Meliyanni Johar and Shiko Maruyama, "Does Coresidence Improve An
Elderly Parent's Health?", Journal of Applied Econometrics, Vol. 29,
No. 6, 2014, pp. 965-983.

The data used in this paper are derived from the Indonesian Family
Life Survey (IFLS) managed by the RAND Corporation. While the data are
free, the IFLS team does not allow distribution of data so any user
must register at http://www.rand.org/labor/FLS/IFLS.html 

The data used in this study are from IFLS2, IFLS3 and IFLS4. The IFLS2
is only used for the community adat variables. In each wave, data for
each booklet (e.g., education) is stored separately so data files have
to be merged to use variables from various booklets. 

IFLS2. 
The data are extracted from c97adat.zip: adat_bl
The data is at the community level and can be merged to household data
  in IFLS3 through community identifier 'commid97'.
The tracking file is hh00_all_dta: htrack.dta
This file contains both household identifier in IFLS3 'hhid00' and
  community identifier in IFLS2 'commid97'.  

IFLS3.
The household data are extracted from hh00_all_dta: b2_hr2.dta
  (house ownership and wealth) 
The household information can be merged to individual data through
  household identifier 'hhid00'

The individual data are extracted from hh00_all_dta: bk_ar1.dta (age,
sex, presence of spouse, spouse's age, education, religion, ethnicity,
coresidence); b3b_kk.dta and b3b_ma1.dta (subjective health, health
conditions); b3a_tk1.dta (work); b3b_pm1.dta (community
participation); b3a_hi.dta (pension); b3b_ba2.dta and bk_ar1.dta
(number of siblings; the first data provides the number of
non-coresiding siblings and the second data which is the household
roster provides the number of coresiding siblings); b3b_ba6.dta,
b4_ba6.dta and bk_ar1.dta (age of youngest child and number of
children; the first data provides the information on non-coresiding
children for men and women age>49, the second data provides
information on non-coresiding children for women age<=49, and the last
data provides information on coresiding children); b3a_kw2.dta (who
chooses spouse).

The location variable is extracted from hh00_all_dta: htrack.dta

The dummy for Java is based on community in IFLS3 ('commid00'): all
codes beginning with '3'.

The remoteness variable (rural) is extracted from cf00_all.dta: bk1.dta

IFLS4. 
The mortality information is extracted from hh07_all_dta: bk_ar1.dta 
The health information is extracted from hh07_all_dta: b3b_kk.dta 

The population of interest is chosen based on IFLS3 with selection
criteria described in Section 4. 

Meliyanni Johar
Meliyanni.Johar [AT] uts.edu.au