] secondary	[2] secondary	[2] secondary	[2] secondary	[2] secondary	[2] secondary	[2] secondary
Bangladesh	v106	3	[3] higher	[3] higher	[3] higher	[3] higher	[3] higher	[3] higher	[3] higher	[3] higher	[3] higher

We can see the value labels of v106 are similar across all the bdbr datasets. Although bdbr 2007 had 5 values as seen in Table 17, it has 4 value labels.

v011 - Date of birth (in CMC)

The v011 variable, which has the dob of mothers in cmc, is a numeric variable. Let’s check the range of these values in further details such as checking for outliers. First let’s create a nested tibble of the summary statistics of v011 variable.

# Create the summary statistics for v011 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(skim_v011 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(v011) |> 
        skim_without_charts() |> 
        as_tibble() |> 
        select(-c(skim_type, complete_rate)) |> 
        rename(
          variable = 1,
          n_miss = 2,
          mean = 3,
          sd = 4,
          min = 5,
          p25 = 6,
          p50 = 7,
          p75 = 8,
          max = 9
        )
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(skim_v011)) |> 
  # Make variable values have one decimal point 
  mutate(
    mean = sprintf("%.1f", mean),
    sd = sprintf("%.1f", sd)
  )

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 19: Data dictionary of v011 across the bdbr rounds

ctr_name	svy_year	variable	n_miss	mean	sd	min	p25	p50	p75	max
Bangladesh	1993	v011	0	712.9	99.2	529	633	712	792	949
Bangladesh	1996	v011	0	750.0	100.4	565	668	748	829	987
Bangladesh	1999	v011	0	778.1	102.1	600	693	776	856	1022
Bangladesh	2004	v011	0	826.7	103.7	650	740	824	908	1073
Bangladesh	2007	v011	0	860.8	103.6	688	776	857	943	1111
Bangladesh	2011	v011	0	914.7	102.6	740	829	912	998	1163
Bangladesh	2014	v011	0	948.2	100.8	776	862	947	1027	1197
Bangladesh	2017	v011	0	983.7	100.6	815	897	981	1060	1237
Bangladesh	2022	v011	0	1041.6	96.5	871	963	1038	1115	1293

v501 - Mother’s marital status

We check the value labels of v501 variable which gives the current marital status of mother. First we create a nested tibble of v501’s value labels.

# Create the data dictionary for v501 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_v501 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(v501) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_v501)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "v501", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 20: Data dictionary of v501 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	v501	0	[0] never married	[0] never married	[0] never married	[0] never married	[0] never married	[0] never in union	[0] never in union	[0] never in union	[0] never in union
Bangladesh	v501	1	[1] married	[1] married	[1] married	[1] married	[1] married	[1] married	[1] married	[1] married	[1] married
Bangladesh	v501	2	[2] living together	[2] living together	[2] living together	[2] living together	[2] living together	[2] living with partner	[2] living with partner	[2] living with partner	[2] living with partner
Bangladesh	v501	3	[3] widowed	[3] widowed	[3] widowed	[3] widowed	[3] widowed	[3] widowed	[3] widowed	[3] widowed	[3] widowed
Bangladesh	v501	4	[4] divorced	[4] divorced	[4] divorced	[4] divorced	[4] divorced	[4] divorced	[4] divorced	[4] divorced	[4] divorced
Bangladesh	v501	5	[5] not living together	[5] not living together	[5] not living together	[5] not living together	[5] not living together	[5] no longer living together/separated	[5] no longer living together/separated	[5] no longer living together/separated	[5] no longer living together/separated

All the bdbr rounds have the same number of value label codes. The bdbr 1993, 1996, 1999, 2004 and 2007 have the same set of value label texts. Then bdbr 2011, 2014, 2017 and 2022 have another set of similar value labels.

v701 - Husband/Partner’s education level

We check the value labels of v701 variable which gives the current marital status of mother. First we create a nested tibble of v701’s value labels.

# Create the data dictionary for v701 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_v701 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(v701) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_v701)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "v701", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 21: Data dictionary of v701 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	v701	0	[0] no education	[0] no education	[0] no education	[0] no education	[0] no education	[0] no education	[0] no education	[0] no education	[0] no education
Bangladesh	v701	1	[1] primary	[1] primary	[1] primary	[1] primary	[1] primary	[1] primary	[1] primary	[1] primary	[1] primary
Bangladesh	v701	2	[2] secondary	[2] secondary	[2] secondary	[2] secondary	[2] secondary	[2] secondary	[2] secondary	[2] secondary	[2] secondary
Bangladesh	v701	3	[3] higher	[3] higher	[3] higher	[3] higher	[3] higher	[3] higher	[3] higher	[3] higher	[3] higher
Bangladesh	v701	8	[8] don't know	[8] don't know	[8] don't know	[8] don't know	[8] don't know	[8] don't know	[8] don't know	[8] don't know	[8] don't know

We see that all the bdbr rounds have 5 value labels and they are same across all rounds.

v025 - Type of place of residence

We check the value labels of v025 variable which shows if a household belongs to rural or urban psu. First we create a nested tibble of v025’s value labels.

# Create the data dictionary for v025 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_v025 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(v025) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(c(ctr_name, svy_year, lookfor_v025)) |> 
  unnest(cols = c(lookfor_v025)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "v025", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 22: Data dictionary of v025 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	v025	1	[1] urban	[1] urban	[1] urban	[1] urban	[1] urban	[1] urban	[1] urban	[1] urban	[1] urban
Bangladesh	v025	2	[2] rural	[2] rural	[2] rural	[2] rural	[2] rural	[2] rural	[2] rural	[2] rural	[2] rural

The values labels and codes for v025 are same across all the bdbr rounds.

v151 - Sex of household head

We check the value labels of v151 variable which gives the sex of the household head. First we create a nested tibble of v151’s value labels.

