Wednesday, November 26, 2014

Factors Influencing Student Outcomes and How They Interact - 15 - Family Income

In response to the feedback received on "Educational Funding and Student Outcomes: The Relationship as Evidenced by State-Level Data," the KASB Research Department is working on a "Part II" which will dig further into other factors that impact student outcomes, and how funding impacts when these other factors are taken into consideration.

In this series of blog posts, I will describe the preliminary correlation analysis comparing these factors (independent variables) with each other in an attempt to show how closely tied to each other they are.  

Today's topic is Family Income

The Bureau of Economic Analysis provides an average per capita personal income by year and state.  The US Census Bureau provides a median household income by year and state.  Both measures were used to indicate family income; as actual dollar amounts, amounts adjusted for inflation over time using the Bureau of Labor Statistics' Consumer Price Index, and amounts adjusted for state cost of living using the Bureau of Economic Analysis' Regional Price Parity calculation.

In terms of other independent variables, I will list the interactions in terms of the strength of the highest correlation between variables observed; Strong (+/- 1.0 to 0.5), Moderate (+/- 0.5 to 0.3), Weak (+/- .03 to 0.1), and None or Very Weak (+/- 0.1 to 0.0).

Strong

  • There is a strong positive correlation between RPP and personal and household income.  This correlation remains strong when adjusted for inflation (CPI2014), and becomes weak when adjusted for RPP; which means that even when the dollar amounts are adjusted for state cost of living, the relationship remains statistically significant between RPP and personal and household income.
  • There is a moderate to strong negative correlation between percent of students eligible for free or reduced-price lunch and both personal and household income.  This holds true when income is adjusted for inflation (CPI2014) and state cost of living (RPP).  
  • There is a moderate to strong positive correlation between teacher salaries and both per capita and household income; indicating that states with higher teacher salaries also have higher per capita and household incomes.
  • There is a weak to strong positive correlation between school spending and per capita or household income; indicating states that spend more per pupil have higher average individual and family income.  
  • There is a moderate to strong positive correlation between educational attainment and per capita and household income; indicating states with higher education levels also have higher income levels.
  • There is a moderate to strong negative correlation between poverty and personal or family income; with higher percents of poverty tied to lower income.  This holds true when the income amounts are adjusted both for inflation (CPI2014) and for regional cost of living (RPP).
  • There is a weak to strong positive correlation between population per square mile and family income; indicating that incomes are lower in less densely populated states. 

Moderate

  • There is a moderate positive correlation between period and per capita personal income; indicating that individual incomes are increasing over time.  This relationship becomes non-significant when controlled for inflation (CPI2014), but becomes a strong positive correlation when adjusted for cost of living by state (RPP). There is a weak positive correlation between period and household income; indicating that family incomes are not increasing as noticeably over time as personal incomes.  This relationship becomes a weak negative relationship when controlled for inflation (CPI2014); implying that family incomes are actually decreasing over time, and increases slightly when adjusted for cost of living by state (CPI2014). 
  • There is a moderate negative correlation between percent taking the ACT and personal or household income, and a moderate positive correlation between percent taking the SAT and personal or household income.  This is consistent when adjusting for inflation (CPI2014) and cost of living (RPP); meaning students from states with lower average income are more likely to take the ACT and less likely to take the SAT. 
  • There is a weak negative to no correlation between percent of white students and personal or household income, a weak to moderate negative correlation between percent of black students and personal or household income, a weak positive to no correlation between percent of Hispanic students and personal or household income, and no correlation between percent of American Indian or Alaska native and personal or household income when looking at actual dollar amounts and amounts adjusted for inflation (CPI2014).  When adjusting for state cost-of-living (RPP), there is a weak positive correlation between the percent of white students and personal or household income, a weak to moderate negative correlation between percent of black students and personal or household income, and no correlation between the percent of Hispanic and American Indian or Alaska native students and personal or household income.  This suggests that, beyond regional cost differences, states with higher percents of white students and lower percents of black students have higher personal or household incomes.
  • There is a weak to moderate negative correlation between student teacher ratio and per capita personal income but no correlation between student teacher ratio and median household income; suggesting states with more students per teacher are also likely to have lower per capita personal income.  

Weak

  • There is no correlation between percent of students served under IDEA and the median household income (actual, CPI2014, and RPP), but there is a weak positive correlation between the percent of special education students and the per capita personal income (CPI2014 and RPP, but not actual); indicating that states with higher percents of special education students have higher per capita personal incomes when adjusted for inflation or cost of living.
  • There is no correlation between percent of ELL students and per capita personal income when looking at actual dollars, a weak positive correlation when looking at personal income adjusted for inflation (CPI2014), and a weak negative correlation when looking at personal income adjusted for state cost-of-living (RPP).  There is a weak positive correlation between between percent of ELL students and household income when looking at actual dollars and dollars adjusted for inflation (CPI2014), and no correlation when looking at household income adjusted for state cost-of-living (RPP).  This suggests that states with higher percents of ELL students have higher incomes except when taking state cost-of-living into account, though this relationship is weak.
  • There is a weak positive correlation between instruction as a percent of current spending and per capita or household income, and no correlation between instruction as a percent of total revenue and per capita or household income; indicating that states spending more on expenses categorized as "instruction" also have higher income levels. 

None or Very Weak

  • There were no correlations related to educational attainment yielding very weak or non-significant results.

Tuesday, November 25, 2014

Factors Influencing Student Outcomes and How They Interact - 14 - Population Density

In response to the feedback received on "Educational Funding and Student Outcomes: The Relationship as Evidenced by State-Level Data," the KASB Research Department is working on a "Part II" which will dig further into other factors that impact student outcomes, and how funding impacts when these other factors are taken into consideration.

In this series of blog posts, I will describe the preliminary correlation analysis comparing these factors (independent variables) with each other in an attempt to show how closely tied to each other they are.  

Today's topic is Population Density

The US Census bureau provides information on the population per square mile by state and year.  This is used as a measure of overall population density for each state.

In terms of other independent variables, I will list the interactions in terms of the strength of the highest correlation between variables observed; Strong (+/- 1.0 to 0.5), Moderate (+/- 0.5 to 0.3), Weak (+/- .03 to 0.1), and None or Very Weak (+/- 0.1 to 0.0).

Strong

  • There is a strong positive correlation between RPP and population per square mile; indicating states with more people per mile have higher cost of living.
  • There is a moderate negative correlation between percent taking the ACT and population per square mile, and a strong positive correlation between percent taking the SAT and population per square mile; indicating that students from less densely populated states are more likely to take the ACT and less likely to take the SAT. 
  • There is a strong positive correlation between teacher salary and population per square mile; indicating states with higher teacher salaries are more densely populated.
  • There is a moderate to strong correlation between school spending and population per square mile; indicating states that spend more per pupil are more densely populated.
  • There is a weak to strong positive correlation between population per square mile and family income; indicating that incomes are lower in less densely populated states. 

Moderate

  • There is a moderate positive correlation between percent of students served under IDEA and population per square mile; indicating that more densely populated states have higher percentages of special education students.  
  • There is a weak to moderate negative correlation between poverty and population per square mile; with higher percents of poverty tied to lower population density.  This implies that there is more poverty in less densely populated states.

