196 - Childhood Exposure to Public Spending on Community-Based Resources During Childhood and Adverse and Positive Childhood Experiences

Publication Number: 196.4137

Jaime La Charite, University of California, Los Angeles David Geffen School of Medicine, Redondo Beach, CA, United States; Frederick Zimmerman, University of California, Los Angeles David Geffen School of Medicine, Los Angeles, CA, United States; Kristen Choi, UCLA, Los Angeles, CA, United States; Adam Schickedanz, UCLA Mattel Childrens Hospital, Los Angeles, CA, United States; Nicholas Jackson, University of California, Los Angeles David Geffen School of Medicine, Los Angeles, CA, United States; Mitchell D. Wong, University of California, Los Angeles David Geffen School of Medicine, Los Angeles, CA, United States; Keren Chen, University of California, Los Angeles David Geffen School of Medicine, Los Angeles, CA, United States; Rebecca N. Dudovitz, UCLA, Los Angeles, CA, United States

Background: Public community resources (education, libraries, parks and recreation, community development and housing) may enhance community wellbeing by reducing family stress and promoting resiliency. It is unknown whether such public spending is associated with adverse (ACEs) and positive childhood experiences (PCEs).

Objective: To examine associations between childhood exposure to local (combined city, county, school district) public spending on community resources and ACE and PCEs.

Design/Methods: We conducted a retrospective cohort study of data from the Panel Study of Income Dynamics (PSID, 2014-2021) linked to public spending data from the Fiscally Standardized Cities dataset (1977-2017) and city characteristics from the US Census. We constructed ACEs counts from two PSID samples: the 2014 Childhood Retrospective Circumstance Study (CRCS) and the 2017-2021 waves of the Transition into Adulthood Study (TAS). Average PCE scores were constructed with CRCS data only (not available in TAS). The independent variable is a city-level per capita sum of city, county, and school district operational spending on community resources. We linked datasets using the childhood city of residence and year when the PSID respondent was age nine or closest available age – limiting to childhood cities with 150,000 residents or more. We conducted weighted Poisson and linear regressions, adjusting for demographics, childhood city characteristics, and remaining public spending categories (e.g., hospitals).

Results: There were 1,009 adult CRCS respondents (mean age 36, SD 9 yrs) and 1,384 adult TAS respondents (mean age 25, SD 4 yrs). In both samples, about 80% reported one or more ACEs and an average PCE score of 3.11 (SD 0.52, max possible 4.0) in the CRCS sample. There were no significant associations between childhood community resource spending and ACEs in either sample. In contrast, every 10% increase in per-capita community resource spending was associated with a 0.06 standard deviation decrease in PCE score (95% CI: -0.12, -0.01). Post hoc stratified analyses suggest this negative association is primarily among respondents born 1960-1979 (aged 34-53) and/or with childhood household income above the median.

Conclusion(s): Overall greater operational public spending on community resources is not associated with ACEs, but– contrary to our hypothesis – is associated with lower PCEs, which may differ by cohort and household income. Further research is needed to understand this unexpected relationship and explore how investments in specific public programs, rather than overall spending, may be leveraged to support life course health.

Table 1. Characteristics of Eligible Respondents in CRCS (N=1,009) and TAS (N=1,384) Study Samples
CRCS = Childhood Retrospective Circumstance Study. TAS = Transition to Adulthood Supplement. FPL = federal poverty limit; m = mean; SD = standard deviation; ACE = adverse childhood experience; PCE = positive childhood experience

(a) CRCS Sample drawn from Childhood Retrospective Circumstance Study supplemental survey fielded to a subsample of the Panel Study of Income Dynamics English speaking panel members in 2014.

(b) TAS Sample drawn from the Transition to Adulthood supplemental surveys fielded to young adult (aged 18-28 years) panel members of the Panel Study of Income Dynamics in 2017, 2019, and 2021.

(c) Dollars adjusted to be comparable to 2021 dollars

(d) Community Resource Spending included operational spending per capita on public primary and secondary education, libraries, parks and recreation, community development and housing sectors.

(e) All Other Sector Spending included the residual operational spending per capita after removing the operational spending on community resource sectors. This other sector spending variable included operational spending on other welfare, higher education, hospitals, public health, transportation, public safety, natural resources, sewage, solid waste management, governmental administration, interest on general debt, utilities, liquor store, employee retirement trust expenditures sectors.

Table 2. Adjusted Incident Rate Ratios and Beta Coefficients for Poisson and Linear Multivariate Regression Models of ACE Score and PCE Score on FiSC Operational Expenditures Per Capita on Community Resource Spending
FiSC = fiscally standardized cities; aIRR = adjusted incident rate ratio; CI = confidence interval; FPL = federal poverty limit

Regression models were adjusted for the covariates listed in Table 2 and survey weighted using PSID-CRCS cross-sectional survey weights or PSID average cross-sectional weight from 2017, 2019, and 2021 PSID surveys.

