Parsing Through Predictions
Abstract/Contents
- Abstract
- Can algorithms send us to jail? This research analyzes the role predictive algorithms play in sorting defendants in and out of custody. Specifically, it investigates the Santa Clara County pretrial risk assessment instrument and how judges and legal professionals interpret its scores to make pretrial decisions. Are risk assessment scores the primary determinants of pretrial decisions or just consulted as supplementary information? Moreover, how much do judges trust the information from risk assessment tools given their history of racially prejudiced training data and scores? 14 interviews with legal professionals in Santa Clara County and Orange County were conducted to tackle these questions. After the analysis, a hierarchy of information became apparent whereby some forms of information had more bearing on a judge's decision than risk assessment scores. Judges prioritize information in the following order: 1) Swing facts or improvements to a defendant's behavior or lifestyle had the most bearing on pretrial decisions. They have the potential for judges to go against pretrial services recommendations. 2) Simple facts about a defendant's case such as, how many failures to appear, were very indicative of a defendant's future behavior for a judge. They are the most reliant forms of information for judges. 3) Lastly, risk assessment scores or indicators that aggregate the aforementioned facts into a number or score are consulted. These are treated as supplemental and somewhat nebulous forms of information on which to base a decision. Overall, the risk assessment score was not the determining factor for pretrial custody decisions. I make a call for algorithmic literacy, weeding out factors and algorithmic transparency to make risk assessment scores more useful for pretrial decisions.
Description
| Type of resource | text |
|---|---|
| Date created | June 1, 2019 |
Creators/Contributors
| Author | Williams, Zora |
|---|---|
| Advisor | Angèle Christin |
| Degree granting institution | Stanford University, Program in Science Technology and Society |
Subjects
| Subject | risk assessment tools |
|---|---|
| Subject | risk assessment scores |
| Subject | pretrial decisions |
| Subject | judges |
| Subject | predictive algorithms |
| Subject | courts |
| Subject | justice |
| Subject | judicial discretion |
| Subject | Science Technology and Society |
| Genre | Thesis |
Bibliographic information
| Related item | |
|---|---|
| Location | https://purl.stanford.edu/hj402wb7969 |
Access conditions
- Use and reproduction
- User agrees that, where applicable, content will not be used to identify or to otherwise infringe the privacy or confidentiality rights of individuals. Content distributed via the Stanford Digital Repository may be subject to additional license and use restrictions applied by the depositor.
- License
- This work is licensed under a Creative Commons Attribution 3.0 Unported license (CC BY).
Preferred citation
- Preferred Citation
- Williams, Zora. (2019). Parsing Through Predictions: A Critical Analysis of the Credibility and Application of Pretrial Risk Assessment Scores. Unpublished Honors Thesis. Stanford University, Stanford CA. Available at https://purl.stanford.edu/hj402wb7969.
Collection
Stanford University, Program in Science, Technology and Society, Honors Theses
View other items in this collection in SearchWorksContact information
- Contact
- zorazambezi@gmail.com
Also listed in
Loading usage metrics...