Weak

  • There is a weak negative correlation between percent of students eligible for free or reduced-price lunch and population per square mile; indicating that states with lower percents of at-risk students have higher numbers of persons per square mile. 
  • There is a weak negative correlation between population per square mile and percent of white and American Indian or Alaska native students, a weak positive correlation between population per square mile and percent of black students, and no correlation between population per square mile and Hispanic students.  This indicates states with lower percents of white and American Indian or Alaska native students and higher percents of black students have more people per square mile. 
  • There is a weak negative correlation between student teacher ratio and persons per square mile; indicating that state with more student per teacher also tend to be less densely populated.
  • There is a weak positive to no correlation between population density and the percent of the population with at least a high school diploma, and a strong positive correlation between percent of 25 year olds and up with college degrees and population per square mile; indicating more densely populated states tend to have higher percents of the population with bachelors and graduate degrees. 

None or Very Weak

  • There is no correlation between school year (period) and population density.
  • There is no correlation between percent of students participating in programs for English Language Learners and population density.
  • There is no correlation between spending on instruction as a percent and population density.

Monday, November 24, 2014

Factors Influencing Student Outcomes and How They Interact - 13 - Poverty

In response to the feedback received on "Educational Funding and Student Outcomes: The Relationship as Evidenced by State-Level Data," the KASB Research Department is working on a "Part II" which will dig further into other factors that impact student outcomes, and how funding impacts when these other factors are taken into consideration.

In this series of blog posts, I will describe the preliminary correlation analysis comparing these factors (independent variables) with each other in an attempt to show how closely tied to each other they are.  

Today's topic is Poverty

This study reviewed six measures of poverty; five measures of the percent of children at certain levels of poverty (50%, 100%, 150%, 200%, and 250%) as reported by the Anne E. Casey foundation in their annual Kids Count Data Book (based on U.S. Census data), and the percent of people with income below the poverty level in the past twelve months as reported by the U.S. Census bureau.  

In terms of other independent variables, I will list the interactions in terms of the strength of the highest correlation between variables observed; Strong (+/- 1.0 to 0.5), Moderate (+/- 0.5 to 0.3), Weak (+/- .03 to 0.1), and None or Very Weak (+/- 0.1 to 0.0).

Strong

  • There is a strong positive correlation between the percent in poverty and the percent eligible for free or reduced-price lunch, as would be expected.  
  • There is a moderate to strong negative correlation between poverty and teacher salaries; with higher percents of poverty tied to lower teacher salaries.  This does not hold true, however, on the teacher salary amounts adjusted for regional cost of living (RPP); suggesting both poverty and teacher salaries are related to regional cost of living but not necessarily directly related to each other. 
  • There is a moderate to strong negative correlation between poverty and school spending; with higher percents of poverty tied to lower school spending.  This holds true even with the funding amounts adjusted for regional cost of living (RPP); suggesting poverty and school funding may be directly related to each other.  
  • There is a moderate to strong negative correlation between poverty and educational attainment; with higher percents of poverty tied to lower educational attainment.  
  • There is a moderate to strong negative correlation between poverty and personal or family income; with higher percents of poverty tied to lower income.  This holds true when the income amounts are adjusted both for inflation (CPI2014) and for regional cost of living (RPP).

Moderate

  • Poverty has a moderate negative correlation with RPP (regional price parity, or cost of living by state); suggesting that as the cost of living increases, the percent in poverty decreases.  
  • There is a weak positive correlation between the percent of students who are Hispanic and the percent in poverty, though this correlation is not as strong as that between the percent of students who are black and the percent in poverty.  The percent of students who are white is negatively correlated with the percent in poverty, but again this is a moderate correlation.  There is no apparent correlation between the percent of students who are American Indian or Alaska Native and the percent in poverty, though it is important to note that this population is very small in many states.  
  • There is a moderate negative correlation between poverty and spending on instruction as a percent of current spending per pupil; indicating that areas with higher poverty are putting a smaller percent of current spending towards instruction than those with lower poverty.  There is no significant relationship between poverty and instruction as a percent of total revenue per pupil. 
  • There is a weak to moderate negative correlation between poverty and population per square mile; with higher percents of poverty tied to lower population density.  This implies that there is more poverty in less densely populated states.

Weak

  • Poverty has a weak positive correlation with period (or year); suggesting that the percent of the population in poverty is increasing over time.  
  • There is a weak negative correlation between the percent in poverty and the percent of children served under IDEA, which is somewhat counter-intuitive (unless the availability of services is perhaps a significant factor). 
  • There is a weak positive correlation between poverty and the student-teacher ratio; with higher percents of poverty tied to higher numbers of students per teacher, but this relationship is not statistically significant across all poverty measures.

None or Very Weak

  • There is no significant correlation between the percent in poverty and the percent participating in programs for English Language Learners (ELL), which is contrary to the argument that students from homes where English is not the primary language spoken are more likely to also come from lower income homes.  

Thursday, November 20, 2014

Factors Influencing Student Outcomes and How They Interact - 12 - Educational Attainment

In response to the feedback received on "Educational Funding and Student Outcomes: The Relationship as Evidenced by State-Level Data," the KASB Research Department is working on a "Part II" which will dig further into other factors that impact student outcomes, and how funding impacts when these other factors are taken into consideration.

In this series of blog posts, I will describe the preliminary correlation analysis comparing these factors (independent variables) with each other in an attempt to show how closely tied to each other they are.  

Today's topic is Educational Attainment

The US Census Bureau provides statistics on the percent of 18 to 24 year olds in a state with at least a high-school diploma, and also the percent of 25 year olds and up with at least a high school diploma, a bachelors degree and up, and a graduate degree and up.  These provide a good set of measures indicating the educational attainment within a state.  The measure of 18 to 24 year olds with at least a high school diploma is also included as a dependent variable; or student outcome, as it can also represent an extended graduation rate for the state.

In terms of other independent variables, I will list the interactions in terms of the strength of the highest correlation between variables observed; Strong (+/- 1.0 to 0.5), Moderate (+/- 0.5 to 0.3), Weak (+/- .03 to 0.1), and None or Very Weak (+/- 0.1 to 0.0).