# Create the data dictionary for v151 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_v151 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(v151) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_v151)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "v151", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 23: Data dictionary of v151 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	v151	1	[1] male	[1] male	[1] male	[1] male	[1] male	[1] male	[1] male	[1] male	[1] male
Bangladesh	v151	2	[2] female	[2] female	[2] female	[2] female	[2] female	[2] female	[2] female	[2] female	[2] female

The values labels and codes for v151 are same across all the bdbr rounds.

v152 - Age of household head

Interestingly, we see v152 (a continuous variable) has value labels for all rounds except bdbr 1996 and 1999. Therefore, we check the value labels of v152 for those rounds. First we create a nested tibble of v152’s value labels.

# Create the data dictionary for v152 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_v152 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(v152) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_v152)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "v152", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 24: Data dictionary of v152 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	v152	96	[96] 96+
Bangladesh	v152	97				[97] 97+	[97] 97+	[97] 97+	[97] 97+	[97] 97+	[97] 97+
Bangladesh	v152	98				[98] dk	[98] dk	[98] don't know	[98] don't know	[98] don't know	[98] don't know
Bangladesh	v152

We can see that the value labels of v152 are mostly for missing values. However, since v152 has no missing values across the bdbr rounds, we need not be concerned about them.

Checking the Social group variables before harmonization

Now we document the social group variables and then harmonize them. Upon manually checking the full data dictionaries of each bdbr dataset we find the following variables - religion and ethnicity. First we will check the data dictionary of these social group variables. Then we will check them variable wise.

Note: Only bdbr 2017 and 2022 have the native language of hh respondent variable. Since, they are not available for majority of rounds we do not use it.

# We check the social group vars in all bdbr datasets.
# First we create the data dictionary in nested tibble.
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |>
  mutate(lookfor_socgrp = map(
    bdbr_data,
    \(df) {
      df |> 
        # select the common independent variables
        select(v130, v131) |> 
        lookfor(details = "full") |> 
        select(-c(levels:n_na)) |> 
        # For correctly viewing the range column in data dictionary
        convert_list_columns_to_character()
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary 
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_socgrp)) |> 
  arrange(pos)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 25: Data dictionary of social group variables across the bdbr rounds

ctr_name	svy_year	pos	variable	label	col_type	missing	unique_values	range
Bangladesh	1993	1	v130	religion	dbl+lbl	0	5	1 - 5
Bangladesh	1996	1	v130	religion	dbl+lbl	6	5	1 - 4
Bangladesh	1999	1	v130	religion	dbl+lbl	1	6	1 - 6
Bangladesh	2004	1	v130	religion	dbl+lbl	17	5	1 - 4
Bangladesh	2007	1	v130	religion	dbl+lbl	0	6	1 - 99
Bangladesh	2011	1	v130	religion	dbl+lbl	0	4	1 - 4
Bangladesh	2014	1	v130	religion	dbl+lbl	3	6	1 - 96
Bangladesh	2017	1	v130	religion	dbl+lbl	0	4	1 - 4
Bangladesh	2022	1	v130	religion	dbl+lbl	0	5	1 - 96
Bangladesh	1993	2	v131	ethnicity -na	dbl+lbl	32581	1
Bangladesh	1996	2	v131	ethnicity na	dbl	29344	1
Bangladesh	1999	2	v131	ethnicity na	dbl	31906	1
Bangladesh	2004	2	v131	ethnicity -na	dbl	33597	1
Bangladesh	2007	2	v131	[na] ethnicity	dbl	30527	1
Bangladesh	2011	2	v131	na - ethnicity	dbl	45833	1
Bangladesh	2014	2	v131	na - ethnicity	dbl	43772	1
Bangladesh	2017	2	v131	na - ethnicity	dbl	47828	1
Bangladesh	2022	2	v131	na - ethnicity	dbl	64722	1

The above table gives an overall snapshot of the social group variables. Note that, although there is an ethnicity variable, all its values are missing. Effectively, only the religion variable is available for use in case of Bangladesh. Note that, the religion variable has some missing values in the bdbr 1996, 1999, 2004 and 2014 datasets. Next, we look at just the religion variable for examining the value labels across the bdbr datasets.

v130 - Religion of hh head

We check the value labels of the first social group variable v130, which gives the religion of household head. First we create a nested tibble of v130’s value labels.

# Create the data dictionary for v130 in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_v130 = map(
    bdbr_data,
    \(df) {
      df |> 
        select(v130) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_v130)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdbr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "v130", .before = 2)

# Convert the tibble to flextable for easy viewing
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 26: Data dictionary of v130 across the bdbr rounds

ctr_name	var_name	label_num	bdbr_1993	bdbr_1996	bdbr_1999	bdbr_2004	bdbr_2007	bdbr_2011	bdbr_2014	bdbr_2017	bdbr_2022
Bangladesh	v130	1	[1] islam	[1] islam	[1] islam	[1] islam	[1] islam	[1] islam	[1] islam	[1] islam	[1] islam
Bangladesh	v130	2	[2] christianity	[2] hinduism	[2] hinduism	[2] hinduism	[2] hinduism	[2] hinduism	[2] hinduism	[2] hinduism	[2] hinduism
Bangladesh	v130	3	[3] hinduism	[3] buddhism	[3] buddhism	[3] buddhism	[3] buddhism	[3] buddhism	[3] buddhism	[3] buddhism	[3] buddhist
Bangladesh	v130	4	[4] buddhism	[4] christianity	[4] christianity	[4] christianity	[4] christianity	[4] christianity	[4] christianity	[4] christianity	[4] christianity
Bangladesh	v130	5	[5] other	[5] other	[5]
Bangladesh	v130	6			[6] other	[6] other
Bangladesh	v130	96					[96] other	[96] other	[96] other	[96] other	[96] others

Evidently, the values labels of v130 are different across the bdbr 1993, 1996, 1999 and 2004 rounds. The value labels of bdbr 2007, 2011, 2014, 2017 and 2022 are similar.