(a) TAS Sample drawn from the Transition to Adulthood supplemental surveys fielded to young adult (aged 18-28 years) panel members of the Panel Study of Income Dynamics in 2017, 2019, and 2021.

(b) CRCS Sample drawn from Childhood Retrospective Circumstance Study supplemental survey fielded to a subsample of the Panel Study of Income Dynamics English speaking panel members in 2014.

(c) Community Resource Spending included operational spending per capita on public primary and secondary education, libraries, parks and recreation, community development and housing sectors.

(d) All Other Sector Spending included the residual operational spending per capita after removing the operational spending on family targeted community resource sectors. This other sector spending variable included operational spending on other welfare, higher education, hospitals, public health, transportation, public safety, natural resources, sewage, solid waste management, governmental administration, interest on general debt, utilities, liquor store, employee retirement trust expenditures sectors.

Table 3. Adjusted Beta Coefficients for Linear Multivariate Regression Models of PCE Scores on FiSC Operational Expenditures on Community Resource Spending, Stratified by Childhood Household Income, Age, and Presence of a Move During Childhood for CRCS Sample
(a) All shown analyses adjusted for the following covariates (parent highest parental educational attainment, childhood household income, childhood move, childhood city population size, city demographics [% less than 18 years, poverty, unemployed, non-Hispanic White, foreign born], population density, median rent, age in 2013, sex, race/ethnicity.
(b) Regression models were survey weighted using PSID-CRCS cross-sectional survey weight.

Table 1. Characteristics of Eligible Respondents in CRCS (N=1,009) and TAS (N=1,384) Study Samples
CRCS = Childhood Retrospective Circumstance Study. TAS = Transition to Adulthood Supplement. FPL = federal poverty limit; m = mean; SD = standard deviation; ACE = adverse childhood experience; PCE = positive childhood experience

(a) CRCS Sample drawn from Childhood Retrospective Circumstance Study supplemental survey fielded to a subsample of the Panel Study of Income Dynamics English speaking panel members in 2014.

(b) TAS Sample drawn from the Transition to Adulthood supplemental surveys fielded to young adult (aged 18-28 years) panel members of the Panel Study of Income Dynamics in 2017, 2019, and 2021.

(c) Dollars adjusted to be comparable to 2021 dollars

(d) Community Resource Spending included operational spending per capita on public primary and secondary education, libraries, parks and recreation, community development and housing sectors.

(e) All Other Sector Spending included the residual operational spending per capita after removing the operational spending on community resource sectors. This other sector spending variable included operational spending on other welfare, higher education, hospitals, public health, transportation, public safety, natural resources, sewage, solid waste management, governmental administration, interest on general debt, utilities, liquor store, employee retirement trust expenditures sectors.

Table 2. Adjusted Incident Rate Ratios and Beta Coefficients for Poisson and Linear Multivariate Regression Models of ACE Score and PCE Score on FiSC Operational Expenditures Per Capita on Community Resource Spending
FiSC = fiscally standardized cities; aIRR = adjusted incident rate ratio; CI = confidence interval; FPL = federal poverty limit

Regression models were adjusted for the covariates listed in Table 2 and survey weighted using PSID-CRCS cross-sectional survey weights or PSID average cross-sectional weight from 2017, 2019, and 2021 PSID surveys.

(a) TAS Sample drawn from the Transition to Adulthood supplemental surveys fielded to young adult (aged 18-28 years) panel members of the Panel Study of Income Dynamics in 2017, 2019, and 2021.

(b) CRCS Sample drawn from Childhood Retrospective Circumstance Study supplemental survey fielded to a subsample of the Panel Study of Income Dynamics English speaking panel members in 2014.

(c) Community Resource Spending included operational spending per capita on public primary and secondary education, libraries, parks and recreation, community development and housing sectors.

(d) All Other Sector Spending included the residual operational spending per capita after removing the operational spending on family targeted community resource sectors. This other sector spending variable included operational spending on other welfare, higher education, hospitals, public health, transportation, public safety, natural resources, sewage, solid waste management, governmental administration, interest on general debt, utilities, liquor store, employee retirement trust expenditures sectors.

Table 3. Adjusted Beta Coefficients for Linear Multivariate Regression Models of PCE Scores on FiSC Operational Expenditures on Community Resource Spending, Stratified by Childhood Household Income, Age, and Presence of a Move During Childhood for CRCS Sample
(a) All shown analyses adjusted for the following covariates (parent highest parental educational attainment, childhood household income, childhood move, childhood city population size, city demographics [% less than 18 years, poverty, unemployed, non-Hispanic White, foreign born], population density, median rent, age in 2013, sex, race/ethnicity.
(b) Regression models were survey weighted using PSID-CRCS cross-sectional survey weight.