Strong

  • There is a weak positive correlation between RPP and the percent of 18-24 year olds and 25 year olds and older with at least a high school diploma, but a strong positive correlation between RPP and the percent of 25 year olds and older with a bachelors, graduate degree, or higher.  This indicates that states with higher cost of living are also more likely to have citizens with some form of college degree.  
  • There is no correlation between the percent taking the ACT and the percent with at least a high school diploma, but there is a moderate negative correlation between the percent taking the ACT and the percent with at least a bachelors or at least a graduate degree; indicating that students from states with a higher percent of college degree earners are less likely to take the ACT.  There is no correlation to a weak positive correlation between the percent taking the SAT and the percent with at least a high school diploma, and a strong positive correlation between the percent taking the SAT and the percent with at least a bachelors or at least a graduate degree; indicating that students from states with a higher percent of college degree earners are more likely to take the SAT. 
  • There is a moderate to strong negative correlation between percent of students eligible for free or reduced-price lunch and educational attainment in the state; indicating that states with lower percents of at-risk students have higher percent of people achieving high school diplomas and college degrees.  
  • There is a moderate positive correlation between percent of white students and percent of the population with at least a high school diploma, a moderate to strong negative correlation between the percent of black and Hispanic students and the percent of the population with at least a high school diploma, and a weak negative correlation between percent of 18-24 year olds with at least a high school diploma and the percent of American Indian or Alaska native students, but a weak positive correlation between percent of 25 year olds and older with at least a high school diploma.  There is a weak negative correlation between percent of black students and percent of people with at least a bachelor's degree, a weak positive correlation between percent of Hispanic students and percent of people with at least a bachelor's degree, and no correlation between percent of white and American Indian or Alaska native students and percent of people with at least a bachelor's degree.  There is a weak negative correlation between the percent of white and American Indian or Alaska native students and the percent of people with a graduate degree, a weak correlation between the percent of Hispanic students and percent of people with a gradate degree, and no correlation between percent of black students and percent of population with a graduate degree.  
  • There is a moderate positive correlation between teacher salaries and percent of the population 18-14 with at least a high school diploma, a moderate to strong positive correlation between teacher salaries and percent of the population 25 and up with at least a bachelor's degree, a strong positive correlation between teacher salaries and percent of the population 25 and up with a graduate degree, and no correlation between teacher salaries and percent of the population 25 and up with at least a high school diploma; indicating that states with higher teacher salaries also tend to have a more educated populace.  
  • There is a weak to strong positive correlation between school spending and educational attainment; indicating states that spend more per pupil also have more highly educated populations.
  • There is a moderate to strong negative correlation between poverty and educational attainment; with higher percents of poverty tied to lower educational attainment.  
  • There is a moderate to strong positive correlation between educational attainment and per capita and household income; indicating states with higher education levels also have higher income levels.

Moderate

  • There is a weak to moderate positive correlation between period and educational attainment; indicating that as time goes on a higher percent of the population is earning high school diplomas, bachelors degrees, and graduate degrees.
  • There is a weak to moderate negative correlation between ELL program participation and the percent of the population with at least a high school diploma, and a weak positive correlation between ELL program participation and percent of the population with a college degree; indicating that states with higher percents of ELL students have lower percents of high school completers but higher percents of college graduates.  
  • There is a moderate negative relationship between the number of students per teacher and percent of 18 to 24 year olds with at least a high school diploma and no correlation between the student teacher ratio and all other measures of educational attainment.  This indicates that states with fewer students per teacher have more residents ages 18 to 24 who have graduated from high school.

Weak

  • There is a weak positive correlation between percent of students served under IDEA and the percent of the population with at least a high school diploma, but no correlation between percent of student served under IDEA and percent of the population with a college degree; indicating that states with higher percents of special education students also have higher percents of people with at least a high school diploma.  
  • There is a weak positive to no correlation between spending on instruction as a percent and educational attainment; suggesting that states putting more spending towards expenses categorized as "instruction" also tend to have a more educated population.  
  • There is a weak positive to no correlation between population density and the percent of the population with at least a high school diploma, and a strong positive correlation between percent of 25 year olds and up with college degrees and population per square mile; indicating more densely populated states tend to have higher percents of the population with bachelors and graduate degrees. 

None or Very Weak

  • There were no correlations related to educational attainment yielding very weak or non-significant results.

Wednesday, November 19, 2014

Factors Influencing Student Outcomes and How They Interact - 11 - Instruction as a Percent of Spending

In response to the feedback received on "Educational Funding and Student Outcomes: The Relationship as Evidenced by State-Level Data," the KASB Research Department is working on a "Part II" which will dig further into other factors that impact student outcomes, and how funding impacts when these other factors are taken into consideration.

In this series of blog posts, I will describe the preliminary correlation analysis comparing these factors (independent variables) with each other in an attempt to show how closely tied to each other they are.  

Today's topic is Instruction as a Percent of Current Spending or Total Revenue

The US Census Bureau provides information on per student total revenue, current spending, and spending on instruction by state and year.  Looking at spending on instruction as a percent of either current spending or total revenue is an attempt to parse out the amount of funding going directly towards classroom instruction and what is going to school, district, and state administration.  

In terms of other independent variables, I will list the interactions in terms of the strength of the highest correlation between variables observed; Strong (+/- 1.0 to 0.5), Moderate (+/- 0.5 to 0.3), Weak (+/- .03 to 0.1), and None or Very Weak (+/- 0.1 to 0.0).

Strong

  • There were no comparisons to school spending for which the strongest correlation was above 0.5.

Moderate

  • There is a weak positive correlation between percent of white students and spending on instruction as a percent of current spending, and a moderate negative correlation between spending on instruction as a percent of current spending and percent of American Indian or Alaska native students, but no correlation between spending on instruction and percent of black or Hispanic students.  There is a weak positive correlation between percent of white students and spending on instruction as a percent of total revenue, weak positive correlations between percent of Hispanic and American Indian or Alaska native students and spending on instruction as a percent of total revenue, and no correlation between the percent of black students and spending on instruction as a percent of total revenue.  This suggests that states with higher percents of white students and lower percents of American Indian or Alaska native students allocate more funds to expenses classified as "instruction," and when looking at total spending, states with higher percents of Hispanic students put more of their total funding amounts towards "instruction."  
  • There is  a moderate positive correlation to no correlation between instruction as a percent of current spending or total revenue and school spending in terms of dollars; indicating that states with higher per pupil spending also tend to put a higher percent of spending towards expenses categorized as "instruction."  
  • There is a moderate negative correlation between poverty and spending on instruction as a percent of current spending per pupil; indicating that areas with higher poverty are putting a smaller percent of current spending towards instruction than those with lower poverty.  There is no significant relationship between poverty and instruction as a percent of total revenue per pupil. 

Weak

  • There is a weak positive correlation between RPP and instruction as a percent of current spending; indicating that states with higher cost of living spend a greater percent of current spending on expenses categorized under "instruction."  There is no correlation between RPP and instruction as a percent of total revenue. 
  • There is a weak negative correlation between the percent taking the ACT and spending on instruction as a percent of current spending, and a weak positive correlation between the percent taking the SAT and spending on instruction as a percent of current spending; meaning students coming from states where more of the overall spending goes to expenses classified as "instruction" are less likely to take the ACT and more likely to take the SAT.  There is no correlation between the percent of students taking the ACT and spending on instruction as a percent of total revenue, and a weak positive correlation between percent taking the SAT and spending on instruction as a percent of total revenue. 
  • There is no correlation between percent of students served under IDEA and spending on instruction as a percent of current spending, and a weak positive correlation between percent of students served under IDEA and spending on instruction as a percent of total spending; indicating states with higher percents of special education students put more of total spending towards instruction.
  • There is a weak negative correlation between percent of students eligible for free or reduced-price lunch and spending on instruction as a percent of current spending, and no correlation between percent of students eligible for free and reduced-price lunch and spending on instruction as a percent of total revenue; indicating that states with higher percents of at-risk students have lower percents of current spending  going to expenses categorized as "instruction."  
  • There is no correlation between ELL program participation and spending on instruction as a percent of current spending and a weak negative correlation between ELL program participation and spending on instruction as a percent of total revenue; suggesting states with higher percents of ELL students allocate smaller percents of spending towards expenses categorized as "instruction."
  • There is a weak negative correlation between student-teacher ratio and instruction spending as a percent of both current spending and total revenue; indicating that states with more students per teacher allocate less to expenses classified as "instruction." 
  • There is a weak positive to no correlation between spending on instruction as a percent and educational attainment; suggesting that states putting more spending towards expenses categorized as "instruction" also tend to have a more educated population.  
  • There is a weak positive correlation between instruction as a percent of current spending and per capita or household income, and no correlation between instruction as a percent of total revenue and per capita or household income; indicating that states spending more on expenses categorized as "instruction" also have higher income levels. 