To harmonize the religion variable we can use the following value labels -

• 1 Islam
• 2 Others

Correcting year-related variables

The year-related variables might have different formatting in each survey. Therefore, we need to check and harmonize them before appending the datasets.

# First we create the data dictionary of year-related vars in nested tibble
bdbr1_pre_tmp1 <- bdbr1_pre_tmp0 |> 
  mutate(lookfor_year = map(
    bdbr_data,
    \(df) {
      df |> 
        select(c(b2, v007, v010)) |> 
        look_for(details = "full") |> 
        # For correctly viewing the range column in data dictionary
        convert_list_columns_to_character() |> 
        select(-c(levels:n_na))
    }
  ))
bdbr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary 
bdbr1_pre_tmp2 <- bdbr1_pre_tmp1 |> 
  select(-c(unf, bdbr_data, n_births)) |> 
  unnest(cols = c(lookfor_year)) |> 
  arrange(pos)
# Convert and view the tibble as flextable
bdbr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 27: Data dictionary of year-related variables across the bdbr rounds

ctr_name	svy_year	pos	variable	label	col_type	missing	unique_values	range
Bangladesh	1993	1	b2	year of birth	dbl	0	38	57 - 94
Bangladesh	1996	1	b2	year of birth	dbl	0	38	60 - 97
Bangladesh	1999	1	b2	year of birth	dbl+lbl	0	38	0 - 99
Bangladesh	2004	1	b2	year of birth	dbl	0	39	1966 - 2004
Bangladesh	2007	1	b2	year of birth	dbl	0	38	1970 - 2007
Bangladesh	2011	1	b2	year of birth	dbl	0	38	1974 - 2011
Bangladesh	2014	1	b2	year of birth	dbl	0	37	1978 - 2014
Bangladesh	2017	1	b2	year of birth	dbl	0	38	1981 - 2018
Bangladesh	2022	1	b2	year of birth	dbl	0	39	1984 - 2022
Bangladesh	1993	2	v007	year of interview	dbl	0	2	93 - 94
Bangladesh	1996	2	v007	year of interview	dbl	0	2	96 - 97
Bangladesh	1999	2	v007	year of interview	dbl+lbl	0	2	0 - 99
Bangladesh	2004	2	v007	year of interview	dbl	0	1	2004 - 2004
Bangladesh	2007	2	v007	year of interview	dbl+lbl	0	1	2007 - 2007
Bangladesh	2011	2	v007	year of interview	dbl	0	1	2011 - 2011
Bangladesh	2014	2	v007	year of interview	dbl	0	1	2014 - 2014
Bangladesh	2017	2	v007	year of interview	dbl	0	2	2017 - 2018
Bangladesh	2022	2	v007	year of interview	dbl	0	1	2022 - 2022
Bangladesh	1993	3	v010	respondent's year of birth	dbl	0	36	44 - 79
Bangladesh	1996	3	v010	respondent's year of birth	dbl	0	36	47 - 82
Bangladesh	1999	3	v010	respondent's year of birth	dbl	0	37	49 - 85
Bangladesh	2004	3	v010	respondent's year of birth	dbl	0	36	1954 - 1989
Bangladesh	2007	3	v010	respondent's year of birth	dbl	0	36	1957 - 1992
Bangladesh	2011	3	v010	respondent's year of birth	dbl	0	36	1961 - 1996
Bangladesh	2014	3	v010	respondent's year of birth	dbl	0	36	1964 - 1999
Bangladesh	2017	3	v010	respondent's year of birth	dbl	0	37	1967 - 2003
Bangladesh	2022	3	v010	respondent's year of birth	dbl	0	36	1972 - 2007

We find that the year variables before 2001 do not have correct values. Just look at the year of interview variable. We see that the year of interview values have only the last two year digits corresponding to survey years for rounds before 2001. Therefore, we need to reformat them before appending the datasets.

Also, a few year variables in some rounds have value labels. We examine them below.

b2 - Year of birth

We see that the b2 variable has some value labels only for bdbr 1999. Therefore, we check those value labels for anything strange.

# Check the labels for b2 in bdbr 1999
bdbr1_pre_tmp0$bdbr_data$bdbr_1999 |> 
  select(b2) |> 
  look_for(details = "full") |> 
  lookfor_to_long_format() |>
  convert_list_columns_to_character() |>
  select(-c(pos, levels, class:n_na)) |>
  qflextable() |> 
  autofit()

Table 28: Data dictionary of b2 in bdbr 1999

variable	label	col_type	missing	value_labels	unique_values	range
b2	year of birth	dbl+lbl	0	[0] year 2000	38	0 - 99

Well, the year 2000 is denoted as 0 with a value label of “2000”. We will correct this while harmonizing by simply denoting 0 as 2000.

v007 - Year of interview

We see that the v007 variable has some value labels only for bdbr 1999 and 2007. Therefore, we check those value labels for anything strange.

# Check the labels for v007 in bdbr 1999
bdbr1_pre_tmp0$bdbr_data$bdbr_1999 |> 
  select(v007) |> 
  look_for(details = "full") |> 
  lookfor_to_long_format() |>
  convert_list_columns_to_character() |>
  select(-c(pos, levels, class:n_na)) |>
  qflextable() |> 
  autofit()

Table 29: Data dictionary of v007 in bdbr 1999

variable	label	col_type	missing	value_labels	unique_values	range
v007	year of interview	dbl+lbl	0	[0] year 2000	2	0 - 99
v007	year of interview	dbl+lbl	0	[99]	2	0 - 99

# # Check the labels for v007 in bdbr 2007
bdbr1_pre_tmp0$bdbr_data$bdbr_2007 |> 
  select(v007) |> 
  look_for(details = "full") |> 
  lookfor_to_long_format() |>
  convert_list_columns_to_character() |>
  select(-c(pos, levels, class:n_na)) |>
  qflextable() |> 
  autofit()

Table 30: Data dictionary of v007 in bdbr 2007

variable	label	col_type	missing	value_labels	unique_values	range
v007	year of interview	dbl+lbl	0	[2007]	1	2007 - 2007

Note that in bdbr 1999 the year 2000 is denoted as 0 with a value label of “2000”. Also, the year 1999 is denoted by 99 has value label 1999. We will simply correct this while harmonizing by denoting 0 as 2000 and 99 as 1999. Similarly, in bdbr 2007, the year 2007 has an empty value label. Here, we will just remove the value label.

