Primary
Middle
High
Overall
Analysis of the data:
The regression model predicted district accountability indices using percent free and reduced meal rate (a proxy for composite SES status of the student population). Mealrate accounted for nearly 60% (r2 = .5788) of the variance in the dependent variable, a robust relationship by any standard.
SES is the single strongest predictor of achievement among students.
Moreover, it correlates closely other kinds of social stressors like mobility, parental education, etc.
Note, prior to performing the regression analysis, the distribution of values for the variables of interest was reviewed using descriptive statistics (i.e., skewness and kurtosis), histograms, and Q-Q plots.
Results indicated a normal distribution for both variables.
Moreover, the presence of speciously influential observations, or "outliers," was also investigated. Specifically, an OLS regression was performed in order to compute two statistics commonly used for the identification of outliers: Cook's distance and Mahalanobis distance. No outliers were identified. Additionally, and following the method reported by Cronbach (1987), both variables of interest were centered in order to reduce the collinearity of related independent variables.
and why regression analysis was deemed appropriate for this kind of study.
Regression analysis is the standard approach for investigating the relationship between one or more independent variables and a continuous dependent variable (see, e.g., Best & Hahn, 2006; Harris, 1998; Myrtler & Charles, 2005; Vogt, 2007).
It would seem to me that the fundamental premise is flawed if you are comparing the SES subgroup mean score to the district's score since the higher the percentage of disadvantaged students the closer that mean score will be to the district mean score or if the overall district score is low the less of a gap (i.e., overachieving, underachieving designation).
The SES subgroup mean score was not used in the analysis. The achievement measure was the overall district accountability index.
Was distance from the index goal of 100 factored in to the formula?
No.
Was the percentage of students performing at or above the proficient level?
No.
Was any other assessment other than CATS data used?
No.
