* stata_codebook.do - attach long-form notes to the .dta files (run once in Stata).
* Generated by build_data_dictionary.py - do not edit by hand.

* ---- raw_data.dta ----
use "raw_data.dta", clear
label data "Full Linden-Rockoff transaction file (NC home sales)"
note _dta: Full home-transaction file from Butts's (2023) replication archive of Linden & Rockoff (2008); one row per sale, 1994-2004. distance is zero for sales outside every offender's catchment. This is a CSV dump of the original Stata source (linden_rockoff.dta); the .dta column labels are reused here.
note sale_month: Calendar month in which the home transaction closed.. Construction: From the transaction record.. Units: 1-12. Source: Linden & Rockoff (2008)
note sale_year: Calendar year in which the home transaction closed.. Construction: From the transaction record.. Units: year. Source: Linden & Rockoff (2008)
note cpi_value: Consumer Price Index value used to deflate sale prices.. Construction: From the transaction record (price-deflation series).. Units: index. Source: Linden & Rockoff (2008)
note correction_term_dec_2004: Inflation-correction factor expressing prices in December-2004 dollars.. Construction: Derived from the CPI series, base = December 2004.. Units: factor. Source: Linden & Rockoff (2008)
note geo_code: Census geography identifier: state, county, tract, block group.. Construction: Geocoded census block-group code.. Units: code. Source: Linden & Rockoff (2008)
note parcel: Unique tax-parcel identifier for the property.. Construction: From the county property records.. Units: string. Source: Linden & Rockoff (2008)
note amt_Price: Recorded sale price of the home.. Construction: From the transaction record.. Units: US$. Source: Linden & Rockoff (2008)
note sale_date: Date the home transaction closed.. Construction: From the transaction record.. Units: date. Source: Linden & Rockoff (2008)
note year_built: Year the property was constructed.. Construction: From the county property records.. Units: year. Source: Linden & Rockoff (2008)
note HEATED: Heated floor area of the home.. Construction: From the county property records.. Units: sq ft. Source: Linden & Rockoff (2008)
note BATHS: Count of bathrooms in the home.. Construction: From the county property records.. Units: count. Source: Linden & Rockoff (2008)
note BEDROOMS: Count of bedrooms in the home.. Construction: From the county property records.. Units: count. Source: Linden & Rockoff (2008)
note neighborhood: Assessor neighborhood identifier (the clustering unit for SEs).. Construction: From the county property records.. Units: code. Source: Linden & Rockoff (2008)
note distance: Distance from the home to the nearest offender's eventual address; zero if the home is in no offender's area. In data_prepared this is rescaled to miles (raw distance / 3).. Construction: Geocoded straight-line distance; the analysis sample rescales to miles via distance / 3.. Units: miles (prepared) / raw units (raw_data). Source: Linden & Rockoff (2008)
note srn: Identifier of the registered offender whose catchment the home falls in.. Construction: Sex-offender registry record number.. Units: string. Source: Linden & Rockoff (2008)
note offender_address_date: Date the offender's address was published / the offender moved in.. Construction: From the sex-offender registry.. Units: date. Source: Linden & Rockoff (2008)
note building_grade: Assessor construction-quality grade of the home.. Construction: From the county property records.. Units: category. Source: Linden & Rockoff (2008)
note offender: 1 if the home is within 1/3 mile of some offender's eventual address (i.e. enters the analysis sample); else 0.. Construction: Flag derived from distance to nearest offender.. Units: 0/1. Source: Linden & Rockoff (2008)
note streetname: Street on which the property is located.. Construction: From the county property records.. Units: string. Source: Linden & Rockoff (2008)
note BQM1: First component of the assessor building-quality grade (1-6).. Construction: From the county property records.. Units: 1-6. Source: Linden & Rockoff (2008)
note BQM2: Second component of the assessor building-quality grade (1-6).. Construction: From the county property records.. Units: 1-6. Source: Linden & Rockoff (2008)
note AIRCOND: 1 if the home has air conditioning, else 0.. Construction: From the county property records.. Units: 0/1. Source: Linden & Rockoff (2008)
note HEIGHT: Coded number of stories / height class (1, 1.5, 2, 3, 8, 9).. Construction: From the county property records.. Units: code. Source: Linden & Rockoff (2008)
note WALL: Exterior wall-material category (e.g. ALUM,VINYL; FACE BRICK; WOOD SHINGLE).. Construction: From the county property records.. Units: category. Source: Linden & Rockoff (2008)