Bangladesh HH dataset use for variable creation

Checking the ID variables before harmonization

Here we check the formatting of the constituent variables with which we will prepare the ID variables for the pooled Bangladesh household recode (hr) dataset. We will use the following constituent variables for creating the ID variables for the pooled dataset:

# We check the var type of ID vars in all bdhr datasets.
# First we create a data dictionary of the bdhr datasets in nested tibble.
bdhr1_pre_tmp1 <- bdhr1_pre_tmp0 |>
  mutate(lookfor_idvars = map(
    bdhr_data,
    \(df) {
      df |> 
        select(hv001, hv002) |> 
        lookfor(details = "full") |> 
        select(-c(levels:n_na)) |> 
        # For correctly viewing the range column in data dictionary
        convert_list_columns_to_character()
    }
  ))
bdhr1_pre_tmp1

# Now we unnest the tibble and output the pooled data dictionary 
bdhr1_pre_tmp2 <- bdhr1_pre_tmp1 |> 
  select(c(ctr_name, svy_year, lookfor_idvars)) |> 
  unnest(cols = c(lookfor_idvars)) |> 
  arrange(pos)

# Convert and view the tibble as flextable
bdhr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 31: Data dictionary of variables to be used for ID creation across the bdhr rounds

ctr_name	svy_year	pos	variable	label	col_type	missing	unique_values	range
Bangladesh	1993	1	hv001	cluster number	dbl	0	301	101 - 573
Bangladesh	1996	1	hv001	cluster number	dbl	0	313	101 - 630
Bangladesh	1999	1	hv001	cluster number	dbl	0	341	3 - 500
Bangladesh	2004	1	hv001	cluster number	dbl	0	361	1 - 550
Bangladesh	2007	1	hv001	cluster number	dbl	0	361	1 - 361
Bangladesh	2011	1	hv001	cluster number	dbl	0	600	1 - 600
Bangladesh	2014	1	hv001	cluster number	dbl	0	600	1 - 600
Bangladesh	2017	1	hv001	cluster number	dbl	0	672	1 - 675
Bangladesh	2022	1	hv001	cluster number	dbl	0	674	1 - 675
Bangladesh	1993	2	hv002	household number	dbl	0	544	1 - 615
Bangladesh	1996	2	hv002	household number	dbl	0	549	1 - 658
Bangladesh	1999	2	hv002	household number	dbl	0	500	1 - 547
Bangladesh	2004	2	hv002	household number	dbl	0	246	1 - 290
Bangladesh	2007	2	hv002	household number	dbl	0	198	1 - 252
Bangladesh	2011	2	hv002	household number	dbl	0	184	1 - 217
Bangladesh	2014	2	hv002	household number	dbl	0	204	1 - 222
Bangladesh	2017	2	hv002	household number	dbl	0	255	1 - 299
Bangladesh	2022	2	hv002	household number	dbl	0	206	1 - 225

From the above we can see that both the hv001 and hv002 are of numeric class with no missing values. These variables can be used for preparing the ID variables after finding the maximum length of their largest value. Note that survey year is also a constituent ID variable of 4-digits and we need not check it.

# We thought to process the above nested tibble further by decomposing the 
# "range" col into min and max values using separate_wider_regex().
# However, we hit a roadblock as pattern did not identify the max values in 
# some bdhr rounds correctly
bdhr1_pre_tmp3 <- bdhr1_pre_tmp0 |> 
  # Generate the summary stats for id vars
  mutate(skim_idvars = map(bdhr_data, \(df) {
    df |> 
      select(hv001, hv002) |> 
      skim_without_charts()
  })) |> 
  # Pool the summary stats for all bdhr rounds
  select(c(ctr_name, svy_year, skim_idvars)) |> 
  unnest(cols = c(skim_idvars)) |> 
  arrange(skim_variable, svy_year) |> 
  # Group and generate the max and min values for each variable
  group_by(variable = skim_variable) |> 
  summarize(
    min_val = min(numeric.p0),
    max_val = max(numeric.p100)
  ) |> 
  # calculate the num of digits in the maximum values
  mutate(
    max_digits = nchar(as.character(max_val))
  ) |>
  # add variable labels and relocate it after variable name
  bind_cols(vlabel = c("cluster number", "household number")) |>
  relocate(vlabel, .after = 1)

# Convert the tibble to flextable for easy viewing
bdhr1_pre_tmp3 |>
  qflextable() |>
  align(align = "left", part = "all") |>
  autofit()

Table 32: The maximum length of constituent ID variables to be set across the bdhr rounds

variable	vlabel	min_val	max_val	max_digits
hv001	cluster number	1	675	3
hv002	household number	1	658	3

The above table gives the required length of the constituent ID variables to be set, so that we can correctly concatenate them to create the ID variables. The required length of the ID variables are given in max_digits column. Note that survey year is also a constituent ID variable of 4-digits.

Checking HH-level variables before harmonization

Here we check the wealth quintile variable before harmonizing them. Note in Bangladesh 1993, 1996 and 1999 the wealth quintile variables are provided in separate datasets. Therefore we join those variables to the hh file before proceeding with the checking.

Upon manually checking the full data dictionaries we find the variable names. Now we will check the data dictionary of these hh-level variables. Then we will check their value labels variable wise.

