CT Safe Connect Repeat Callers by Demographics and Services

by Alberlis Hernandez

Last updated on January 18th, 2022

for Data Visualization for All
with Prof. Jack Dougherty
Trinity College, Hartford CT, USA

Story

The Connecticut Coalition Against Domestic Violence (CCADV) introduced a project called Safe Connect two years ago. This was developed so victims of domestic violence may have easier access to resources or assistance, via their new statewide phone/text/chat hotline. My class, Data Visualization for All, has partnered with CT Safe connect to analyze and visualize their data within the past two years. I will be focusing my data around the proportion of individuals that interact with Safe Connect multiple times since 2019. I will call them repeat callers, defined as case individuals who appeared 10 or more times in the data set. By setting this distinction, I hope to learn more about any patterns that differ repeat callers from individuals that do not interact with Safe Connect as much. Thus, I will answer the question: What can we learn about repeat callers according to region, demographics, and types of services? To do this, I will break up the proportion of repeat callers by said variables. Additionally, to normalize my data I will then compare the percentages of repeat callers within each variable with the percentages of callers who interacted with Safe Connect less than 10 times (non-repeat callers). It is relevant that I make this comparison so that we may identify if the percentage for repeat callers is unusually high or low. With this method, we can point to outliers in the data set as well as ensure our findings are consistent with what we may or may not expect based on what we know about non-repeat callers. Visualizing the data for repeat callers matters because we may realize what support by Safe Connect is being utilized the most and possible reasons that individuals may seek out constant help.

From my data, I learned that repeat callers are more likely to not identify their race, more likely to be 35 years of age or older, and are not requesting extra support from Safe Connect. Further down I will break down my findings using data visualizations I created.

Data

I created my data story by first defining what it means to be a repeat caller. I chose callers who have appeared in the data set at least 10 times in the two year time period, as it allows room for distinction between individuals who need more than one instance to get the resources needed, and those who request resources consistently. From there, I began to break up my data with pivot tables. I applied a filter on my tables so that my data would collect information about repeat callers and non-repeat callers separately. Each individual table I made focused on one variable at a time- age, type of service, or race. The purpose of me having my variables separate is so that I can analyze each section, and make it easier to understand commonalities or factors that may or may not exist for repeat callers.

In Figure A, we read that the majority of individuals who are repeat callers identify as Other or Unknown for their race, meaning many indivudals likely are choosing to not provide this information. This group accounted for 34.48% of cases who did not have constant contact with Safe Connect while it compromised 26.29% of repeat callers. This is a difference of 8.19 percentage points. A second insightful comment to mention from this table is that individuals who identify as Black represented the fourth highest percentage of non-repeat callers with 15.28%. Yet, this percentage rose for repeat callers, showing a 7.44 percentage increase. These two races had the largest differences between repeat callers and non-repeat callers.

Race % Repeat Contact % Non-repeat Contact Percentage Point Difference
Other-Unknown 26.29 34.48 -8.19
Black 22.72 15.28 7.44
Asian 2.78 1.34 1.44
Latino 21.14 22.06 -0.92
Multiple 2.97 3.04 -0.08
White 24.10 23.79 0.31

Figure A: Distribution of Race by Percentage

In Figure B, we notice that the gap between individuals who identify as 35 years old or older takes the majority in the distribution of age for both repeat callers and non-repeat callers. We can also notice that this age group increased, and accounted for double or more than the representation of other age groups for repeat callers.

Figure B: This chart displays the percentage that age groups make up for repeat and non-repeat callers.

Figure C shows us a map providing the percentage point difference between repeat callers and non-repeat callers who received counseling support from Safe Connect across the regions of Connecticut. This difference was positive across every region, meaning that repeat callers consistently had sought out more counseling support throughout Connecticut than non-repeat callers. This finding is not unusual, as counseling support is defined in the data set to be when a caller spends more than 10 minutes on the phone with Safe Connect. Therefore we expect repeat callers to have a higher percentage of counseling support than non-repeat callers. The region with the highest percentage point difference was Torrington, with 2.12. We can conclude that individuals have a need for greater time with Safe Connect than non-repeat callers in that region. In contrast, Waterbury region has the lowest percentage point different of 0.26, meaning repeat callers do not spend a significant increase in communication with Safe Connect than non-repeat callers.

Figure C: This map provides detailed, regional data about the percentage of callers who received counseling support from Safe Connect.

To continue, Figures D-G shows the percentage point difference for types of other services that Safe Connect provides. These services are Civil Legal, Criminal Justice, Safety Planning, and Victim Advocacy support respectively. Figure F has a negative percentage point difference for every single region, while Figures D and E only have one region with a positive percentage point difference (Enfield and Killingly, respectively), yet Figure G had mixed values. I was surprised to encounter so many negative values, as this means repeat callers are requesting extra services from Safe Connect less often than non-repeat callers. I originally assumed that repeat callers are having a large number of interactions with Safe Connect because they seek out some of their services, but the data showed otherwise. From this, we can question the motives for individuals who are repeat callers as to why they have been continuously in contact with Safe Connect as well as the efficiency of the services offered.

