prepare a MICS dataset for analysis

Usage

prep_mics_sib_histories(
  mm.df,
  survey,
  wm.df = NULL,
  varmap = sibhist_varmap_mics6,
  id.vars = c("hh1", "hh2", "ln"),
  doi.var = "wdoi",
  doi.ym = c("wm6y", "wm6m"),
  dob.var = "wdob",
  lowercase = TRUE,
  weight.scale = 1,
  add_maternal = FALSE,
  style = c("mics6", "mics4"),
  na.action = NULL,
  preg.window = c("2months", "42days"),
  keep_missing = FALSE,
  keep_varmap_only = FALSE,
  death.exposure = c("dhs", "mics"),
  verbose = TRUE
)

Arguments

mm.df: the MICS maternal mortality file (mm.sav), one row per reported sibling
survey: string identifying the survey, eg "ZW2019". Required: MICS has no v000 equivalent, so this cannot be derived; see Details
wm.df: optional women's file (wm.sav), joined on id.vars to bring across respondent attributes that mm.df does not carry
varmap: see Details; defaults to sibhist_varmap_mics6
id.vars: columns of mm.df that together identify a respondent. Combined to form caseid
doi.var: column holding the date of interview as a CMC. Used when present; otherwise doi.ym is used
doi.ym: year and month columns from which to construct a CMC date of interview, when doi.var is absent
dob.var: column holding the respondent's date of birth as a CMC, used to derive respondent age
lowercase: lowercase all variable names before matching the varmap? See Details
weight.scale: divide the weight by this number. Defaults to 1, since MICS weights are already normalized; see Details
add_maternal: should maternal/pregnancy-related death info be added?
style: which MICS roster coding the data use: "mics6" (the default, also correct for MICS7) or "mics4" (also correct for MICS5). Only used when add_maternal = TRUE
na.action: required when add_maternal = TRUE; see add_maternal_deaths
preg.window: width of the postpartum window for the pregnancy-related column; "42days" reproduces published MICS tables. See add_maternal_deaths
keep_missing: should we keep reported sibs that are missing sex or survival status?
keep_varmap_only: should we only keep ego variables mentioned in the varmap?
death.exposure: whether a sibling who died contributes the month of death as exposure; "dhs" (default) counts it, "mics" stops the month before. See get_sib_df
verbose: report detailed summaries?

Value

a list; see Details

Details

MICS publishes the sibling history as a dedicated file with one row per reported sibling, unlike the DHS, which publishes a wide women's file with one column per sibling attribute per sibling. So no reshape is needed, and this function does not call attributes.to.long.

survey has no default because MICS has no v000 equivalent. Supplying it explicitly keeps MICS survey ids comparable with the DHS codes.

weight.scale defaults to 1 because MICS weights (wmweight) are already normalized to average 1. Do not pass 1e6 here: that is a DHS convention.

MICS .sav files use mixed case – questionnaire items and link keys are uppercase (MM15, HH1, WDOI) while derived and design variables are lowercase (wmweight, psu, welevel) – so lowercase = TRUE normalizes them and every shipped MICS varmap is written in lowercase.

Several columns the rest of the pipeline requires are not in every MICS file and are constructed here when absent:

caseid – from id.vars; MICS has no single respondent id
doi – from doi.var, else doi.ym. Must end up a CMC
age – respondent age, from (doi - dob.var) / 12
psu – from the first of id.vars (the cluster) when no psu column exists, which is the case in most MICS surveys
sex – constant 'f'; MICS interviews only women

sib.dob and sib.death.date are used when the varmap supplies them (MM17C/MM18C in MICS6, MM7C/MM8C in MICS4/5) and approximated from reported ages and years-since-death otherwise.

varmap has the same format as the DHS varmaps: columns orig.varname, new.varname and sibvar.

Returns a list whose entries include

ego.dat - dataset with information about the survey respondents
sib.dat - dataset with information about the reported siblings
summ - a one-row tibble with a summary of the data

Examples

  # TODO - write example code