Introduction

This project examines child abuse and neglect investigation data for Bexar County, Texas. The analysis focuses on county-level victim counts reported in completed child abuse and neglect investigations across fiscal years 2016 through 2025.

Data

The data used in this project come from the Texas Department of Family and Protective Services (DFPS). The dataset contains county-level information on victims reported in completed child abuse and neglect investigations.

The dataset is accompanied by documentation describing the measures and definitions used in the data. The dataset is publicly available and contains aggregated information by county, fiscal year, and program rather than individual-level identifying information.

The dataset is relevant to Public Administration because it contains information about child abuse and neglect investigations administered through a state government agency, the Texas Department of Family and Protective Services.

For this analysis, the data were limited to Bexar County and fiscal years 2016 through 2025. The dataset contains two programs, DCI and RCI, and distinguishes between not-confirmed and confirmed victims.

The original Bexar County subset contained 40 records. After restructuring the data so that not-confirmed and confirmed victim counts could be compared for each fiscal year and program, the final analytic dataset contained 20 program-year observations.

Data Source and Suitability

This dataset is appropriate for an exploratory Public Administration analysis because it describes the operation of a public child welfare program administered through the Texas Department of Family and Protective Services. The data are reported at the county, fiscal-year, and program level rather than identifying individual children.

The dataset is listed as public in the Data.gov catalog and provides a CSV download. The dataset documentation also provides definitions for the victim categories and describes the scope of completed investigations. Data.gov

Data Cleaning

The original dataset was filtered to include only observations from Bexar County. The victim-count variable was converted from character data to numeric data so that statistical calculations could be performed.

The data were then reorganized so that not-confirmed and confirmed victim counts could be compared for each fiscal year and program. This resulted in 20 program-year observations representing 10 fiscal years and two programs.

There were no missing values in the fiscal year, not-confirmed victim count, or confirmed victim count variables used in the analysis.

Missing Data

There were no missing values in the variables used for the analysis.

colSums(is.na(bexar_wide[, c(
  "Fiscal.Year",
  "Not.Confirmed",
  "Confirmed"
)]))
##   Fiscal.Year Not.Confirmed     Confirmed 
##             0             0             0

Descriptive Statistics

There were 20 observations in the analysis, covering fiscal years 2016 through 2025 across the DCI and RCI programs.

The mean number of not-confirmed victims was 249.85, while the median was 191.5. The values ranged from 77 to 756.

The mean number of confirmed victims was 21.2, while the median was 17. The values ranged from 7 to 74.

Summary Statistics

summary(bexar_wide$Not.Confirmed)
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    77.0   102.5   191.5   249.8   351.5   756.0
summary(bexar_wide$Confirmed)
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##    7.00   14.25   17.00   21.20   24.75   74.00

Histograms

The distributions of not-confirmed and confirmed victims were examined using histograms.

hist(
  bexar_wide$Not.Confirmed,
  main = "Not Confirmed Victims in Bexar County",
  xlab = "Number of Not Confirmed Victims"
)

hist(
  bexar_wide$Confirmed,
  main = "Confirmed Victims in Bexar County",
  xlab = "Number of Confirmed Victims"
)

Relationship Between Variables

A scatterplot was used to examine the relationship between not-confirmed and confirmed victim counts.

plot(
  bexar_wide$Not.Confirmed,
  bexar_wide$Confirmed,
  main = "Not Confirmed and Confirmed Victims in Bexar County",
  xlab = "Number of Not Confirmed Victims",
  ylab = "Number of Confirmed Victims"
)

Correlation

The correlation between not-confirmed and confirmed victim counts was calculated using Pearson’s correlation coefficient.

cor(
  bexar_wide$Not.Confirmed,
  bexar_wide$Confirmed,
  use = "complete.obs"
)
## [1] 0.3724998

The correlation was approximately 0.37, indicating a positive association between not-confirmed and confirmed victim counts in the Bexar County observations. This correlation describes an association and does not establish that one variable causes the other.

Discussion

The analysis examined the relationship between not-confirmed and confirmed victim counts across 20 Bexar County program-year observations from fiscal years 2016 through 2025.

The Pearson correlation was 0.3725. This represents a positive association between the two variables in the observations examined. In other words, observations with higher not-confirmed victim counts also tended, on average, to have higher confirmed victim counts. However, the correlation is not especially close to 1, so the relationship is not a perfect linear relationship.

The descriptive statistics also show substantial differences in the two variables. The mean number of not-confirmed victims was 249.85, compared with a median of 191.5. Confirmed victims had a mean of 21.2 and a median of 17. The difference between the means and medians, particularly for not-confirmed victims, is consistent with the right-skew visible in the histograms.

These findings describe an association in the selected Bexar County observations. They do not establish that not-confirmed victim counts cause confirmed victim counts to increase or decrease.

Limitations

There are several limitations to this analysis. First, the dataset is aggregated at the county and program level rather than containing individual-level records. Therefore, the analysis describes patterns in the available county-level observations and should not be interpreted as representing individual children or individual investigations.

Because there are only 20 program-year observations, the results should be interpreted as exploratory rather than as a basis for broad generalizations.

Second, the analysis is observational. The correlation between not-confirmed and confirmed victim counts does not establish causation. Other factors may contribute to differences in victim counts across years and programs.

Finally, the analysis includes only Bexar County observations from fiscal years 2016 through 2025 and therefore should not automatically be generalized to other Texas counties or other time periods.

Conclusion

This exploratory analysis examined the relationship between not-confirmed and confirmed victim counts in Bexar County child abuse and neglect investigation data from fiscal years 2016 through 2025.

The analysis found a positive correlation of 0.3725 between the two variables. The descriptive statistics and graphs also showed differences in the distributions of not-confirmed and confirmed victim counts.

The results indicate a positive association between the two variables within the observations examined. However, the findings should be interpreted cautiously because the analysis uses aggregated, observational data and only 20 program-year observations. The results therefore describe the selected Bexar County observations rather than establishing a causal relationship or necessarily representing other counties or time periods.