Update 2021, panel

Author: gronert-m

GLD/IND/IND_2017-2023_PLFS-Urban-Panel/IND_2017-2023_PLFS-Urban-Panel_V02_M/Programs/IND_2017.do

@@ -0,0 +1,68 @@
+
+***************************************
+** Code for IND PLFS Urban Panel     **
+** Master file that defines paths,   **
+** calls other functions to do work  **
+***************************************
+
+* List files in folder
+filelist, dir("${path_input}") pat("*.dta")
+
+* Create var that takes out survey core (i.e., without version numbers)
+gen survey_core = regexs(1) if regexm(filename, "(^[a-zA-Z][a-zA-Z][a-zA-Z]_[0-9][0-9][0-9][0-9]_[a-zA-Z0-9-]+)(_[vV][0-9][0-9]_M_[vV][0-9][0-9]_A)_([a-zA-Z][a-zA-Z][a-zA-Z]+)_([ALL]+)\.dta")
+
+* By construction of the name (surveycore_v##_M_v##_A), the alphanumeric order will
+* put the most recent one last. Thus, surveycore_v01_M_v03_A is after surveycore_v01_M_v02_A,
+* both are before surveycore_v02_M_v01_A.
+
+* We need to make sure the filename is entirely capitalized; otherwse, not sort properly
+replace filename = upper(filename)
+ 
+sort filename
+egen survey_number = seq(), by(survey_core)
+egen survey_numb_max = max(survey_number), by(survey_core)
+keep if survey_numb_max == survey_number
+		
+* Create relevant variables to match
+gen country = substr(survey_core,1,3)
+gen years = substr(survey_core,5,4)
+gen survname = regexs(1) if regexm(survey_core, "^[a-zA-Z][a-zA-Z][a-zA-Z]_[0-9][0-9][0-9][0-9]_([a-zA-Z0-9-]+)")
+
+* Keep relevant variables, save
+keep country years survname dirname filename
+
+* Create variable for full name
+gen fullname = dirname + "/" + filename
+
+levelsof fullname, local(year_files)
+clear
+foreach file of local year_files {
+	
+    append using "`file'", force
+	
+}
+
+* In the case of India, panel can only be formed using the urban sample
+keep if urban == 1
+
+* Diagnostic: Check wave and visit consistency. For a given HH wave- visit_no should be unique. Only applies when visit_no is available. 
+	
+capture confirm variable visit_no
+	
+if _rc == 0 {
+	
+	gldpanel_wave_visit_check
+	drop vw_tag
+	
+}
+		
+* Here there are 329,076 cases of individual-waves where the household-wave info is mapped to more than one visit_no. There are two possibilities: (1) visit is assigned based on individual appearance not the household's; (2) the hhid is  not unique for a given round-year. 
+
+* Diagnostic: Check for re-use of HHID outside panel
+gldpanel_id_check
+	
+* The results above show that there is a prevalent re-use of HHIDs in non-subsequent years!
+* Also, there is a re-use of HHID within a year! 
+* This means we cannot use HHID or PID to create the panel variable!
+
+*save "${path_output}/paneldata_${country}.dta", replace

GLD/IND/IND_2017-2023_PLFS-Urban-Panel/IND_2017-2023_PLFS-Urban-Panel_V02_M/Programs/IND_2018.do

