JDR Analysis 1

Published

March 6, 2024

Almost all of the courts consistently have counts for Summons + Postin that are greater than the summation of its subcategories. Is this error random or systematic?

Factors in consideration:

  • Court size
  • JMS system
  • ???

Let’s take a look at historical accuracies across all counties…

Does accuracy vary by court size?

I recieved questions about accuracy possibly being JMS specific…

Histocial JMS accuracy:

# A tibble: 7 × 2
  jms_system             avg
  <chr>                <dbl>
1 JMIS                 0.947
2 JSI                  0.937
3 Agile Jury           0.929
4 Custom               0.787
5 Other                0.753
6 Judicial Systems Inc 0.706
7 Web Enhanced JS      0.540

The data does not indicate a relationship between any specific JMS and accuracy rates

Other stats…

Least accurate counties this FY

# A tibble: 58 × 3
# Groups:   county [58]
   county         jms_system                                            avg_prop
   <chr>          <chr>                                                    <dbl>
 1 Inyo           Jury Systems Incorporated                                0.206
 2 Monterey       Jury Systems Incorporated                                0.295
 3 San Bernardino Web Based Jury Enhancement System created by San Ber…    0.588
 4 Placer         Tyler Technologies                                       0.645
 5 Shasta         Jury Systems Incorporated                                0.701
 6 Colusa         Jury Systems Incorporated                                0.727
 7 Orange         Avenu- Clearview                                         0.862
 8 Yolo           Jury Systems Incorporated and Tyler Technologies         0.894
 9 San Mateo      Jury Systems Incorporated                                0.902
10 Marin          Jury Systems Incorporated                                0.904
# ℹ 48 more rows

Conclusions:

Prelimenary analysis shows that smaller courts tend to have lower accuracy. Reporting accuracy also seems to be JMS agnostic. JSI is the most commonly used system, but accuracy still varies per court.

Additionally, the table of historical accuracy by county shows that only the bottom ~10% are below 90% accuracy. These courts should be focused on. In addition to paying more attention to these specific courts, further analysis should probe the subcategories of each element in Summons + Postin to identify possible systematic reporting bias per question.