note AGE: Age of the home at the time of sale.. Construction: sale_year minus year_built.. Units: years. Source: Linden & Rockoff (2008)
note NEW: 1 if the sale is of newly built property, else 0.. Construction: From the county property records.. Units: 0/1. Source: Linden & Rockoff (2008)
note orig_distance: Distance to the offender's address before limiting the sample to the 4-year arrival window.. Construction: Geocoded distance, pre-windowing.. Units: raw units. Source: Linden & Rockoff (2008)
note close_offender: 1 if the home is in the inner treated ring (within 0.1 mile of the offender's address), else 0.. Construction: Indicator distance <= 0.1 mile.. Units: 0/1. Source: Butts (2023)
note post_move: 1 if the sale closed after the offender's arrival, else 0 (pre-arrival).. Construction: Indicator that the sale date is after offender_address_date.. Units: 0/1. Source: Butts (2023)
note close_post_move: Interaction close_offender x post_move; its coefficient is the parametric ring-DiD ATT.. Construction: close_offender * post_move.. Units: 0/1. Source: Butts (2023)
note neighbor_year: Group identifier for neighborhood crossed with sale year.. Construction: group(neighborhood, sale_year).. Units: group ID. Source: Linden & Rockoff (2008)
note CNTLBQM1_2: 1 if building-quality part 1 equals 2, else 0 (hedonic control).. Construction: Indicator BQM1 == 2.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLBQM1_3: 1 if building-quality part 1 equals 3, else 0 (hedonic control).. Construction: Indicator BQM1 == 3.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLBQM1_4: 1 if building-quality part 1 equals 4, else 0 (hedonic control).. Construction: Indicator BQM1 == 4.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLBQM1_5: 1 if building-quality part 1 equals 5, else 0 (hedonic control).. Construction: Indicator BQM1 == 5.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLBQM1_6: 1 if building-quality part 1 equals 6, else 0 (hedonic control).. Construction: Indicator BQM1 == 6.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_2: 1 if exterior wall is asbestos shingle/siding, else 0 (hedonic control).. Construction: Indicator WALL == 'ASB SHNG/SDG'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_3: 1 if exterior wall is cedar/redwood, else 0 (hedonic control).. Construction: Indicator WALL == 'CEDAR,RDWD'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_4: 1 if exterior wall is exterior plywood, else 0 (hedonic control).. Construction: Indicator WALL == 'EXT PLYWOOD'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_5: 1 if exterior wall is face brick, else 0 (hedonic control).. Construction: Indicator WALL == 'FACE BRICK'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_6: 1 if exterior wall is masonite, else 0 (hedonic control).. Construction: Indicator WALL == 'MASONITE'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_7: 1 if exterior wall is other material, else 0 (hedonic control).. Construction: Indicator WALL == 'OTHER'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_8: 1 if exterior wall is stucco/hardcoat, else 0 (hedonic control).. Construction: Indicator WALL == 'STUCCO HRDCT'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_9: 1 if exterior wall is wood on sheathing, else 0 (hedonic control).. Construction: Indicator WALL == 'WOOD ON SHTG'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_10: 1 if exterior wall is wood shingle, else 0 (hedonic control).. Construction: Indicator WALL == 'WOOD SHINGLE'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLHEIGHT_2: 1 if building height class is 1.5, else 0 (hedonic control).. Construction: Indicator HEIGHT == 1.5.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLHEIGHT_3: 1 if building height class is 2, else 0 (hedonic control).. Construction: Indicator HEIGHT == 2.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLHEIGHT_4: 1 if building height class is 3, else 0 (hedonic control).. Construction: Indicator HEIGHT == 3.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLHEIGHT_5: 1 if building height class is 8, else 0 (hedonic control).. Construction: Indicator HEIGHT == 8.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLHEIGHT_6: 1 if building height class is 9, else 0 (hedonic control).. Construction: Indicator HEIGHT == 9.. Units: 0/1. Source: Linden & Rockoff (2008)
note log_price: Natural log of the home sale price (the regression outcome).. Construction: log(amt_Price).. Units: log US$. Source: Linden & Rockoff (2008)
save "raw_data.dta", replace

* ---- data_prepared.dta ----
use "data_prepared.dta", clear
label data "Analysis sample: offender catchment sales (<= 1/3 mile)"
note _dta: Analysis sample: rows with offender == 1 (within 1/3 mile of some offender's eventual address). Built from raw_data by rescaling distance to miles and adding dist_post, post, srn_year, and offdays for the estimators.