Figure D: Distribution of Civil Legal Support by Percentage 2020-21 Figure E: Distribution of Criminal Justice Support by Percentage 2020-21
Region % Repeated contacts % Non-repeat contacts Percentage Point Difference Region % Repeated contacts % Non-repeat contacts Percentage Point Difference
Ansonia region 15.45 16.61 -1.16 Ansonia region 0.00 1.52 -1.52
Bridgeport region 12.75 18.49 -5.74 Bridgeport region 0.34 1.00 -0.66
Danbury region 13.28 18.30 -5.01 Danbury region 0.00 0.83 -0.83
Enfield region 23.81 19.42 4.39 Enfield region 1.19 1.25 -0.06
Hartford region 13.69 16.73 -3.04 Hartford region 0.22 1.07 -0.85
Killingly region 19.44 23.60 -4.15 Killingly region 2.78 0.95 1.83
Mansfield region 18.33 18.71 -0.38 Mansfield region 0.00 1.43 -1.43
Meriden region 30.83 20.61 10.22 Meriden region 0.00 1.38 -1.38
Middletown region 5.56 17.34 -11.78 Middletown region 1.01 1.68 -0.67
New Britain region 17.28 17.83 -0.55 New Britain region 0.28 1.35 -1.06
New Haven region 9.48 17.24 -7.76 New Haven region 0.38 0.99 -0.61
New London region 14.12 20.82 -6.70 New London region 0.59 1.07 -0.48
Stamford region 16.47 16.82 -0.35 Stamford region 0.00 0.93 -0.93
Torrington region 2.86 15.04 -12.18 Torrington region 0.00 1.50 -1.50
Waterbury region 6.84 16.24 -9.40 Waterbury region 0.85 1.01 -0.15
#N/A 12.18 13.31 -1.13 #N/A 0.96 0.66 0.30
Grand Total 12.86 17.47 -4.61 Grand Total 0.44 1.09 -0.66
Figure F: Distribution of Cases that Received Safety Planning by Percentage 2020-21 Figure G: Distribution of Victim Advocacy Support by Percentage 2020-21
Region % Repeated contacts % Non-repeat contacts Percentage Point Difference Region % Repeated contacts % Non-repeat contacts Percentage Point Difference
Ansonia region 40.65 43.73 -3.08 Ansonia region 42.28 48.22 -5.95
Bridgeport region 41.95 45.58 -3.63 Bridgeport region 46.98 48.03 -1.05
Danbury region 32.81 44.91 -12.09 Danbury region 37.50 46.02 -8.52
Enfield region 33.33 46.59 -13.26 Enfield region 45.24 47.53 -2.30
Hartford region 37.71 45.15 -7.44 Hartford region 42.20 47.58 -5.38
Killingly region 33.33 43.86 -10.53 Killingly region 50.00 49.10 0.90
Mansfield region 38.33 44.68 -6.35 Mansfield region 41.67 47.96 -6.29
Meriden region 28.33 42.69 -14.36 Meriden region 50.00 49.98 0.02
Middletown region 34.85 44.94 -10.10 Middletown region 51.01 45.05 5.96
New Britain region 38.24 48.05 -9.80 New Britain region 46.18 44.31 1.86
New Haven region 34.37 45.04 -10.67 New Haven region 37.75 47.18 -9.43
New London region 33.53 46.13 -12.60 New London region 34.71 45.37 -10.66
Stamford region 30.59 45.10 -14.52 Stamford region 46.47 50.50 -4.03
Torrington region 35.71 41.92 -6.20 Torrington region 40.00 47.46 -7.46
Waterbury region 33.76 42.97 -9.21 Waterbury region 48.72 48.24 0.48
#N/A 41.03 41.92 -0.89 #N/A 8.65 19.25 -10.60
Grand Total 36.18 44.82 -8.64 Grand Total 40.20 45.49 -5.29

The above tables are data for the current types of services Safe Connect has avaliable and the percentage of individuals who use them, by region.

Access to my masked raw data story can be found at my Google Sheets.

Caution and Uncertainties

The largest caution to address for my data is that I am limited by the accuracy of the data set I was provided from Safe Connect. There may be common irregularities caused by human error so my data or findings are not a complete representation of the callers Safe Connect has had. This is evident because many callers did not provide personal demographic or regional information which has the possibility of skewing the numbers. I also do not wish for readers to assume that each caller is a victim of domestic abuse. Our partners at CT Coalition Against Domestic Violence mentioned that they do have interactions with family members of domestic violence and do not only serve this community. The data we were given did not collect the reasoning behind any individual case or caller, so it would be useful to have more data to truly understand the intentions behind repeat callers CCADV has had in the past 24 months. Finally, I looked at my variables (demographics and types of services) separately, therefore my data is not meant to look at the relationship between each other but what they each can help us conclude about repeat callers as a population in Safe Connect.

Credits

I would like to give thanks to Maria Guzman and Joanne Vitarelli, our CT Coalition Against Domestic Violence partners. I also thank my instructors for the course, Jack Dougherty and Myri Ayala.