Low-Income Litigants in the Sandbox: Court Record Data and the Legal Technology A2J Market
Notice bibliographique
Résumé
(Excerpt) Katrina was a community college student with two children, trying to juggle work, childcare, and school. During class in the spring of 2018, her phone buzzed incessantly. She looked down to see a message from her roommate saying a process server had shown up at the house to deliver a summons and complaint, naming Katrina in a lawsuit filed in county court by a debt collection company she had never heard of. Katrina turned to the internet for help and found herself overwhelmed with advertisements that began to pop up in her social media feeds trying to get her to enroll in debt settlement companies, or offering help filing bankruptcy, with or without a lawyer. Katrina didn't know which of these tools to trust, and the court self-help website was overwhelming and full of confusing information that was hard to read on her mobile phone. Katrina is one of the estimated 71 million people in the United States with debt in collections and was one of almost a quarter of a million Californians sued for debt in 2018, almost all of whom have to navigate a state civil court system as unrepresented litigants against professional debt collection lawyers. Consumer debt collection cases comprise an increasing percentage of the dockets of most state civil courts in the United States. In California, over the last ten years, debt collection cases totaled an average of 20% of all cases filed, with debt cases rising to 37% of all civil filings in 2019. It is estimated that of the 71 million consumers who have debt in collections, 15% were sued in the last year. That means, according to the research of the Aspen Institute, an estimated 12 million people were sued across the United States to collect a consumer debt in the last year (most commonly credit card, medical debt, auto deficiency, or other consumer unsecured debt). The exact number of people sued on consumer debt cases in state courts each year is not known, because these data points are lost in a myriad of state court case management systems. Researchers and advocates know the exact number of businesses and consumers litigating in federal court, through the unified federal court management system PACER, and a rising number of data analytics companies, from Bloomberg to Lex Machina and Ravel Law, promise law firms and corporations ever-detailed information about judicial behavior and case trends. Also available through the federal PACER system is docket-level information about consumer bankruptcy filings, leading to empirical analysis of legislative changes to the bankruptcy code, but in the area of civil justice, as administered by state courts, there is a "severe data deficit." Recently, states have moved to obtain better criminal case record data in recognition of the necessity for empirical data as a predicate for crafting criminal justice policy, but for many types of civil cases, access-to-justice scholars and other vital stakeholders do not know what is happening in many state courts, particularly in states with disaggregated case management systems. Recently released tools such as FastCase, Docket Alarm, Lexis Advance Courtlink, and Westlaw Docket Search have access to some state civil filings, but these databases of state court records are only as complete as the records states choose to make available through online access. Access may diverge by county within a state, including California, where some counties and case types are available online while others are not.
Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.
Comment cette classification a été obtenuedéplier
Prédiction machine sur la base complète
Imitation des enseignantsNi prévalence calibrée, ni vérité terrain. Validation humaine à venir. Le volet Gemma est une étiquette directe du modèle pour chaque travail de la base, lue sur la notice réduite au titre. Le volet Codex est un classifieur appris des 10 348 étiquettes directes de Codex et calibré sur les taux pondérés de l'échantillon; les champs sans appui suffisant ne portent aucun appel Codex. Le mode candidate est l'union des deux volets; le consensus est leur intersection. Ces sorties portent le statut machine_predicted_unvalidated et ne sont pas des étiquettes humaines.
Scores du classifieur distillé par catégorie (deux têtes)
| Catégorie | Codex | Gemma |
|---|---|---|
| Métarecherche | 0,003 | 0,026 |
| Méta-épidémiologie (sens strict) | 0,000 | 0,000 |
| Méta-épidémiologie (sens large) | 0,000 | 0,000 |
| Bibliométrie | 0,003 | 0,005 |
| Études des sciences et des technologies | 0,005 | 0,001 |
| Communication savante | 0,010 | 0,008 |
| Science ouverte | 0,001 | 0,003 |
| Intégrité de la recherche | 0,003 | 0,003 |
| Charge utile insuffisante (le modèle a refusé de juger) | 0,098 | 0,017 |
Scores machine (provisoires)
Les deux têtes enseignantes du modèle étudiant, lues sur ce travail. Un score ordonne la base pour la relecture; il n'affirme jamais une catégorie, et le statut de validation accompagne chaque rangée tel quel.
Scores de référence d'un modèle non mature (critères de maturité non atteints, 7 itérations). Un score ordonne; il n'affirme jamais une catégorie.
score_only:v0-immature-baseline · tel quel depuis la passe de notation : score_only signifie que le nombre peut ordonner les travaux, et qu'aucune étiquette de catégorie n'en découleClassification
machine, non validéePrédiction automatique; un appel candidat d’une seule source (Gemma direct ou Codex distillé), pas un consensus.
Le détail, modèle par modèle et score par score, se trouve en fin de page sous « Comment cette classification a été obtenue ».