None or Very Weak

  • There is no correlation between period and percent spent on instruction.  
  • There is no correlation between teacher salaries and instruction as a percent of either current spending or total spending.  
  • There is no correlation between spending on instruction as a percent and population density.

Tuesday, November 18, 2014

Factors Influencing Student Outcomes and How They Interact - 10 - School Spending

In response to the feedback received on "Educational Funding and Student Outcomes: The Relationship as Evidenced by State-Level Data," the KASB Research Department is working on a "Part II" which will dig further into other factors that impact student outcomes, and how funding impacts when these other factors are taken into consideration.

In this series of blog posts, I will describe the preliminary correlation analysis comparing these factors (independent variables) with each other in an attempt to show how closely tied to each other they are.  

Today's topic is School Spending

The US Census Bureau provides information on per student total revenue, current spending, and spending on instruction by state and year.  These values are included in the study as actual dollar amounts, amounts adjusted for inflation over time using the Bureau of Labor Statistics' Consumer Price Index, and amounts adjusted for state cost of living using the Bureau of Economic Analysis' Regional Price Parity calculation.

In terms of other independent variables, I will list the interactions in terms of the strength of the highest correlation between variables observed; Strong (+/- 1.0 to 0.5), Moderate (+/- 0.5 to 0.3), Weak (+/- .03 to 0.1), and None or Very Weak (+/- 0.1 to 0.0).

Strong

  • There is a strong positive correlation between RPP and school spending; indicating that states with higher cost of living also spend more money per pupil than those with lower cost of living.  This relationship remains strong when adjusted for inflation (CPI2014), and becomes moderate when adjusted for RPP; which means that even when the dollar amounts are adjusted for state cost of living, the relationship remains statistically significant between RPP and per pupil spending.
  • There is a moderate negative relationship between the percent of graduates taking the ACT exam and per pupil spending, and a moderate to strong positive relationship between the percent of students taking the SAT and per pupil spending; meaning students from states spending less on education are more likely to take the ACT and less likely to take the SAT.  This relationship remains strong when adjusted for inflation (CPI2014), but becomes moderate when adjusted for state cost of living (RPP).  
  • There is a moderate to strong negative correlation between student-teacher ratio and all measures of school spending; indicating that states with higher per pupil spending also have fewer students per teacher.  
  • There is a strong positive correlation between teacher salaries and school spending; indicating that states that spend more per pupil also have higher teacher salaries.
  • There is a weak to strong positive correlation between school spending and educational attainment; indicating states that spend more per pupil also have more highly educated populations.
  • There is a moderate to strong correlation between school spending and population per square mile; indicating states that spend more per pupil are more densely populated.
  • There is a weak to strong positive correlation between school spending and per capita or household income; indicating states that spend more per pupil have higher average individual and family income.  

Moderate

  • There is a moderate positive correlation between percent of students served under IDEA and education funding with and without adjustments for inflation (CPI2014) and state cost of living (RPP); indicating that states with higher percents of special education students spend more on average per pupil.  
  • There is a moderate negative correlation between percent of students eligible for free and reduced-price lunch and education spending.  This is true for actual dollars and dollars adjusted for inflation (CPI2014) and state cost of living (RPP); suggesting that states with more at-risk students (largely due to poverty) also have lower per pupil spending.  
  • There is a weak negative correlation between percent of students participating in ELL programs and education spending when looking at actual dollar amounts and inflation-adjusted amounts, and a moderate negative correlation when looking at state cost-of-living-adjusted amounts; suggesting that states with higher percents of ELL students have lower per pupil spending, and this trend is even more apparent when accounting for regional cost differences.
  • There is  a moderate positive correlation to no correlation between instruction as a percent of current spending or total revenue and school spending in terms of dollars; indicating that states with higher per pupil spending also tend to put a higher percent of spending towards expenses categorized as "instruction."  

Weak

  • There is a weak positive correlation between period and school spending; indicating that as time passes, more money per pupil is being spent on education.  This relationship becomes very weak to weak when adjusted for inflation (CPI2014), but becomes weak to moderate when adjusted for regional cost of living (RPP). This suggests that overall education spending per pupil may be increasing nationwide at a rate slightly higher than inflation.  
  • There is no correlation between school spending and percent of white, black, or American Indian or Alaska native students and a weak negative correlation for some measures of school spending and percent of Hispanic students when looking at actual dollars and dollars adjusted for inflation (CPI2014).  When adjusting for state cost-of-living (RPP), there is a weak positive correlation between the percent of white students and school spending, a weak negative correlation between the percent of Hispanic students and school spending, and no correlation between school spending and the percent of black and American Indian or Alaska native students.  This suggests that beyond regional cost differences, states with higher percents of white students and lower percents of Hispanic students have higher per pupil education spending.  

None or Very Weak

  • There were no comparisons to school spending for which the strongest correlation was very weak or non-significant.

Monday, November 17, 2014

Factors Influencing Student Outcomes and How They Interact - 09 - Teacher Salaries

In response to the feedback received on "Educational Funding and Student Outcomes: The Relationship as Evidenced by State-Level Data," the KASB Research Department is working on a "Part II" which will dig further into other factors that impact student outcomes, and how funding impacts when these other factors are taken into consideration.

In this series of blog posts, I will describe the preliminary correlation analysis comparing these factors (independent variables) with each other in an attempt to show how closely tied to each other they are.  

Today's topic is Teacher Salaries

The US Census Bureau provides information on the average teacher salary across all K-12 schools and also broken out by primary teachers and secondary teachers.  These values are included in the study as actual dollar amounts, amounts adjusted for inflation over time using the Bureau of Labor Statistics' Consumer Price Index, and amounts adjusted for state cost of living using the Bureau of Economic Analysis' Regional Price Parity calculation.

In terms of other independent variables, I will list the interactions in terms of the strength of the highest correlation between variables observed; Strong (+/- 1.0 to 0.5), Moderate (+/- 0.5 to 0.3), Weak (+/- .03 to 0.1), and None or Very Weak (+/- 0.1 to 0.0).