note sale_month: Calendar month in which the home transaction closed.. Construction: From the transaction record.. Units: 1-12. Source: Linden & Rockoff (2008)
note sale_year: Calendar year in which the home transaction closed.. Construction: From the transaction record.. Units: year. Source: Linden & Rockoff (2008)
note cpi_value: Consumer Price Index value used to deflate sale prices.. Construction: From the transaction record (price-deflation series).. Units: index. Source: Linden & Rockoff (2008)
note correction_term_dec_2004: Inflation-correction factor expressing prices in December-2004 dollars.. Construction: Derived from the CPI series, base = December 2004.. Units: factor. Source: Linden & Rockoff (2008)
note geo_code: Census geography identifier: state, county, tract, block group.. Construction: Geocoded census block-group code.. Units: code. Source: Linden & Rockoff (2008)
note parcel: Unique tax-parcel identifier for the property.. Construction: From the county property records.. Units: string. Source: Linden & Rockoff (2008)
note amt_Price: Recorded sale price of the home.. Construction: From the transaction record.. Units: US$. Source: Linden & Rockoff (2008)
note sale_date: Date the home transaction closed.. Construction: From the transaction record.. Units: date. Source: Linden & Rockoff (2008)
note year_built: Year the property was constructed.. Construction: From the county property records.. Units: year. Source: Linden & Rockoff (2008)
note HEATED: Heated floor area of the home.. Construction: From the county property records.. Units: sq ft. Source: Linden & Rockoff (2008)
note BATHS: Count of bathrooms in the home.. Construction: From the county property records.. Units: count. Source: Linden & Rockoff (2008)
note BEDROOMS: Count of bedrooms in the home.. Construction: From the county property records.. Units: count. Source: Linden & Rockoff (2008)
note neighborhood: Assessor neighborhood identifier (the clustering unit for SEs).. Construction: From the county property records.. Units: code. Source: Linden & Rockoff (2008)
note distance: Distance from the home to the nearest offender's eventual address; zero if the home is in no offender's area. In data_prepared this is rescaled to miles (raw distance / 3).. Construction: Geocoded straight-line distance; the analysis sample rescales to miles via distance / 3.. Units: miles (prepared) / raw units (raw_data). Source: Linden & Rockoff (2008)
note srn: Identifier of the registered offender whose catchment the home falls in.. Construction: Sex-offender registry record number.. Units: string. Source: Linden & Rockoff (2008)
note offender_address_date: Date the offender's address was published / the offender moved in.. Construction: From the sex-offender registry.. Units: date. Source: Linden & Rockoff (2008)
note building_grade: Assessor construction-quality grade of the home.. Construction: From the county property records.. Units: category. Source: Linden & Rockoff (2008)
note offender: 1 if the home is within 1/3 mile of some offender's eventual address (i.e. enters the analysis sample); else 0.. Construction: Flag derived from distance to nearest offender.. Units: 0/1. Source: Linden & Rockoff (2008)
note streetname: Street on which the property is located.. Construction: From the county property records.. Units: string. Source: Linden & Rockoff (2008)
note BQM1: First component of the assessor building-quality grade (1-6).. Construction: From the county property records.. Units: 1-6. Source: Linden & Rockoff (2008)
note BQM2: Second component of the assessor building-quality grade (1-6).. Construction: From the county property records.. Units: 1-6. Source: Linden & Rockoff (2008)
note AIRCOND: 1 if the home has air conditioning, else 0.. Construction: From the county property records.. Units: 0/1. Source: Linden & Rockoff (2008)
note HEIGHT: Coded number of stories / height class (1, 1.5, 2, 3, 8, 9).. Construction: From the county property records.. Units: code. Source: Linden & Rockoff (2008)
note WALL: Exterior wall-material category (e.g. ALUM,VINYL; FACE BRICK; WOOD SHINGLE).. Construction: From the county property records.. Units: category. Source: Linden & Rockoff (2008)
note AGE: Age of the home at the time of sale.. Construction: sale_year minus year_built.. Units: years. Source: Linden & Rockoff (2008)
note NEW: 1 if the sale is of newly built property, else 0.. Construction: From the county property records.. Units: 0/1. Source: Linden & Rockoff (2008)
note orig_distance: Distance to the offender's address before limiting the sample to the 4-year arrival window.. Construction: Geocoded distance, pre-windowing.. Units: raw units. Source: Linden & Rockoff (2008)
