"Dead Ends" dataset and analysis code

Zenodo (CERN European Organization for Nuclear Research)

Abstract

This repository contains the final dataset and analysis code for the paper "Dead Ends and Policy Change in Chinese Experimental Governance", forthcoming in the journal Governance. The file "fuzzy matches - final.db" contains the main sqlite3 dataset: table "exploration_results" contains all unique experiments 'query' as str: the experiment as extracted from the document 'fuzzy_matches' as list: table "matches" contains all matched phrases in policy documents "query_phrase" as str: the original experiment being traced "effective_date" as timestamp: the publication date of matching document "admin_lvl" as str: the administrative level of the agency issuing the matching document "sentence" as str: the matched sentence "title" as str: the title of the matched document "issuer_name" as str: name of the issuing agency "cat_tier_1" through, "cat_tier_3": three-tiered document hierarchy courtesey of Reading China Better "issuer_level_1" through "issuer_level_4": used for locating the precise administrative hierarchy of an issuing agency table "aggregated_results_with_flags" contains the aggregation of all matches "query_phrase" as str: the original experiment being traced total_references as int: count of unique matches first_effective_date as timestamp: first reference date first_issuer_name as str: name of first agency referencing the experiment first_issuer_level_1 through 4 as str: corresponding hierarchy last_effective_date: as timestamp date the experiment was last mentioned lifespan_days as float: difference in days between first and last effective date lifespan_years as float: difference in years between first and last effective date modal_issuer_name as str: most common issuer count_T0 through count_T8 as int: total number of references to experiment from T0 (corresponding to the year of first_effective_date) to eight years afterwards class_T0 through class_T8 as str: dominant status of implementation (Explore, Implement, Adjust, Abolish) scale_T0 through scale_T8 as str: scale (locally, large-scale, or nationwide, determined by how many unique issuing agencies refer to the experiment and at what levels) count_central_T0 through count_central_T8 as int: number of references by central agencies only has_received_central_support as int: binary 0/1 for whether the initiative received references in count_central matched_province as str: province where the experiment originated (if applicable) party_change_T0 through party_change_T6: binary 0/1 whether or not there was a change in party secretary in the province between T0-T6 (only applicable if experiment originated at provincial level) The two Juptyer Notebooks contain the python code used for analysis: DEADENDS-NOTEBOOK contains all code used to generate all Figures and Tables, relying on the database file and the .xlsx version of aggregated_results_with_flags. DEADENDS-FINAL-FUZZY MATCHING is provided for reference only to show the fuzzy matching technique (the raw data used for this is not provided, as it relies on proprietary data from Reading China Better).

Authors 1

1. Vincent Brussee corresponding Aachen

   Leiden University

   Affiliation as printed

   Leiden University