Data Story

Regional Consistency

Are SafeConnect support services provided with the same frequency across Connecticut?

The Connecticut Coalition Against Domestic Violence (CCADV) is the state's central domestic violence hotline providing 7 different services over the phone to over 25,000 CT residents in need a year, those seven consisting of counseling, victim advocacy, information providing, civil legal support, referrals, criminal justice support, and safety planning. After looking over the CCADV data from the past 2 years, we wanted to see where the agency’s services were being provided most, in order to see if there are any inconsistencies between regions. More specifically, we wanted to know if any of the 7 services might be recommended disproportionately in certain areas.

Our findings bode well for the CCADV, as there is a remarkable consistency in counseling servces, victim advocacy, safety planning, information provision, and referrals across each state region. While in terms of overall services provided, New Haven and Middletown are at the lower end of the spectrum and Stamford is at the higher end, the spread between them is incredibly tight. Most notably, criminal justice support is inconsistent across the regions, and civil/legal support is to a lesser degree, which we focus on in Maps 3 and 4 and Tables 1 and 2. In short, the frequency of how these services are provided is strongly consistent, with some variation in the criminal justice and civil/legal fields.

Inconsistency in the Criminal Justice and Civil/Legal Services

We decided to focus the rest of our charts on the services that had the most variation from region to region, as these services have more significant ramifications for individual clients and require more time and resources from the CCADV. We first considered the Criminal Justice data found in chart 1. Although the service is recommended in the lowest percentage of total calls, it has by far the most variance in between regions. Most interestingly, Middletown, the region with the fewest overall services recommended per call also is by far the most likely to recommend Criminal Justice services while Stamford, the region recommending the most overall services per call is one of the least likely to recommend criminal justice support. There is also a significant variation in the Torrington and Danbury regions, which, despite having relatively similar demographics and number of services recommended, have wildly different percentages of calls offering Criminal Justice support.

This criminal justice trend continues throughout the sheet, with a handful of outliers. Oddly enough, the trend does not continue into the civil/legal instances chart. We think that this is due to the lower overall percentage of criminal justice instances; we trust they are more reliable numbers as their lower values end up influencing the overall services provided values less. This would explain why the trend does not follow in other provided services. Still, there are interesting bits of data in the civil/legal section as well: Torrington and Killingly have a similar frequency of calls, yet their civil/legal support numbers are on opposite ends of the spectrum. For reference, Killingly has a higher number of total service provision than Torrington. Perhaps these are coincidences, but we found it interesting enough to mention.

While we had access to a lot of raw data, normalizing it in such a way that made it meaningful was a good bit of work. You can view the work we did on our google sheet, which will be linked at the end of the section. Essentially, we began with the masked CCADV data and the Connecticut census data and went from there, making pivot tables, normalized data tables, and contextualizing the data in any possible way we thought would be useful. Even the slightest chance of using the data later meant it got attention. The most in-depth we went was with some of the CCADV data, in which we tracked the action taken post-call by the CCADV, and organized it by race, region, and proportionally to residents in both of those categories. We found out fairly late in the process that we'd based a lot of our data on a simple miscalculation early on, which led to a lot of repeat work and edited visualizations and findings.

While we are fairly confident in our data analysis, there are some uncertainties in our final work. One of our more recent findings, the discussion of criminal justice trends in maps 3 and 4 and tables 1 and 2, needs more fleshing out. There are a number of other potentially influencing factors we did not explore, and criminal justice is a small sample size of instances to be investigating for trends, but the trend was interesting enough to note. We did not include data on victims from outside of Connecticut, as it did not fit well into charts or maps that were specifically state-based. We also had some miscalculations in our earlier drafts of this report, and while we fixed each of those, mistakes have been made before. As always, it is important to note that while we believe there’s a correlation here, there are a number of other factors influencing all of this information, and so the causal relationship we are looking for may not exist in the way we think it does.

Credits

The Connecticut Coalition Against Domestic Violence. CCADV SafeConnect 2020 and 2021 data provided by Joanne Viratelli to Jack Dougherty, Trinity College, Dec 2021 & Jan 2022.

SocialExplorer American Community Survey, 2015-2019 Census data estimates.