note close_offender: 1 if the home is in the inner treated ring (within 0.1 mile of the offender's address), else 0.. Construction: Indicator distance <= 0.1 mile.. Units: 0/1. Source: Butts (2023)
note post_move: 1 if the sale closed after the offender's arrival, else 0 (pre-arrival).. Construction: Indicator that the sale date is after offender_address_date.. Units: 0/1. Source: Butts (2023)
note close_post_move: Interaction close_offender x post_move; its coefficient is the parametric ring-DiD ATT.. Construction: close_offender * post_move.. Units: 0/1. Source: Butts (2023)
note neighbor_year: Group identifier for neighborhood crossed with sale year.. Construction: group(neighborhood, sale_year).. Units: group ID. Source: Linden & Rockoff (2008)
note CNTLBQM1_2: 1 if building-quality part 1 equals 2, else 0 (hedonic control).. Construction: Indicator BQM1 == 2.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLBQM1_3: 1 if building-quality part 1 equals 3, else 0 (hedonic control).. Construction: Indicator BQM1 == 3.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLBQM1_4: 1 if building-quality part 1 equals 4, else 0 (hedonic control).. Construction: Indicator BQM1 == 4.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLBQM1_5: 1 if building-quality part 1 equals 5, else 0 (hedonic control).. Construction: Indicator BQM1 == 5.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLBQM1_6: 1 if building-quality part 1 equals 6, else 0 (hedonic control).. Construction: Indicator BQM1 == 6.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_2: 1 if exterior wall is asbestos shingle/siding, else 0 (hedonic control).. Construction: Indicator WALL == 'ASB SHNG/SDG'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_3: 1 if exterior wall is cedar/redwood, else 0 (hedonic control).. Construction: Indicator WALL == 'CEDAR,RDWD'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_4: 1 if exterior wall is exterior plywood, else 0 (hedonic control).. Construction: Indicator WALL == 'EXT PLYWOOD'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_5: 1 if exterior wall is face brick, else 0 (hedonic control).. Construction: Indicator WALL == 'FACE BRICK'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_6: 1 if exterior wall is masonite, else 0 (hedonic control).. Construction: Indicator WALL == 'MASONITE'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_7: 1 if exterior wall is other material, else 0 (hedonic control).. Construction: Indicator WALL == 'OTHER'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_8: 1 if exterior wall is stucco/hardcoat, else 0 (hedonic control).. Construction: Indicator WALL == 'STUCCO HRDCT'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_9: 1 if exterior wall is wood on sheathing, else 0 (hedonic control).. Construction: Indicator WALL == 'WOOD ON SHTG'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLWALL_10: 1 if exterior wall is wood shingle, else 0 (hedonic control).. Construction: Indicator WALL == 'WOOD SHINGLE'.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLHEIGHT_2: 1 if building height class is 1.5, else 0 (hedonic control).. Construction: Indicator HEIGHT == 1.5.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLHEIGHT_3: 1 if building height class is 2, else 0 (hedonic control).. Construction: Indicator HEIGHT == 2.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLHEIGHT_4: 1 if building height class is 3, else 0 (hedonic control).. Construction: Indicator HEIGHT == 3.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLHEIGHT_5: 1 if building height class is 8, else 0 (hedonic control).. Construction: Indicator HEIGHT == 8.. Units: 0/1. Source: Linden & Rockoff (2008)
note CNTLHEIGHT_6: 1 if building height class is 9, else 0 (hedonic control).. Construction: Indicator HEIGHT == 9.. Units: 0/1. Source: Linden & Rockoff (2008)
note log_price: Natural log of the home sale price (the regression outcome).. Construction: log(amt_Price).. Units: log US$. Source: Linden & Rockoff (2008)
note dist_post: Distance interaction term used in the analysis sample.. Construction: distance * 10 * close_post_move (distance in miles).. Units: miles x 10. Source: This study (analysis sample)
note post: String label of the period: 'Pre' (before arrival) or 'Post' (after arrival).. Construction: ifelse(post_move == 1, 'Post', 'Pre').. Units: Pre/Post. Source: This study (analysis sample)
note srn_year: Neighborhood (offender) crossed with sale year; the fixed-effect cell in the parametric model.. Construction: paste(srn, sale_year, sep = '-').. Units: string. Source: This study (analysis sample)
note offdays: Days between the sale date and the offender's arrival date (negative = before arrival).. Construction: as.numeric(sale_date - offender_address_date).. Units: days. Source: This study (analysis sample)
save "data_prepared.dta", replace