# We check the hh-level vars in all bdhr datasets.
# First we create the data dictionary in nested tibble.
bdhr1_pre_tmp1 <- bdhr1_pre_tmp0 |>
  mutate(lookfor_hhvars = map(
    bdhr_data,
    \(df) {
      df |> 
        # select the common independent variables
        select(
          matches("^wlthind5$|^hv270$")
        ) |> 
        lookfor(details = "full") |> 
        select(-c(levels:n_na)) |> 
        # For correctly viewing the range column in data dictionary
        convert_list_columns_to_character()
    }
  ))
bdhr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary 
bdhr1_pre_tmp2 <- bdhr1_pre_tmp1 |> 
  select(c(svy_year, lookfor_hhvars)) |> 
  unnest(cols = c(lookfor_hhvars)) |> 
  arrange(pos, svy_year)

# Convert the tibble to flextable for easy viewing
bdhr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 33: Data dictionary of hh-level variables across the bdhr rounds

svy_year	pos	variable	label	col_type	missing	unique_values	range
1993	1	wlthind5	quintiles of wealth index	dbl+lbl	0	5	1 - 5
1996	1	wlthind5	quintiles of wealth index	dbl+lbl	0	5	1 - 5
1999	1	wlthind5	quintiles of wealth index	dbl+lbl	0	5	1 - 5
2004	1	hv270	wealth index	dbl+lbl	0	5	1 - 5
2007	1	hv270	wealth index	dbl+lbl	0	5	1 - 5
2011	1	hv270	wealth index	dbl+lbl	0	5	1 - 5
2014	1	hv270	wealth index	dbl+lbl	0	5	1 - 5
2017	1	hv270	wealth index combined	dbl+lbl	0	5	1 - 5
2022	1	hv270	wealth index combined	dbl+lbl	0	5	1 - 5

The above table gives an overall snapshot of the hh-level variables. All the variables are of labelled class and have the same number of value labels across all the bdhr datasets. Next, we compare the value labels of the wealth quintile variable across the bdhr datasets.

Wealth index quintile variable

Next, we check the value labels of the household wealth quintile variable. The variable names of this variable differs across the bdbr datasets. First we create a nested tibble of the value labels.

# Create the data dictionary in nested tibble
bdhr1_pre_tmp1 <- bdhr1_pre_tmp0 |> 
  mutate(lookfor_wiqt = map(
    bdhr_data,
    \(df) {
      df |> 
        select(matches("^wlthind5$|^hv270$")) |> 
        look_for() |> 
        lookfor_to_long_format() |> 
        select(value_labels)
    }
  ))
bdhr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdhr1_pre_tmp2 <- bdhr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(c(ctr_name, svy_year, lookfor_wiqt)) |> 
  unnest(cols = c(lookfor_wiqt)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdhr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "Wealth index quintiles", .before = 2)

# Convert the tibble to flextable for easy viewing
bdhr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 34: Data dictionary of wealth quintiles across the bdhr rounds

ctr_name	var_name	label_num	bdhr_1993	bdhr_1996	bdhr_1999	bdhr_2004	bdhr_2007	bdhr_2011	bdhr_2014	bdhr_2017	bdhr_2022
Bangladesh	Wealth index quintiles	0	[0]	[0]	[0]
Bangladesh	Wealth index quintiles	1	[1] lowest quintile	[1] lowest quintile	[1] lowest quintile	[1] poorest	[1] poorest	[1] poorest	[1] poorest	[1] poorest	[1] poorest
Bangladesh	Wealth index quintiles	2	[2] second quintile	[2] second quintile	[2] second quintile	[2] poorer	[2] poorer	[2] poorer	[2] poorer	[2] poorer	[2] poorer
Bangladesh	Wealth index quintiles	3	[3] middle quintile	[3] middle quintile	[3] middle quintile	[3] middle	[3] middle	[3] middle	[3] middle	[3] middle	[3] middle
Bangladesh	Wealth index quintiles	4	[4] fourth quintile	[4] fourth quintile	[4] fourth quintile	[4] richer	[4] richer	[4] richer	[4] richer	[4] richer	[4] richer
Bangladesh	Wealth index quintiles	5	[5] highest quintile	[5] highest quintile	[5] highest quintile	[5] richest	[5] richest	[5] richest	[5] richest	[5] richest	[5] richest

The value labels are similar across the bdhr rounds with minor differences. Note, the value labels for bdhr 1993, 1996 and 1996 have an empty label code 0. The value label texts are similar for the set of bdhr 1993, 1996 and 1996, then they are same for bdhr 2004, 2007, 2011, 2014, 2017 and 2022 rounds. Therefore, we need to be mindful of this during harmonization.

Bangladesh PR dataset use for family structure variables creation

Checking the ID variables before harmonization

Here we check the formatting of the constituent variables with which we will prepare the ID variables for the pooled Bangladesh person recode (pr) dataset. We will use the following constituent variables for creating the ID variables for the pooled dataset:

# We check the var type of ID vars in all bdpr datasets.
# First we create a data dictionary of the bdpr datasets in nested tibble.
bdpr1_pre_tmp1 <- bdpr1_pre_tmp0 |>
  mutate(lookfor_idvars = map(bdpr_data, \(df) {
    df |> 
      select(hv001, hv002, hvidx) |> 
      lookfor(details = "full") |> 
      select(-c(levels:n_na)) |> 
      # For correctly viewing the range column in data dictionary
      convert_list_columns_to_character()
  }))
bdpr1_pre_tmp1

# Now we unnest the tibble and output the pooled data dictionary 
bdpr1_pre_tmp2 <- bdpr1_pre_tmp1 |> 
  select(c(ctr_name, svy_year, lookfor_idvars)) |> 
  unnest(cols = c(lookfor_idvars)) |> 
  arrange(pos)