Strong

  • There is a moderate to strong positive correlation between period and teacher salaries; indicating that as time passes teachers are paid, on average, more.  This correlation is actually stronger when the dollar amounts are adjusted for regional cost of living (RPP) or inflation (CPI2014).  
  • There is a strong positive correlation between RPP and teacher salaries; indicating that states with higher cost of living also have higher average teacher salaries.  This relationship remains strong when adjusted for inflation (CPI2014), and becomes a weak positive correlation when the dollar amounts are adjusted for RPP; which means that even when the dollar amounts are adjusted for state cost of living, the relationship remains statistically significant between RPP and teacher salary.  
  • There is a moderate negative relationship between the percent of graduates taking the ACT exam and teacher salary, and a moderate to strong positive relationship between the percent of graduates taking the SAT exam and teacher salary; meaning students from states with higher average teacher salaries are less likely to take the ACT and more likely to take the SAT.  This relationship remains moderate when adjusted for inflation (CPI2014), but becomes weak or non-significant when controlled for state cost of living (RPP).  
  • There is a strong positive correlation between teacher salaries and school spending; indicating that states that spend more per pupil also have higher teacher salaries.
  • There is a moderate positive correlation between teacher salaries and percent of the population 18-14 with at least a high school diploma, a moderate to strong positive correlation between teacher salaries and percent of the population 25 and up with at least a bachelor's degree, and a strong positive correlation between teacher salaries and percent of the population 25 and up with a graduate degree, and no correlation between teacher salaries and percent of the population 25 and up with at least a high school diploma; indicating that states with higher teacher salaries also tend to have a more educated populace.  
  • There is a moderate to strong negative correlation between poverty and teacher salaries; with higher percents of poverty tied to lower teacher salaries.  This does not hold true, however, on the teacher salary amounts adjusted for regional cost of living (RPP); suggesting both poverty and teacher salaries are related to regional cost of living but not necessarily directly related to each other. 
  • There is a strong positive correlation between teacher salary and population per square mile; indicating states with higher teacher salaries are more densely populated.
  • There is a moderate to strong positive correlation between teacher salaries and both per capita and household income; indicating that states with higher teacher salaries also have higher per capita and household incomes.

Moderate

  • There were no comparisons for which the strongest correlation was moderate.

Weak

  • There is no correlation between the percent of special education students and teacher salaries (in actual dollars and inflation adjusted), but a weak positive correlation when the salary figures are adjusted for regional price parity (RPP); indicating that there is a relationship between the percent served under IDEA and teacher salaries when controlled for state cost of living with more special education students tied to higher cost of living.
  • There is a weak negative correlation between percent of students eligible for free or reduced price lunch and average teacher salary (raw and CPI2014),  but this correlation becomes non-significant when adjusting for regional price parity (RPP); suggesting that higher percents of at-risk students and lower teacher salaries are both related to lower state cost of living and not directly related to each other.  
  • There is a weak positive correlation between percent of students participating in ELL programs and teacher salaries when looking at actual dollar amounts and inflation-adjusted amounts, but not when adjusted for state cost of living; suggesting that higher percents of ELL students and higher teacher salaries are both related to lower state cost of living and not directly related to each other.
  • There is a weak to moderate negative correlation between teacher salaries and percent of white students, a weak positive correlation between teacher salaries and percent of Hispanic students, a weak negative correlation between teacher salaries and percent of American Indian or Alaska native students, and no correlation between teacher salaries and percent of black students when looking at actual dollar amounts and amounts adjusted for inflation (CPI2014).  When looking at teacher salaries adjusted for state cost-of-living (RPP), there is no correlation between teacher salaries and the percent of white and the percent of American Indian or Alaska native students (with the exception of percent white and secondary teacher salaries with a weak negative correlation), no correlation between teacher salaries and percent of Hispanic students, and a weak positive correlation between teacher salaries and percent of black students.  This suggests that, in general, states with higher percents of white and Native American or Alaska native students and higher have lower average teacher salaries, but that the nature of this relationship changes when state cost-of-living is taken into consideration; in which case states with higher percents of black students and lower percents of Native American or Alaska native students have higher teacher salaries.

None or Very Weak

  • There is no correlation between student-teacher ratio and average teacher salaries.  
  • There is no correlation between teacher salaries and instruction as a percent of either current spending or total spending. 

Friday, November 14, 2014

Factors Influencing Student Outcomes and How They Interact - 08 - Student-Teacher Ratio

In response to the feedback received on "Educational Funding and Student Outcomes: The Relationship as Evidenced by State-Level Data," the KASB Research Department is working on a "Part II" which will dig further into other factors that impact student outcomes, and how funding impacts when these other factors are taken into consideration.

In this series of blog posts, I will describe the preliminary correlation analysis comparing these factors (independent variables) with each other in an attempt to show how closely tied to each other they are.  

Today's topic is Student to Teacher Ratio

NCES provides information on the number of students and the number of teachers by state and year; which can be used to calculate the average number of students per teacher.  

In terms of other independent variables, I will list the interactions in terms of the strength of the highest correlation between variables observed; Strong (+/- 1.0 to 0.5), Moderate (+/- 0.5 to 0.3), Weak (+/- .03 to 0.1), and None or Very Weak (+/- 0.1 to 0.0).

Strong

  • There is a strong negative correlation between the percent of special education students and the student teacher ratio; indicating that states with fewer special education students have more students per teacher. 
  • There is a strong positive correlation between percent of students participating in ELL programs and student-teacher ratio; indicating that states with higher percents of students receiving English language learner services have more students per teacher. 
  • There is a moderate to strong negative correlation between student-teacher ratio and all measures of school spending; indicating that states with higher per pupil spending also have fewer students per teacher.  

Moderate

  • There is a moderate negative relationship between the number of students per teacher and percent of 18 to 24 year olds with at least a high school diploma and no correlation between the student teacher ratio and all other measures of educational attainment.  This indicates that states with fewer students per teacher have more residents ages 18 to 24 who have graduated from high school.
  • There is a weak to moderate negative correlation between student teacher ratio and per capita personal income but no correlation between student teacher ratio and median household income; suggesting states with more students per teacher are also likely to have lower per capita personal income.  

Weak

  • There is no correlation between percent of graduates taking the ACT exam and student-teacher ratio, but there is a weak negative correlation between percent of graduates taking the SAT and student teacher ratios; indicating graduates from states with more students per teacher are less likely to take the SAT.
  • There is a weak positive correlation between percent of students eligible for free or reduced price lunch and the student-teacher ratio; indicating that states with higher percents of at-risk students also have more students per teacher. 
  • There is a weak negative correlation between student-teach ratio and the percent of white students, a moderate positive correlation between student-teacher ratio and the percent of Hispanic students, and no correlation between student-teacher ratio and percent of black or American Indian or Alaska Native students; indicating that states with higher percents of white students and lower percents of Hispanic students have fewer students per teacher.
  • There is a weak negative correlation between student-teacher ratio and instruction spending as a percent of both current spending and total revenue; indicating that states with more students per teacher allocate less to expenses classified as "instruction." 
  • There is a weak positive to no correlation between student teacher ratio and poverty; indicating that states with more students per teacher are somewhat more likely to have higher poverty levels.
  • There is a weak negative correlation between student teacher ratio and persons per square mile; indicating that state with more student per teacher also tend to be less densely populated.

None or Very Weak

  • There is no correlation between period and student-teacher ratio; indicating that there is not a consistent trend of either increases or decreases in the number of students per teacher over time.
  • There is no correlation between state cost-of-living (RPP) and student-teacher ratios.
  • There is no correlation between student-teacher ratio and average teacher salaries.  