@@ -0,0 +1,98 @@
+
+***************************************
+** Code for IND PLFS Urban Panel     **
+** Create the panel vars from the    **
+** appended file                     **
+***************************************
+
+*<_panel_>
+
+	capture confirm variable panel
+	if _rc==0 {
+		display "panel is already available in the data. Do not create"
+	}
+	else {
+		
+		* Create panel variable out of visit numbers
+		* Credits to Nils Enevoldsen for this code:
+		destring wave, ignore("Q") gen(wave_num)
+
+		gen quarter_monotonic = wave_num + 8 * floor((year - 2017) / 2)
+
+		// Generate Panel ID, numbered as in PLFS documentation
+		// This logic might break with future PLFS releases. In particular, it assumes that
+		// the "tens place" increments every two rounds, and the "ones place" resets every
+		// two rounds.
+		tempvar tens ones
+		gen `tens' = 1 + floor((quarter_monotonic - visit_no) / 8)
+		gen `ones' = 1 + mod(quarter_monotonic - visit_no, 8)
+		gen panel_num = `tens' * 10 + `ones'
+		*tostring panel_num, gen(panel)
+		
+		* Slight deviation from Nils, instead adopt panel nomenclature used in the report (which starts with P)
+		gen panel = "P" +string(panel_num)
+
+		* Display unique panels and tabulate year and wave for each panel
+		quietly levelsof panel, local(panels)
+		foreach panel of local panels {
+			display "* Displaying panel: `panel'",
+			dis "This is panel: `panel'"
+			tab year wave if panel == "`panel'"
+		}
+	}
+
+	
+	*drop wave_num panel_num
+*</_panel_>
+
+*<_visit_no_>
+
+	capture confirm variable visit_no
+	if _rc==0 {
+		display "visit_no is already available in the data. Do not create"
+	}
+		
+*</_visit_no_>
+
+
+*<_hhid_panel_>
+	gen hhid_panel = hhid + panel
+	label var hhid_panel "Household ID (panel)"
+*</_hhid_panel_>
+
+
+*<_pid_panel_>
+* Create PID variable adding panel information
+	* First extract the individual number
+	gen rosternum = substr(pid, -2, 2)
+	gen pid_panel = hhid_panel + rosternum
+	drop rosternum
+	
+	isid pid_panel wave
+	label var pid_panel "Person ID (panel)"
+*</_pid_panel_>
+
+* Clean up
+
+quietly{
+
+drop quarter_monotonic
+order panel, before(visit_no)
+order hhid_panel, after(hhid)
+order pid_panel, after(pid)
+
+}
+
+* Diagnostic: Check for re-use of HHID outside panel
+gldpanel_id_check, hhid(hhid_panel) pid(pid_panel)
+	
+* Check panel formation - if needed
+/*
+levelsof panel, local(panels)
+foreach panel of local panels {
+	
+	dis "This is panel: `panel'"
+	tab year wave if panel == "`panel'"
+	
+}
+*/

GLD/IND/IND_2017-2023_PLFS-Urban-Panel/IND_2017-2023_PLFS-Urban-Panel_V02_M/Programs/IND_2019.do

@@ -0,0 +1,33 @@
+
+***************************************
+** Code for IND PLFS Urban Panel     **
+** Assess Panel Quality              **
+***************************************
+
+
+*---------- Age sex matches ------------------------------*
+*gldpanel_issue_check, hhid(hhid_panel) pid(pid_panel)
+*graph export "${path_work}/age_sex_matches.png", replace
+
+*---------- Sources of mismatch ------------------------------*
+gldpanel_check_source, hhid(hhid_panel) pid(pid_panel)
+graph export "${path_work}/source_mismatches.png", replace
+
+*---------- PID Attrition ------------------------------*
+* Case of India, change wave so that it runs in sets of four
+gen tempwave = wave
+replace tempwave = "Q1" if wave == "Q5"
+replace tempwave = "Q2" if wave == "Q6"
+replace tempwave = "Q3" if wave == "Q7"
+replace tempwave = "Q4" if wave == "Q8"
+
+gldpanel_attrition, hhid(hhid_panel) pid(pid_panel) wave(tempwave) consecutive_waves
+graph export "${path_work}/attrition_consecutive_waves.png", replace name(Graph)
+
+gldpanel_attrition, hhid(hhid_panel) pid(pid_panel) wave(tempwave) any_wave
+graph export "${path_work}/attrition_any_wave.png", replace
+
+gldpanel_attrition, hhid(hhid_panel) pid(pid_panel) wave(tempwave) all_waves
+graph export "${path_work}/attrition_all_waves.png", replace
+
+cap drop tempwave