============================================================================================================================== = This is the README file for replicating Intergenerational Support During the Rise of Mobile Telecommunication in Indonesia = ============================================================================================================================== I. ======================= ReadMe ===================================================== ***** The .do files were written and run with STATA/SE 15.1 *********** ***** Constructing the sample and Running the Analysis **************** 1. Do-file: "data cleaning" This is the STATA program file for generating the sample used in the paper “Intergenerational Support During the Rise of Mobile Telecommunication in Indonesia”. 2. Do-file: “estimation_period” This is the STATA program file for running the analysis using the sample as repeated cross-sectional. For details, please see section 3. Data in the manuscript. 3. Do-file: “descriptive_period” This is the STATA program file for generating the descriptive table (Table 2) using the sample as repeated cross-sectional. 4. Do-file: “estimation_cohort” This is the STATA program for the analysis using the sample as longitudinal data. For details, please see section 3. Data in the manuscript. 5. Do-file: “fakeoutcome_height” This is the STATA program for running the placebo test (Appendix table 8). Running the placebo test requires to construct a new sample to include an additional outcome variable (height) used in the analysis. II. ======================== MetaInformation ======================================== ****Accessing the data: the Indonesia Family Life Survey *** Interested users can register with RAND, which manages the data. You can register by visiting the website: (https://www.rand.org/well-being/social-and-behavioral-policy/data/FLS/IFLS.html) and directions for registering and accessing the data are provided under the section "Register and Download Data". The contact information for the researchers who manage the data is available on the website under the section "Contact Us". To directly use the .do files provided in this replication package, the data files should be stored in folders with paths as: IFLSx/cf/ for community survey and IFLSx/hh/ for household survey, where the “x” in IFLSx indicates the wave of the survey. ****Accessing the data: The Village Potential Statistics (PODES) **** PODES data are managed by Statistics Indonesia (BPS). Interested users can apply for the data by visiting the website: (https://www.bps.go.id/). Directions for applying and purchasing the data can be found under the “Service” – “Purchasing Service” section. **** Variables and Records Used in the Analysis ******************** 1. Test the effect of signal coverage on telecommunication consumption and cellphone ownership: "telcons_d"(0 = no telecommunication consumption; 1 = have telecommunication consumption), coded from ks06 in household survey book 1; "sig_q" measures signal coverage, (0 = incomplete signal coverage in the community within the subdistrict; 1 = complete signal coverage in the community within the subdistrict (>97%)), coded from R910 in PODES data; "cell" measures cellphone ownership, (0 = no; 1 = yes), coded from DLA03b in household survey Book 3B Number of records used: 2,770 records in IFLS 4 and 5 to predict telecommunication consumption, and 1,507 records in IFLS5 to predict cellphone ownership; 2. Test the impact of signal coverage on weekly contact frequency: "contact2" measures weekly contact frequency, (0 = less than weekly; 1 = at least weekly ), coded from ba84b in household survey Book 4 Number of records used: 2,756 records in IFLS 4 and 5 to predict contact frequency. 3. Test the impact of signal coverage on transfer from older parents/adult children to adult children/older parents: "lntr_down2" and "lntr_up2" measures downward and upward transfers, respectively; they are both continuous variables; coded from ba88 and ba88a in household survey Book4; Number of records used: 4,810 records in IFLS 2,3,4,5 are used to predict transfer amount. 4. Set of individual and family characteristics: "edu" measures the highest level of education of a respondent (0 < junior high; 1 >= junior high), coded from ar16 in household survey book K; "marst" measures the marital status of a respondent (0 = not in marital union; 1 = in marital union), coded from ar13 in household survey Book K; "cores_bk" measures whether an older adult coresides with adult children (0 = no, 1 = yes), coded from ar02 in data file "ptrack"; "far" measures where the farthest child of an older adult is (0 = within district; 1 = outside of district), coded from Book 4; "pcepc_q" measures the fertile of the monthly consumption of a household (1 = the first tertile, 2 = the second tertile, 3 = the third tertile), coded from data file "pcexx" where "xx" is year; "nrchedu_h" and "nrch_youngest" measures the educational level and the age of the child with the highest education among the siblings, and the age of the yougest child, coded from ar13 and age_xx from Book K. 5. Set of community charateristics: "comm_trans" is a transportation index of a community (0 -7), the value is calculated using questions in Section A of the Community Survey Book 1; "comm_schl" measures the number of school in the community (continuous variable), coded from i13 in community survey Book 1; "garbage" measures whether the garbage in the community is taken away by a designated person (0 = no; 1 = yes),coded from c19a in community survey Book 1; "piped" measures whether the main way of water is piped water in the community (0 = no, 1 = yes); coded from C1_1a in community survey book 1; "healthpost" measures the number of health post in the community; coded from j11 in community survey Book 1; "comm_ind/farm/service" measures whether the main industry in the community is industrial, farming, or service, coded from s40 in community survey Book 2.