Factors Influencing Student Outcomes and How They Interact - 07 - Student Race/Ethnicity

In response to the feedback received on "Educational Funding and Student Outcomes: The Relationship as Evidenced by State-Level Data," the KASB Research Department is working on a "Part II" which will dig further into other factors that impact student outcomes, and how funding impacts when these other factors are taken into consideration.

In this series of blog posts, I will describe the preliminary correlation analysis comparing these factors (independent variables) with each other in an attempt to show how closely tied to each other they are.  

Today's topic is Student Race/Ethnicity

NCES provides statistics on the percent of students in each major race and/or ethnic category based on self-identification.  This study includes information on the four categories in which the vast majority of the states' student populations fall; White, Black, Hispanic, and American Indian or Alaska Native.  

In terms of other independent variables, I will list the interactions in terms of the strength of the highest correlation between variables observed; Strong (+/- 1.0 to 0.5), Moderate (+/- 0.5 to 0.3), Weak (+/- .03 to 0.1), and None or Very Weak (+/- 0.1 to 0.0).

Strong

  • There is a strong negative correlation between percent of students eligible for free or reduced-price lunch and percent of white students, a moderate positive correlation between percent of students eligible for free or reduced-price lunch and percent of black students, and a weak positive correlation between percent of students eligible for free or reduced-price lunch and percent of Hispanic students.  There is no correlation between NSLP status and percent of students who are American Indian or Alaska native.  This indicates that states with larger percents of non-white students also have higher percents of students eligible for free or reduced-price lunch, and further that percent of black students is more predictive of free and reduced price lunch eligible students than is percent of Hispanic students.  
  • There is a strong negative correlation between percent of students participating in ELL programs and percent of white students, a weak negative correlation between percent of students participating in ELL programs and percent of black students, a strong positive correlation between percent of students participating in ELL programs and percent of Hispanic students, and a weak positive correlation between students participating in ELL programs and percent of American Indian and/or Alaska native students; indicating that states with higher percents of students participating in ELL programs have fewer white and (to a lesser degree) black students and more Hispanic and (to a lesser degree) American Indian or Alaska native students.

Moderate

  • There is a moderate negative relationship between percent of white students and RPP, a moderate positive relationship between percent of Hispanic students and RPP, and no correlation between percent of Black or American Indian / Alaska Native students and RPP.  This indicates that states with higher cost of living are likely to have fewer white students and more Hispanic students than state with lower cost of living. 
  • There is a moderate positive correlation between percent of special education students and percent of white students, a moderate negative correlation between percent of special education students and percent of Hispanic students, and no correlation between percent of special education students and percent of black or American Indian or Alaska Native students.  This indicates that states with more white students and less Hispanic students have more students receiving services under IDEA. 
  • There is a weak to moderate negative correlation between teacher salaries and percent of white students, a weak positive correlation between teacher salaries and percent of Hispanic students, a weak negative correlation between teacher salaries and percent of American Indian or Alaska native students, and no correlation between teacher salaries and percent of black students when looking at actual dollar amounts and amounts adjusted for inflation (CPI2014).  When looking at teacher salaries adjusted for state cost-of-living (RPP), there is no correlation between teacher salaries and the percent of white and the percent of American Indian or Alaska native students (with the exception of percent white and secondary teacher salaries with a weak negative correlation), no correlation between teacher salaries and percent of Hispanic students, and a weak positive correlation between teacher salaries and percent of black students.  This suggests that, in general, states with higher percents of white and Native American or Alaska native students and higher have lower average teacher salaries, but that the nature of this relationship changes when state cost-of-living is taken into consideration; in which case states with higher percents of black students and lower percents of Native American or Alaska native students have higher teacher salaries.
  • There is a moderate positive correlation between percent of white students and percent of the population with at least a high school diploma, a moderate to strong negative correlation between the percent of black and Hispanic students and the percent of the population with at least a high school diploma, and a weak negative correlation between percent of 18-24 year olds with at least a high school diploma and the percent of American Indian or Alaska native students, but a weak positive correlation between percent of 25 year olds and older with at least a high school diploma.  There is a weak negative correlation between percent of black students and percent of people with at least a bachelor's degree, a weak positive correlation between percent of Hispanic students and percent of people with at least a bachelor's degree, and no correlation between percent of white and American Indian or Alaska native students and percent of people with at least a bachelor's degree.  There is a weak negative correlation between the percent of white and American Indian or Alaska native students and the percent of people with a graduate degree, a weak correlation between the percent of Hispanic students and percent of people with a gradate degree, and no correlation between percent of black students and percent of population with a graduate degree.  

Weak

  • There is a weak positive correlation between percent of white students in a state and percent taking the ACT, and a weak negative correlation between percent taking the SAT and percent of white students in a state.  This relationship is reversed when looking at percent of Hispanic students in a state; meaning states with more white students and fewer Hispanic students are more likely to take the ACT and less likely to take the SAT.  There is no correlation between percent of black students and ACT or SAT participation.  There is no correlation between percent of Native American or Alaska Native students in a state and percent taking the ACT, but there is a weak negative relationship between percent of Native American or Alaska Native students and percent participation on the SAT; meaning students from states with a larger percent of American Indian or Alaska Native students are less likely to take the SAT. 
  • There is a weak negative correlation between student-teach ratio and the percent of white students, a moderate positive correlation between student-teacher ratio and the percent of Hispanic students, and no correlation between student-teacher ratio and percent of black or American Indian or Alaska Native students; indicating that states with higher percents of white students and lower percents of Hispanic students have fewer students per teacher.
  • There is no correlation between school spending and percent of white, black, or American Indian or Alaska native students and a weak negative correlation for some measures of school spending and percent of Hispanic students when looking at actual dollars and dollars adjusted for inflation (CPI2014).  When adjusting for state cost-of-living (RPP), there is a weak positive correlation between the percent of white students and school spending, a weak negative correlation between the percent of Hispanic students and school spending, and no correlation between school spending and the percent of black and American Indian or Alaska native students.  This suggests that beyond regional cost differences, states with higher percents of white students and lower percents of Hispanic students have higher per pupil education spending.  
  • There is a weak positive correlation between percent of white students and spending on instruction as a percent of current spending, and a moderate negative correlation between spending on instruction as a percent of current spending and percent of American Indian or Alaska native students, but no correlation between spending on instruction and percent of black or Hispanic students.  There is a weak positive correlation between percent of white students and spending on instruction as a percent of total revenue, weak positive correlations between percent of Hispanic and American Indian or Alaska native students and spending on instruction as a percent of total revenue, and no correlation between the percent of black students and spending on instruction as a percent of total revenue.  This suggests that states with higher percents of white students and lower percents of American Indian or Alaska native students allocate more funds to expenses classified as "instruction," and when looking at total spending, states with higher percents of Hispanic students put more of their total funding amounts towards "instruction."  
  • There is a weak negative correlation between percent of white students and poverty, a weak to moderate positive correlation between percent of black students and poverty, a weak positive to no correlation between percent of Hispanic students and poverty, and no correlation between percent American Indian or Alaska native students and poverty (with the exception of a weak positive correlation between this and percent of children below 250% poverty).  This indicates that states with lower percents of white students and higher percents of black and Hispanic students have higher poverty.
  • There is a weak negative correlation between population per square mile and percent of white and American Indian or Alaska native students, a weak positive correlation between population per square mile and percent of black students, and no correlation between population per square mile and Hispanic students.  This indicates states with lower percents of white and American Indian or Alaska native students and higher percents of black students have more people per square mile. 
  • There is a weak negative to no correlation between percent of white students and personal or household income, a weak to moderate negative correlation between percent of black students and personal or household income, a weak positive to no correlation between percent of Hispanic students and personal or household income, and no correlation between percent of American Indian or Alaska native and personal or household income when looking at actual dollar amounts and amounts adjusted for inflation (CPI2014).  When adjusting for state cost-of-living (RPP), there is a weak positive correlation between the percent of white students and personal or household income, a weak to moderate negative correlation between percent of black students and personal or household income, and no correlation between the percent of Hispanic and American Indian or Alaska native students and personal or household income.  This suggests that, beyond regional cost differences, states with higher percents of white students and lower percents of black students have higher personal or household incomes.

