ABSTRACT
We explored factors associated with school approaches to identify and support student social, emotional, and behavioral needs. Hypothesized mediators of the relationship between district demographic characteristics and district academic and behavioral outcomes included district administrator perceptions of problems; use of a universal behavioral support program; primary approach to identifying and supporting student needs; and perceived usability of that approach. We found that district demographic characteristics were highly associated with nearly every other indicator, with potential that district leader knowledge and beliefs could serve as a lever to engaging in more preventive approaches.
Acknowledgments
Special thanks are provided to all survey participants for their time, and appreciation is extended to research team members for their assistance on the overall project and specifically Dan Volk for his work on this manuscript.
Disclosure Statement
No potential conflict of interest was reported by the author(s).
Notes
1. We evaluated the measurement invariance of the measurement model and invariance of the regression coefficients in the structural model based on the district administrator role (superintendent vs non-superintendent; where non-superintendent included roles for director of pupil services, director of special education, director of curriculum, director of assessment and accountability, and other/unspecified). We found that scalar invariance in the measurement model was tenable and that the model fit (and regression coefficients) of the structural model did not change in a multiple-group structural equation model.
2. The SEDA data is from when students were tested in the spring of the 2015–2016 academic year.
3. We used Bayesian and Monte Carlo Integration estimation in Mplus to fit each model described herein treating BAP and BPP as categorical. In both cases, the point estimates for the parameters were nearly identical to the point estimates for the parameters using ML estimation and treating BAP and BPP as continuous. Given that Bayesian and Monte Carlo analyses with categorical indicators do not provide usable measures of model fit for model comparison (e.g., the deviance information criterion is not available in Mplus using Bayesian estimation with categorical indicators), and that the point estimates for the parameters were largely unchanged, we decided to use the results from ML estimation with BAP and BPP treated as continuous. For discussion on the availability of the deviance information criterion for Bayesian estimation with categorical indicators in Mplus, see, Muthén and Muthén (Citation2010).
4. The scale scores for Family-School Collaboration and External Support were dropped from the DURP factor because of low factor loadings in a measurement model for the DURP factor.
5. Due to data privacy, EDFacts sometimes reports graduation rates with a range. In these case, the lowest number of the range was chosen.