| # Codebook — NBA draft × high school pathways panel |
|
|
| File: `data/nba_draft_hs_panel.csv` |
| Unit of observation: one drafted player-pick (typically 60 per draft year). |
|
|
| ## Columns |
|
|
| | Column | Type | Description | |
| | --- | --- | --- | |
| | `draft_year` | int | NBA draft year (2016–2026 in this release). | |
| | `pick` | int | Overall pick number (1–60 typical). | |
| | `round` | int | 1 if pick ≤ 30, else 2. | |
| | `player_name` | string | Player name as in source. | |
| | `nba_team` | string | Drafting franchise abbreviation (Basketball-Reference). Empty for curated 2026 if not yet joined to a team table. | |
| | `college_or_prior` | string | College, G League, or international club / pathway listed as prior stop. | |
| | `high_school_finish` | string | **Finishing** secondary school / academy (last U.S. HS or prep program competed for). Empty if unknown or no U.S. HS. | |
| | `hs_city` | string | City/town of finishing school when available. | |
| | `hs_state` | string | U.S. state (name or postal abbr) or `intl` / non-U.S. label. | |
| | `school_type` | enum | See definitions below. | |
| | `school_type_method` | enum | How `school_type` was assigned (see below). | |
| | `ambiguity_flag` | string | Non-empty when classification is contested or pending review. | |
| | `notes` | string | Free-text caveats, prior schools, tuition citations (curated rows). | |
| | `place_median_hh_income` | number or empty | ACS 2024 5-year median household income for HS city/place (public rows; see `data/public_hs_town_income.csv`). Ecological town proxy — not family income. | |
| | `state_median_hh_income` | number or empty | ACS 2024 5-year state median HH income (paired with place income for public rows). | |
| | `hs_source` | string | Provenance of HS attribution. | |
| | `bbref_player_url` | string | Basketball-Reference player URL when scraped. | |
|
|
| ## `school_type` definitions |
| |
| Mutually exclusive finishing-school buckets (same rules as the 2026 case study): |
| |
| | Value | Meaning | |
| | --- | --- | |
| | `public` | U.S. public district or public magnet high school. | |
| | `local_private` | Catholic / Episcopal / Christian / independent **day** school that is **not** a national boarding / prep magnet. | |
| | `elite_prep` | National boarding prep, postgraduate academy, or alternative academy pathway used as finishing school (e.g. Link, IMG, Montverde, Oak Hill, Brewster, Wasatch, Overtime Elite, Sunrise Christian, Compass Prep). | |
| | `international` | No U.S. high school finish (overseas club / federation / academy pathway). | |
| | `unknown` | U.S. finishing school named (or missing) but **not** yet hand-reviewed for public vs local private; **or** HS attribution missing. | |
|
|
| **Do not treat `unknown` as a substantive pathway category in papers** — it is a data-quality status. |
|
|
| ## `school_type_method` |
|
|
| | Value | Meaning | |
| | --- | --- | |
| | `curated_hand` | Human-reviewed (2026 case study). May include `public` / `local_private`. | |
| | `curated_override_csv` | Matched `sources/curated_overrides/school_type_overrides.csv` (name+state; web-sourced). | |
| | `auto_international_no_us_hs` | No U.S. HS on source and/or clear overseas pathway. | |
| | `auto_elite_prep_dictionary` | Finishing name matched curated elite-prep dictionary (`scripts/classify_school_type.py`). | |
| | `pending_public_private_review` | U.S. HS present; public vs local_private not assigned. | |
| | `missing_hs_attribution` | No finishing school string available. | |
| |
| ## Important measurement notes |
| |
| 1. **Finishing school, not birthplace or first HS.** Prior stops (e.g. public → Link Academy) do not change the bucket. |
| 2. **No invented schools or types.** If sources disagree, keep the BBRef/curated finishing name and flag ambiguity. |
| 3. **Income columns are optional** and currently sparse (2026 public-HS towns only). |
| 4. **2026 coverage** may reflect a hand-built first-round-scale slate from the case study; confirm against the official NBA draft board when locking a paper sample. |
| |