None or Very Weak

  • There is no correlation between percent of students in various race and ethnic categories (White, Black, Hispanic, and American Indian / Alaska Native) and period; indicating these percents are remaining constant on average over time.

Tuesday, November 11, 2014

Factors Influencing Student Outcomes and How They Interact - 06 - English Language Learners

In response to the feedback received on "Educational Funding and Student Outcomes: The Relationship as Evidenced by State-Level Data," the KASB Research Department is working on a "Part II" which will dig further into other factors that impact student outcomes, and how funding impacts when these other factors are taken into consideration.

In this series of blog posts, I will describe the preliminary correlation analysis comparing these factors (independent variables) with each other in an attempt to show how closely tied to each other they are.  

Today's topic is Percent of students participating in programs for English Language Learners (ELL)

NCES reports on the percent of students participating in programs designed for students from homes where English is not the primary language and/or who demonstrate a need for additional assistance with speaking and writing English; referred to English Language Learners (ELL).  

In terms of other independent variables, I will list the interactions in terms of the strength of the highest correlation between variables observed; Strong (+/- 1.0 to 0.5), Moderate (+/- 0.5 to 0.3), Weak (+/- .03 to 0.1), and None or Very Weak (+/- 0.1 to 0.0).

Strong

  • There is a strong negative correlation between percent of students participating in ELL programs and the percent of white students, a weak negative correlation between percent of students participating in ELL programs and the percent of black students, a strong positive correlation between percent of students participating in ELL programs and the percent of Hispanic students, and a weak positive correlation between students participating in ELL programs and the percent of American Indian and/or Alaska native students; indicating that states with higher percents of students participating in ELL programs have fewer white and (to a lesser degree) black students and more Hispanic and (to a lesser degree) American Indian or Alaska native students.
  • There is a strong positive correlation between percent of students participating in ELL programs and student-teacher ratio; indicating that states with higher percents of students receiving English language learner services have more students per teacher. 

Moderate

  • There is a moderate positive correlation between percent of student participating in programs for English Language Learners and RPP, meaning states with lower cost of living have lower percents participating in programs for English Language Learners.  This seems contrary to expectations, and may indicate states with higher cost of living have more ELL programs available than states with lower cost of living.  
  • There is a moderate negative correlation between percent of students participating in English Language Learner programs (ELL) and percent of special education students; indicating that states with larger special education student populations have smaller percents of students from homes where English is not the primary language.
  • There is a weak negative correlation between percent of students participating in ELL programs and education spending when looking at actual dollar amounts and inflation-adjusted amounts, and a moderate negative correlation when looking at state cost-of-living-adjusted amounts; suggesting that states with higher percents of ELL students have lower per pupil spending, and this trend is even more apparent when accounting for regional cost differences.
  • There is a weak to moderate negative correlation between ELL program participation and the percent of the population with at least a high school diploma, and a weak positive correlation between ELL program participation and percent of the population with a college degree; indicating that states with higher percents of ELL students have lower percents of high school completers but higher percents of college graduates.  

Weak

  • There is a weak negative correlation between percent participating in ELL programs and percent taking the ACT, and a weak positive relationship between the percent taking the SAT and the percent participating in ELL programs; meaning students from states with higher percents of students in ELL programs are less likely to take the ACT and more likely to take the SAT.  
  • There is a weak positive correlation between percent of students participating in ELL programs and percent of students eligible for free or reduced-price lunch; indicating that states with higher percents of students identified as in poverty or otherwise at risk also have higher percents of students identified to receive English Language Learner services.  
  • There is a weak positive correlation between percent of students participating in ELL programs and teacher salaries when looking at actual dollar amounts and inflation-adjusted amounts, but not when adjusted for state cost of living; suggesting that higher percents of ELL students and higher teacher salaries are both related to lower state cost of living and not directly related to each other.  
  • There is no correlation between ELL program participation and spending on instruction as a percent of current spending and a weak negative correlation between ELL program participation and spending on instruction as a percent of total revenue; suggesting states with higher percents of ELL students allocate smaller percents of spending towards expenses categorized as "instruction."
  • There is no correlation between percent of ELL students and per capita personal income when looking at actual dollars, a weak positive correlation when looking at personal income adjusted for inflation (CPI2014), and a weak negative correlation when looking at personal income adjusted for state cost-of-living (RPP).  There is a weak positive correlation between between percent of ELL students and household income when looking at actual dollars and dollars adjusted for inflation (CPI2014), and no correlation when looking at household income adjusted for state cost-of-living (RPP).  This suggests that states with higher percents of ELL students have higher incomes except when taking state cost-of-living into account, though this relationship is weak.

None or Very Weak

  • There is no correlation between the percent of students participating in programs for English Language Learners and period; indicating this is remaining fairly constant over time. 
  • There is no correlation between percent of students participating in programs for English Language Learners and poverty.  
  • There is no correlation between percent of students participating in programs for English Language Learners and population density.

Monday, November 10, 2014

Factors Influencing Student Outcomes and How They Interact - 05 - Free and Reduced-Price Lunch

In response to the feedback received on "Educational Funding and Student Outcomes: The Relationship as Evidenced by State-Level Data," the KASB Research Department is working on a "Part II" which will dig further into other factors that impact student outcomes, and how funding impacts when these other factors are taken into consideration.

In this series of blog posts, I will describe the preliminary correlation analysis comparing these factors (independent variables) with each other in an attempt to show how closely tied to each other they are.  

Today's topic is Students Eligible for Free or Reduced-Price Lunch

NCES reports on the percent of students eligible for free lunch or reduced-price lunch under the National School Lunch Program (NSLP).  This is the most commonly used proxy measure for student financial status within schools.

In terms of other independent variables, I will list the interactions in terms of the strength of the highest correlation between variables observed; Strong (+/- 1.0 to 0.5), Moderate (+/- 0.5 to 0.3), Weak (+/- .03 to 0.1), and None or Very Weak (+/- 0.1 to 0.0).