# Convert and view the tibble as flextable
bdpr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 35: Data dictionary of variables to be used for ID creation across the bdpr rounds

ctr_name	svy_year	pos	variable	label	col_type	missing	unique_values	range
Bangladesh	1993	1	hv001	cluster number	dbl	0	301	101 - 573
Bangladesh	1996	1	hv001	cluster number	dbl	0	313	101 - 630
Bangladesh	1999	1	hv001	cluster number	dbl	0	341	3 - 500
Bangladesh	2004	1	hv001	cluster number	dbl	0	361	1 - 550
Bangladesh	2007	1	hv001	cluster number	dbl	0	361	1 - 361
Bangladesh	2011	1	hv001	cluster number	dbl	0	600	1 - 600
Bangladesh	2014	1	hv001	cluster number	dbl	0	600	1 - 600
Bangladesh	2017	1	hv001	cluster number	dbl	0	672	1 - 675
Bangladesh	2022	1	hv001	cluster number	dbl	0	674	1 - 675
Bangladesh	1993	2	hv002	household number	dbl	0	544	1 - 615
Bangladesh	1996	2	hv002	household number	dbl	0	549	1 - 658
Bangladesh	1999	2	hv002	household number	dbl	0	500	1 - 547
Bangladesh	2004	2	hv002	household number	dbl	0	246	1 - 290
Bangladesh	2007	2	hv002	household number	dbl	0	198	1 - 252
Bangladesh	2011	2	hv002	household number	dbl	0	184	1 - 217
Bangladesh	2014	2	hv002	household number	dbl	0	204	1 - 222
Bangladesh	2017	2	hv002	household number	dbl	0	255	1 - 299
Bangladesh	2022	2	hv002	household number	dbl	0	206	1 - 225
Bangladesh	1993	3	hvidx	line number	dbl	0	28	1 - 28
Bangladesh	1996	3	hvidx	line number	dbl	0	29	1 - 29
Bangladesh	1999	3	hvidx	line number	dbl	0	26	1 - 26
Bangladesh	2004	3	hvidx	line number	dbl	0	33	1 - 33
Bangladesh	2007	3	hvidx	line number	dbl	0	40	1 - 40
Bangladesh	2011	3	hvidx	line number	dbl	0	31	1 - 31
Bangladesh	2014	3	hvidx	line number	dbl	0	25	1 - 25
Bangladesh	2017	3	hvidx	line number	dbl	0	30	1 - 30
Bangladesh	2022	3	hvidx	line number	dbl	0	25	1 - 25

From the above table we can see that all the three constituent ID variables are of numeric class with no missing values. These variables can directly be used for preparing the ID variables after finding the maximum length of their largest value. Note that survey year is also a constituent ID variable of 4-digits and we need not check it.

# We thought to process the above nested tibble further by decomposing the 
# "range" col into min and max values using separate_wider_regex().
# However, we hit a roadblock as pattern did not identify the max values in 
# some bdpr rounds correctly
bdpr1_pre_tmp3 <- bdpr1_pre_tmp0 |> 
  # Generate the summary stats for id vars
  mutate(skim_idvars = map(bdpr_data, \(df) {
    df |> 
      select(hv001, hv002, hvidx) |> 
      skim_without_charts()
  })) |> 
  # Pool the summary stats for all bdpr rounds
  select(c(ctr_name, svy_year, skim_idvars)) |> 
  unnest(cols = c(skim_idvars)) |> 
  arrange(skim_variable, svy_year) |> 
  # Group and generate the max and min values for each variable
  group_by(variable = skim_variable) |> 
  summarize(
    min_val = min(numeric.p0),
    max_val = max(numeric.p100)
  ) |> 
  # calculate the num of digits in the maximum values
  mutate(
    max_digits = nchar(as.character(max_val))
  ) |>
  # add variable labels and relocate it after variable name
  bind_cols(vlabel = c("cluster number", "household number", "Persons line number")) |>
  relocate(vlabel, .after = 1)

# Convert the tibble to flextable for easy viewing
bdpr1_pre_tmp3 |>
  qflextable() |>
  align(align = "left", part = "all") |>
  autofit()

Table 36: The maximum length of constituent ID variables to be set across the bdpr rounds

variable	vlabel	min_val	max_val	max_digits
hv001	cluster number	1	675	3
hv002	household number	1	658	3
hvidx	Persons line number	1	40	2

The above table gives the required length of the constituent ID variables to be set, so that we can correctly concatenate them to create the ID variables. The required length of the ID variables are given in max_digits column. Note that survey year is also a constituent ID variable of 4-digits.

Checking Family structure variables before harmonization

Here we check the family structure related variables before harmonizing them. The variable names were collected by manually checking the full data dictionaries. Here we will check the data dictionary of these hh-level variables and focus on the variable types.

# We check the family structure vars in all bdpr datasets.
# First we create the data dictionary in nested tibble.
bdpr1_pre_tmp1 <- bdpr1_pre_tmp0 |>
  mutate(lookfor_famstrvars = map(bdpr_data, \(df) {
    df |> 
      # select the common independent variables
      select(c(hv101, hv102, hv103, hv104, hv105)) |> 
      lookfor(details = "full") |> 
      select(-c(levels:n_na)) |> 
      # For correctly viewing the range column in data dictionary
      convert_list_columns_to_character()
  }))
bdpr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary 
bdpr1_pre_tmp2 <- bdpr1_pre_tmp1 |> 
  select(c(svy_year, lookfor_famstrvars)) |> 
  unnest(cols = c(lookfor_famstrvars)) |> 
  arrange(pos, svy_year)

