|
Notes
| Here, we briefly explain which steps are necessary to reproduce all the results found in the paper “Long-Term Care Across Europe and the U.S.: The Role of Formal and Informal Care”. Throughout the Stata do-files there are numerous and often detailed comments that should allow the reader to follow through each step. It is assumed that all the data files that we made available from the HRS and SHARE are downloaded and placed in a folder. In our case this folder is called DATASET. The code has to be modified in the obvious way if the user chooses a different folder structure. The following instructions should be followed in a step-wise procedure as many of the data sets build on each other. See the (admittedly somewhat chaotic) graph in _OverviewTablesStata.pdf for how the different tables build on each other. ******************************* CREATING THE SAMPLE ******************************* DATASET: folder which contains all original data, samples created, and sub-folders (used subsequently) for samples and results * POOLING * NURHOME * OOP * HEAVY * RESULTS DOFILES: folder which contains do-files 1. LTC_getSample_HRS.do creates HRS data set full_sample_HRS_new.dta (saved in DATASET) 2. LTC_fullSample_HRS.do makes minor modifications to full_sample_HRS_new.dta and saves data set as full_sample_HRS_new_temp.dta (saved in DATASET) 3. PREPARATION_1.do creates data sets PRE1-PRE6 based on SHARE (saved in POOLING). 4. PREPARATION_2.do creates data sets PRE7-PRE11 based on SHARE (saved in POOLING). 5. INSIDE_HH.do creates data sets INSIDEHH, INSIDEHH_SP, INSIDEHH_RS based on SHARE and deals with some inconsistencies within households (saved in POOLING). 6. HOURS_IMPUTATION_HRS.do creates data set HRS_input.dta which is harmonized in terms of variable naming etc. with our SHARE sample (saved in POOLING). 7. LTC_getSample_SHARE.do creates the following data tables on care in SHARE: -> P1.dta-P5.dta from SHARE data (saved in POOLING). -> SHARE.dta from merging all SHARE care modules (saved in POOLING). 8. Pool_Impute_Assign.do pools SHARE.dta and HRS_input.dta, imputes total hours of care according to the three methods Dahl, Grouped OLS, or Pooled OLS described in the paper, and assigns total care hours to care categories and to helpers. -> Pooling step: saved as SHARE_HRS_1.dta -> Imputation step: saved as SHARE_HRS_2_1.dta -> Assignment step: saves as SHARE_HRS_2_2.dta -> Combines data sets and saves as FINAL_* where * depends on the imputation method. (all are saved in POOLING) 9. NURSINGHOME_do creates nursing home sample based on SHARE in steps NH1-NH6 (saved in NURHOME); merges these modules and saves it as SHARE.dta (same name as above! but, saved in different folder, NURHOME). 10. OOP.do creates out-of-pocket expenditures OOP1-OOP6 in SHARE (saved in OOP). ************ RESULTS ************ 1. RESULTS.do first does some more preliminary work using nursing-home data from SHARE and nursing-home data from the OECD -> Get table on representativeness of NH population across countries Uses FINAL_* version created before to merge with additional data and creates . additional useful variables; sort out countries and saves data set as FINAL_*_2 This is now the main source to create almost all other tables and figures from the paper. 2. HeavyHelper.do |