Strong

  • There is a strong negative correlation between percent of students eligible for free or reduced-price lunch and percent of white students, a moderate positive correlation between the percent of students eligible for free or reduced-price lunch and percent of black students, and a weak positive correlation between percent of students eligible for free or reduced-price lunch and percent of Hispanic students.  There is no correlation between NSLP status and percent of students who are American Indian or Alaska native.  This indicates that states with larger percents of non-white students also have higher percents of students eligible for free or reduced-price lunch, and further that percent of black students is more predictive of free and reduced price lunch eligible students than is percent of Hispanic students.  
  • There is a moderate to strong negative correlation between percent of students eligible for free or reduced-price lunch and educational attainment in the state; indicating that states with lower percents of at-risk students have higher percent of people achieving high school diplomas and college degrees.  
  • There is a strong positive correlation between percent of students eligible for free or reduced-price lunch and poverty; indicating that states with higher percent of at-risk students (as defined largely by income level) also have higher poverty levels.
  • There is a moderate to strong negative correlation between percent of students eligible for free or reduced-price lunch and both personal and household income.  This holds true when income is adjusted for inflation (CPI2014) and state cost of living (RPP).  

Moderate

  • There is a moderate negative correlation between the percent of students eligible for free or reduced-price lunch and cost of living by state (RPP); meaning that states with higher cost of living have lower percents of student eligible for the National School Lunch Program's free and reduced-price lunch offerings.
  • There is a moderate positive correlation between the percent eligible for free or reduced-price lunch and percent taking the ACT and  and a weak negative correlation between the percent eligible for free or reduced-price lunch and the percent taking the SAT; meaning students from states with higher percents of free or reduced-price lunch participants are more likely to take the ACT and less likely to take the SAT. 
  • There is a moderate negative correlation between percent of students eligible for free and reduced-price lunch and education spending.  This is true for actual dollars and dollars adjusted for inflation (CPI2014) and state cost of living (RPP); suggesting that states with more at-risk students (largely due to poverty) also have lower per pupil spending.  

Weak

  • There is a weak positive correlation between the percent of students eligible for free or reduced-price lunch and period; indicating that a larger percent of students are eligible each year. 
  • There is a weak negative correlation between percent of students eligible for free or reduced-price lunch and percent of special education students; indicating that states with higher percents of special education students have lower percents of students identified for non-full-price meals. 
  • There is a weak positive correlation between percent of students eligible for free or reduced-price lunch and percent of students participating in ELL programs; indicating that states with higher percents of students identified as in poverty or otherwise at risk also have higher percents of students identified to receive English Language Learner services.  
  • There is a weak positive correlation between percent of students eligible for free or reduced price lunch and the student-teacher ratio; indicating that states with higher percents of at-risk students also have more students per teacher.  
  • There is a weak negative correlation between percent of students eligible for free or reduced price lunch and average teacher salary (raw and CPI2014),  but this correlation becomes non-significant when adjusting for regional price parity (RPP); suggesting that higher percents of at-risk students and lower teacher salaries are both related to lower state cost of living and not directly related to each other.  
  • There is a weak negative correlation between percent of students eligible for free or reduced-price lunch and spending on instruction as a percent of current spending, and no correlation between percent of students eligible for free and reduced-price lunch and spending on instruction as a percent of total revenue; indicating that states with higher percents of at-risk students have lower percents of current spending  going to expenses categorized as "instruction."  
  • There is a weak negative correlation between percent of students eligible for free or reduced-price lunch and population per square mile; indicating that states with lower percents of at-risk students have higher numbers of persons per square mile. 

None or Very Weak

  • No correlations with Students Eligible for Free or Reduced-Price Lunch yielded very weak or non-significant results.

Friday, November 7, 2014

Factors Influencing Student Outcomes and How They Interact - 04 - Special Education

In response to the feedback received on "Educational Funding and Student Outcomes: The Relationship as Evidenced by State-Level Data," the KASB Research Department is working on a "Part II" which will dig further into other factors that impact student outcomes, and how funding impacts when these other factors are taken into consideration.

In this series of blog posts, I will describe the preliminary correlation analysis comparing these factors (independent variables) with each other in an attempt to show how closely tied to each other they are.  

Today's topic is Percent of Students ages 3-21 Served Under IDEA

NCES reports on the number of students with disabilities or special needs served under the Individuals with Disabilities Education Act; providing an estimate of the percent of students requiring special accommodations and services within the schools.

In terms of other independent variables, I will list the interactions in terms of the strength of the highest correlation between variables observed; Strong (+/- 1.0 to 0.5), Moderate (+/- 0.5 to 0.3), Weak (+/- .03 to 0.1), and None or Very Weak (+/- 0.1 to 0.0).

Strong

  • There is a strong negative correlation between the percent of special education students and the student teacher ratio; indicating that states with fewer special education students have more students per teacher.  

Moderate

  • There is a moderate negative correlation between percent of special education students and percent of students participating in English Language Learner programs (ELL); indicating that states with larger special education student populations have smaller percents of students from homes where English is not the primary language.
  • There is a moderate positive correlation between the percent of special education students and the percent of white students, a moderate negative correlation between the percent of special education students and the percent of Hispanic students, and no correlation between the percent of special education students and the percent of black or American Indian or Alaska Native students.  This indicates that states with more white students and less Hispanic students have more students receiving services under IDEA. 
  • There is a moderate positive correlation between percent of students served under IDEA and education funding with and without adjustments for inflation (CPI2014) and state cost of living (RPP); indicating that states with higher percents of special education students spend more on average per pupil.  
  • There is a moderate positive correlation between percent of students served under IDEA and population per square mile; indicating that more densely populated states have higher percentages of special education students.  

Weak

  • There is a weak negative correlation between percent of special education students and time (period); indicating that a smaller percent of students are being served under IDEA each year.
  • There is a weak negative correlation between percent of special education students and ACT participation, and a weak positive correlation between percent of special education and SAT participation; indicating students from states with a higher percent of special education students are less likely to take the ACT and more likely to take the SAT. 
  • There is a weak negative correlation between percent of special education students and percent of students eligible for free or reduced-price lunch; indicating that states with higher percents of special education students have lower percents of students in poverty. 
  • There is no correlation between the percent of special education students and teacher salaries (in actual dollars and inflation adjusted), but a weak positive correlation when the salary figures are adjusted for regional price parity (RPP); indicating that there is a relationship between the percent served under IDEA and teacher salaries when controlled for state cost of living with more special education students tied to higher cost of living.
  • There is no correlation between percent of students served under IDEA and spending on instruction as a percent of current spending, and a weak positive correlation between percent of students served under IDEA and spending on instruction as a percent of total spending; indicating states with higher percents of special education students put more of total spending towards instruction.
  • There is a weak positive correlation between percent of students served under IDEA and the percent of the population with at least a high school diploma, but no correlation between percent of student served under IDEA and percent of the population with a college degree; indicating that states with higher percents of special education students also have higher percents of people with at least a high school diploma.  
  • There is a weak negative correlation between percent of students served under IDEA and poverty ; indicating that states with lower percents of special education students have higher poverty levels.
  • There is no correlation between percent of students served under IDEA and the median household income (actual, CPI2014, and RPP), but there is a weak positive correlation between the percent of special education students and the per capita personal income (CPI2014 and RPP, but not actual); indicating that states with higher percents of special education students have higher per capita personal incomes when adjusted for inflation or cost of living.

None or Very Weak

  •  There is no correlation between percent of special education students and state cost of living (RPP).