# Convert the tibble to flextable for easy viewing
bdpr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 37: Data dictionary of family structure vars across the bdpr rounds

svy_year	pos	variable	label	col_type	missing	unique_values	range
1993	1	hv101	relationship to head	dbl+lbl	10	12	1 - 12
1996	1	hv101	relationship to head	dbl+lbl	5	12	1 - 12
1999	1	hv101	relationship to head	dbl+lbl	6	13	1 - 98
2004	1	hv101	relationship to head	dbl+lbl	5	12	1 - 12
2007	1	hv101	relationship to head	dbl+lbl	0	12	1 - 99
2011	1	hv101	relationship to head	dbl+lbl	0	11	1 - 12
2014	1	hv101	relationship to head	dbl+lbl	0	11	1 - 12
2017	1	hv101	relationship to head	dbl+lbl	0	11	1 - 12
2022	1	hv101	relationship to head	dbl+lbl	0	12	1 - 98
1993	2	hv102	usual resident	dbl+lbl	0	2	0 - 1
1996	2	hv102	usual resident	dbl+lbl	0	2	0 - 1
1999	2	hv102	usual resident	dbl+lbl	23	3	0 - 1
2004	2	hv102	usual resident	dbl+lbl	3	3	0 - 1
2007	2	hv102	usual resident	dbl+lbl	0	3	0 - 9
2011	2	hv102	usual resident	dbl+lbl	0	2	0 - 1
2014	2	hv102	usual resident	dbl+lbl	0	2	0 - 1
2017	2	hv102	usual resident	dbl+lbl	0	2	0 - 1
2022	2	hv102	usual resident	dbl+lbl	0	2	0 - 1
1993	3	hv103	slept last night	dbl+lbl	0	2	0 - 1
1996	3	hv103	slept last night	dbl+lbl	7	3	0 - 1
1999	3	hv103	slept last night	dbl+lbl	34	3	0 - 1
2004	3	hv103	slept last night	dbl+lbl	3	3	0 - 1
2007	3	hv103	slept last night	dbl+lbl	0	3	0 - 9
2011	3	hv103	slept last night	dbl+lbl	0	2	0 - 1
2014	3	hv103	slept last night	dbl+lbl	0	2	0 - 1
2017	3	hv103	slept last night	dbl+lbl	0	2	0 - 1
2022	3	hv103	stayed last night	dbl+lbl	0	2	0 - 1
1993	4	hv104	sex of household member	dbl+lbl	0	2	1 - 2
1996	4	hv104	sex of household member	dbl+lbl	0	2	1 - 2
1999	4	hv104	sex of household member	dbl+lbl	4	3	1 - 2
2004	4	hv104	sex of household member	dbl+lbl	0	2	1 - 2
2007	4	hv104	sex of household member	dbl+lbl	0	2	1 - 2
2011	4	hv104	sex of household member	dbl+lbl	0	2	1 - 2
2014	4	hv104	sex of household member	dbl+lbl	0	2	1 - 2
2017	4	hv104	sex of household member	dbl+lbl	0	2	1 - 2
2022	4	hv104	sex of household member	dbl+lbl	0	2	1 - 2
1993	5	hv105	age of household members	dbl+lbl	3	98	0 - 98
1996	5	hv105	age of household members	dbl+lbl	16	97	0 - 98
1999	5	hv105	age of household members	dbl+lbl	29	96	0 - 98
2004	5	hv105	age of household members	dbl+lbl	8	99	0 - 98
2007	5	hv105	age of household members	dbl+lbl	0	98	0 - 99
2011	5	hv105	age of household members	dbl+lbl	4	98	0 - 96
2014	5	hv105	age of household members	dbl+lbl	6	98	0 - 98
2017	5	hv105	age of household members	dbl+lbl	0	94	0 - 95
2022	5	hv105	age of household members	dbl+lbl	0	96	0 - 95

The above table gives an overall snapshot of the family structure related variables. Interestingly, all the variables including age of hh members (a continuous var) are of labelled class. The relation to head and de facto resident variables have few missing values in bdpr 1996. Note that, the three variables of interest hv101-hv102, two variables hv101 and hv103 have different number of value labels across the bdpr rounds. Next, we compare the value labels of the individual variables across the bdpr datasets.

hv101 - Relationship to head

Next, we check the value labels of the relationship to the household head variable. First we create a nested tibble of the value labels.

# Create the data dictionary in nested tibble
bdpr1_pre_tmp1 <- bdpr1_pre_tmp0 |> 
  mutate(lookfor_hv101 = map(bdpr_data, \(df) {
    df |> 
      select(hv101) |> 
      look_for() |> 
      lookfor_to_long_format() |> 
      select(value_labels)
  }))
bdpr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdpr1_pre_tmp2 <- bdpr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(c(ctr_name, svy_year, lookfor_hv101)) |> 
  unnest(cols = c(lookfor_hv101)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdpr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "hv101", .before = 2)

# Convert the tibble to flextable for easy viewing
bdpr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 38: Data dictionary of relationship to head variable across the bdpr rounds

