{ "study": "W3-2026-10-24-v1.0", "observationDate": "2026-10-24", "title": "Method and codebook", "sections": [ { "heading": "Population, sample and unit", "paragraphs": [ "Target topic: the public instructions an England resident encounters when preparing a planning comment. Observed unit: one fixed official guidance entry URL per council, as listed in 03_SOURCE_REGISTER.json (C01-C08). This is an exploratory purposive sample, not a representative survey or preregistered experiment.", "Sample fixed from W0: Camden, Hackney, Wandsworth, Southwark, Islington, Kensington and Chelsea, plus Redbridge as a nearby editorial comparator and Leeds as a non-London comparator. Seven of eight are London boroughs. No councils were added or removed because their results were interesting or because a register fetch failed.", "Entry URLs are a replication boundary, not a claim that every resident starts there. RBKC's entry is its decision-process page, which includes commenting instructions; this is not perfectly matched to dedicated comment pages elsewhere. All are official pages that explain a comment route.", "Observation date: 24 October 2026. Public web text retrieval was used without a council account. Access timestamps below day precision, device viewport and actual resident browser performance were not captured. Some retrieved text may be cached. Page update labels are recorded separately from observation dates and are not treated as proof of freshness." ] }, { "heading": "Exact replication procedure", "steps": [ "Use the eight C01-C08 URLs in the source register in the stated sample, preserving the sample and existing rows. Open the full entry text; record retrieval date, page title, visible update label and any access limitation. Do not substitute a more informative page into the entry count.", "Code all six fields against the entry text only, using the definitions below. Record the relevant section and a short original paraphrase. If a claim is not explicit, use NOT_STATED rather than NO. If entry text cannot be inspected, use UNOBSERVABLE; do not turn retrieval failure into absence.", "Follow the entry page's identified planning-search/register link once. Record the destination URL and tool result. The recorded hop starts at the fixed entry page, not the council homepage. A returned page is not a completed application journey. Do not register, upload or submit.", "Consult supplementary official pages when needed to interpret a claim. Label these S sources; they can qualify the report but cannot alter the entry-page counts. Keep contradictory wording visible. N01 supplies national context, not a substitute local deadline.", "Leave application_reference and actual_comment_deadline null unless a separately scoped record-level audit actually verifies a named application, notice, date and status. W3 did not do so. Never fill these values by adding 21 days to a receipt, validation or publication date.", "Count YES values across the eight rows separately for each field. Do not sum the fields into a score. Preserve NOT_STATED and UNOBSERVABLE denominators. The included verification script reproduces the counts and asserts that no submission or application-deadline test is claimed.", "Re-open the claimed source sections for a second verification pass. W3 completed a same-session text check of the eight entry pages and inspected the principal privacy/deadline qualifications. This is a ChatGPT self-check, not a separate researcher, professional or independent audit.", "On refresh, retain the old version and produce a row-level change log. If the question or sample changes, give the study a new version and do not silently compare unlike measures." ] }, { "heading": "Codebook", "table": { "headers": [ "Field", "YES rule", "Boundary" ], "rows": [ [ "numeric_period_on_entry", "Entry specifies a numerical duration for commenting, including days or weeks.", "A usual duration is not the actual closing date. No requirement that every application type is covered." ], [ "account_explicit_on_entry", "Entry expressly requires registration or sign-in for the online comment route.", "Does not imply offline comments need an account; silence does not mean no login." ], [ "prepared_file_explicit_on_entry", "Entry instructs the user to prepare/save a document for online upload.", "Do not infer from an attachment option or from what a portal might allow." ], [ "late_after_consultation_explicit", "Entry explicitly discusses comments after the consultation period.", "Both acceptance and warning qualify. 'Before decision' alone, without the post-consultation connection, does not qualify." ], [ "personal_detail_publication_explicit", "Entry states treatment of personal information or identifying details in publication.", "General 'comments published' alone is insufficient; this is not a privacy quality score." ], [ "maintenance_fallback", "Entry displays a maintenance notice and gives an alternative comment route.", "A time-specific wording observation, not evidence of duration, cause or actual service unavailability." ] ] } }, { "heading": "Reproduced findings", "table": { "headers": [ "Entry-page observation", "Count / 8", "Councils" ], "rows": [ [ "Numerical comment period stated", "5", "Wandsworth; Southwark; Kensington and Chelsea; Redbridge; Leeds" ], [ "Online account explicitly required", "3", "Southwark; Redbridge; Leeds" ], [ "Preparation of a saved file explicit", "1", "Wandsworth" ], [ "After-consultation comments explicitly addressed", "3", "Camden; Wandsworth; Redbridge" ], [ "Personal-detail publication addressed", "5", "Camden; Wandsworth; Southwark; Redbridge; Leeds" ], [ "Maintenance notice with alternative route", "1", "Hackney" ] ] }, "paragraphs": [ "Counts describe the selected pages, not England's councils. Each field has eight observable entry pages; remaining values are NOT_STATED. No confidence intervals, composite score, percent ranking or causal inference is justified." ] }, { "heading": "Register-route observations", "paragraphs": [ "Each entry provided a register/search link and one outgoing hop was attempted. Camden, Redbridge and Leeds returned search-page text. Wandsworth returned a research fetch 502, Southwark timed out, Hackney returned 406, Islington and RBKC returned 403. These are limitations of the research retrieval path. They are not evidence of five council outages and must not be used as a public headline.", "The Hackney maintenance observation is separately supported by its official entry-page notice. Even there, the study does not establish outage duration or whether a resident could use the portal at the time. No full commenting transaction was tested.", "A trial Camden application-record retrieval produced no usable text and was excluded from findings. It supplies no deadline, record audit or application sample." ] }, { "heading": "Conflicts and deliberately omitted claims", "paragraphs": [ "Islington's specific development-management privacy notice describes full comments not being published online, while its general process page uses broader publication wording. Both are registered. The report attributes the specific notice but explicitly retains the discrepancy; it promises no confidentiality.", "Redbridge and Leeds list application-type consultation distinctions. W3 does not restate those lists as comprehensive legal rules. In particular, Redbridge's grouping includes listed building consent, creating a reason to avoid broad legal inference from that list. N01 and the actual application publicity govern the limited reader advice.", "RBKC mentions Proposal Details for milestones; W3 did not verify a comment-deadline field there. Leeds' statement about most undecided applications does not meet the explicit post-consultation code by itself. Southwark's at-least-21-day wording is reported as council guidance, not converted into a date calculator.", "Where email addresses are masked in text retrieval, no council address was guessed. The report links to the official route rather than reproducing uncertain contact details." ] } ] }