Paper Description: This quantitative study uses a least-squares regression model to determine the predictive power of socio-economic factors on district-level student achievement on new, Common Core-aligned standardized assessments. We posit that educators may use our methodology and model to control for socio-economic factors, and more equitably compare school district performance.
Purpose
We used a least-squares regression model to determine the predictive power of socio-economic status (SES) on student achievement on Common Core-aligned standardized assessments in the State of Connecticut. The purpose of this case study is to: 1) Determine extent to which SES explains the differences …show more content…
To be included in the database, we determined that the Local Educational Agency (LEA) must be a Connecticut public school district for which we can collect SBA assessment data and SES data at both the grade level and the district level. We set the minimum number of test-takers at 60. We used a least-squares regression model to determine the predictive power of two socio-economic factors on student achievement. The outcome variable is the mean student scale score on the SBA, and the predictive variables are two measures of …show more content…
We fit both a straight and curvilinear model. The adjusted r-squared for each grade level and subject area represents the percentage of variance explained by the model. This approach allowed us to answer the primary research question:
To what extent does SES explain the differences between school district student achievement in the State of Connecticut on Common
Core-aligned assessments?
Results and Conclusions
We begin by reporting the descriptive statistics that show the percent of districts and students that are the subject of analysis.
[Insert Table 1]
We then present the measures of central tendency by grade for both subjects on the SBA.
[Insert Table 2]
The main research question dealt with determining the strength of the relationship between SES and student achievement on SBA. We first illustrate the correlation between SBA and FRPL by reporting the Pearson r values for all grade levels and both subject areas.
[Insert Table 3]
Next, we added the community EAR variable to the model.
[Insert Table 4]
In all cases, adding in the EAR variable increased the strength of the relationship at a statistically significant level. The predictive variables EAR and FRPL are statistically