ctr_name	var_name	label_num	bdpr_1993	bdpr_1996	bdpr_1999	bdpr_2004	bdpr_2007	bdpr_2011	bdpr_2014	bdpr_2017	bdpr_2022
Bangladesh	hv101	1	[1] head	[1] head	[1] head	[1] head	[1] head	[1] head	[1] head	[1] head	[1] head
Bangladesh	hv101	2	[2] wife or husband	[2] wife or husband	[2] wife or husband	[2] wife or husband	[2] wife or husband	[2] wife or husband	[2] wife or husband	[2] wife or husband	[2] wife or husband
Bangladesh	hv101	3	[3] son/daughter	[3] son/daughter	[3] son/daughter	[3] son/daughter	[3] son/daughter	[3] son/daughter	[3] son/daughter	[3] son/daughter	[3] son/daughter
Bangladesh	hv101	4	[4] son/daughter-in-law	[4] son/daughter-in-law	[4] son/daughter-in-law	[4] son/daughter-in-law	[4] son/daughter-in-law	[4] son/daughter-in-law	[4] son/daughter-in-law	[4] son/daughter-in-law	[4] son/daughter-in-law
Bangladesh	hv101	5	[5] grandchild	[5] grandchild	[5] grandchild	[5] grandchild	[5] grandchild	[5] grandchild	[5] grandchild	[5] grandchild	[5] grandchild
Bangladesh	hv101	6	[6] parent	[6] parent	[6] parent	[6] parent	[6] parent	[6] parent	[6] parent	[6] parent	[6] parent
Bangladesh	hv101	7	[7] parent-in-law	[7] parent-in-law	[7] parent-in-law	[7] parent-in-law	[7] parent-in-law	[7] parent-in-law	[7] parent-in-law	[7] parent-in-law	[7] parent-in-law
Bangladesh	hv101	8	[8] brother/sister	[8] brother/sister	[8] brother/sister	[8] brother/sister	[8] brother/sister	[8] brother/sister	[8] brother/sister	[8] brother/sister	[8] brother/sister
Bangladesh	hv101	9	[9] co-spouse	[9] co-spouse	[9] co-spouse	[9] co-spouse	[9] co-spouse	[9] co-spouse	[9] co-spouse	[9] co-spouse	[9] co-spouse
Bangladesh	hv101	10	[10] other relative	[10] other relative	[10] other relative	[10] other relative	[10] other relative	[10] other relative	[10] other relative	[10] other relative	[10] other relative
Bangladesh	hv101	11	[11] adopted/foster child	[11] adopted/foster child	[11] adopted/foster child	[11] adopted/foster child	[11] adopted/foster child	[11] adopted/foster child	[11] adopted/foster child	[11] adopted/foster child	[11] adopted/foster child
Bangladesh	hv101	12	[12] not related	[12] not related	[12] not related	[12] not related	[12] not related	[12] not related	[12] not related	[12] not related	[12] not related
Bangladesh	hv101	13					[13] niece/nephew by blood	[13] niece/nephew by blood	[13] niece/nephew by blood	[13] niece/nephew by blood	[13] niece/nephew by blood
Bangladesh	hv101	14					[14] niece/nephew by marriage	[14] niece/nephew by marriage	[14] niece/nephew by marriage	[14] niece/nephew by marriage	[14] niece/nephew by marriage
Bangladesh	hv101	98	[98] dk	[98] dk	[98] dk	[98] dk	[98] dk	[98] don't know	[98] don't know	[98] don't know	[98] don't know

The above table shows that the value label texts vary across the bdpr rounds. To harmonize the relationship to head variable we will use the following value labels -

• 1 head
• 2 spouse
• 3 child
• 4 child-in-law
• 5 grandchild
• 6 parent
• 7 parent-in-law
• 8 sibling
• 9 others

Here, we merge the “spouse” and “co-spouse” categories into “spouse” category, and the “son/daughter” and “adopted/foster child” categories into “child” category.

hv102 - de jure/usual resident

Next, we check the value labels of the de jure resident variable. This means if a household member is an usual resident of the household. First we create a nested tibble of the value labels.

# Create the data dictionary in nested tibble
bdpr1_pre_tmp1 <- bdpr1_pre_tmp0 |> 
  mutate(lookfor_hv102 = map(bdpr_data, \(df) {
    df |> 
      select(hv102) |> 
      look_for() |> 
      lookfor_to_long_format() |> 
      select(value_labels)
  }))
bdpr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdpr1_pre_tmp2 <- bdpr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(c(ctr_name, svy_year, lookfor_hv102)) |> 
  unnest(cols = c(lookfor_hv102)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdpr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "hv102", .before = 2)

# Convert the tibble to flextable for easy viewing
bdpr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 39: Data dictionary of the De jure resident variable across the bdpr rounds

ctr_name	var_name	label_num	bdpr_1993	bdpr_1996	bdpr_1999	bdpr_2004	bdpr_2007	bdpr_2011	bdpr_2014	bdpr_2017	bdpr_2022
Bangladesh	hv102	0	[0] no	[0] no	[0] no	[0] no	[0] no	[0] no	[0] no	[0] no	[0] no
Bangladesh	hv102	1	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes

The above table shows that hv102 has the same value label texts and codes across the bdpr rounds. Therefore, we can use this variable directly after converting to factor type.

hv103 - de facto resident

Next, we check the value labels of the de facto resident variable. In DHS this means if a household member slept last night in the household. First we create a nested tibble of the value labels.

# Create the data dictionary in nested tibble
bdpr1_pre_tmp1 <- bdpr1_pre_tmp0 |> 
  mutate(lookfor_hv103 = map(bdpr_data, \(df) {
    df |> 
      select(hv103) |> 
      look_for() |> 
      lookfor_to_long_format() |> 
      select(value_labels)
  }))
bdpr1_pre_tmp1

# Now we unnest the tibble and refine the pooled data dictionary
bdpr1_pre_tmp2 <- bdpr1_pre_tmp1 |> 
  # First we select the required cols and unnest()
  select(c(ctr_name, svy_year, lookfor_hv103)) |> 
  unnest(cols = c(lookfor_hv103)) |> 
  # Next we make the num of value labels same across each round
  mutate(label_num = parse_number(value_labels)) |> 
  complete(ctr_name, svy_year, label_num) |>
  # Next we create col of value labels for each survey round
  pivot_wider(
    names_from = svy_year, 
    values_from = value_labels,
    names_prefix = "bdpr_"
  ) |>
  # Show the variable name in a col
  mutate(var_name = "hv103", .before = 2)

# Convert the tibble to flextable for easy viewing
bdpr1_pre_tmp2 |> 
  qflextable() |> 
  align(align = "left", part = "all") |> 
  autofit()

Table 40: Data dictionary of the De facto resident variable across the bdpr rounds

ctr_name	var_name	label_num	bdpr_1993	bdpr_1996	bdpr_1999	bdpr_2004	bdpr_2007	bdpr_2011	bdpr_2014	bdpr_2017	bdpr_2022
Bangladesh	hv103	0	[0] no	[0] no	[0] no	[0] no	[0] no	[0] no	[0] no	[0] no	[0] no
Bangladesh	hv103	1	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes	[1] yes

The above table shows that hv103 has the same value label texts and codes across the bdpr rounds. Therefore, we can use this variable directly after converting to factor type.

START FROM HERE

TASK:

• Handling multiple births in death scarring vars may not be necessary.
• Preceding birth interval construction has changed with DHS-7. We could re-construct it.

TO BE CONTINUED …