Monday, September 28, 2015

The Race Achievement Gap

The following post presents research or analyses from outside KASB and is presented for information purposes.  KASB neither endorses nor refutes the conclusions or recommendations contained herein.
for your consideration.gif


The National Center for Education Statistics recently released a report entitled “School Composition and the Black-White Achievement Gap.” In it, they present results of analyses related to the impact of school composition on student outcomes. Put in plainer English, they looked at how the percent of White and Black students in a school impact student outcomes for Black and White students.  


Initially, the study found:


  • “Achievement for both Black and White students was lower in the highest Black student density schools than in the lowest density schools.”
  • “The achievement gap was not different [for black and white students].”


In other words, the higher the percent of Black students in a school, the lower the achievement scores were for both Black and White students. Why do you suppose that might be?  Here are some quotes from the article that might help explain:


  • “Schools that serve large percentages of Black students are more likely to employ less experienced teachers.”
  • “Schools with a higher percentage of students who are Black tend to have higher shares of low-socioeconomic-status students, who often need additional supports to be successful because Black students are more likely to be in a one-parent/guardian family, to be in a family in poverty, and to have parents with lower levels of education, compared with the parents of White students.”
  • “There is a body of work that explores whether negative educational outcomes, such as lower achievement, that are associated with large concentrations of Black students in schools might be due to an “oppositional culture,” which is a part of contemporary Black culture. This line of research considers student peer effects associated with larger concentrations of Black students where it has been theorized that certain behaviors that are associated with higher achievement are shunned because they involve learning to cope with pressures such as ‘the burden of acting White’.”
  • “Some researchers have considered whether teachers may also have lower expectations for student performance in schools with a high population of Black students, sometimes explained as a “Pygmalion effect.” This research is grounded in the assumption that lower expectations by teachers for students from minority backgrounds may result in lower levels of engagement by both teachers and students, which ultimately may contribute to poorer academic performance. One study found that in predominantly Black elementary schools, Black and White students tend to score lower and eventually are placed on a lower track in high school, and this tracking can start in elementary school.”
  • “At the high school level, some research shows that the tracking of Black students tends to differ by the density of Black students in the school. One study found that that Black students are more likely to be in high-track courses (e.g., taking algebra in the eighth grade rather than the ninth grade) in predominantly Black schools than in lower density schools. Another study found that even when controlling for achievement, more racial-ethnic and socioeconomic diversity are related to more “de facto” tracking.”
  • “The number of school disciplinary reports increases as the percentage of students in a school who are Black increases, and Black students are more likely than White students to face school discipline or office referrals (Rocque and Paternoster 2011), which is relevant because higher rates of out-of-school suspension are related with lower achievement.”


Based on these items, it would seem that there are many factors that could cause the achievement gap between schools that are predominantly Black and those that are predominantly White.  


The study also found that:
  • “Black students are, on average, in schools that are 48 percent Black, whereas White students are, on average, in schools that are 9 percent Black.”
  • “Schools in the highest Black student density category are mostly located in the South, with very few in the West.”
  • “Schools in the highest Black student density category are mostly in cities, but this varies by region.”
  • “Schools with higher Black student density also have higher percentages of students with low socioeconomic status.”


So, Black students are more likely to go to predominantly Black schools than White students are, schools in the South are more likely to be predominantly Black, and schools in cities are more likely to be predominantly Black than rural schools.  


But more significantly, there is a correlation between the density of Black students and the overall socioeconomic status of the school. Because a lot of previous research has shown a correlation between poverty and student achievement, it is possible that the connection between Black student density and student achievement could be based on their mutual connection to poverty.  


Because of this, the researchers went on to control for socioeconomic status (via regression analysis), and found that “the previously observed relationship between Black student density and achievement disappeared for Whites but not for Blacks.” When they controlled for socioeconomic status, student, teacher, and school characteristics, “ the achievement gap was greater among schools with the highest Black student density than the schools with the lowest.” So the achievement gap between Black and White students is higher in schools with a higher percent of Black students, even when controlling for poverty and the characteristics of the students, teachers, and schools.  


The study goes on to discuss differences in gender and between-school versus within-school differences.  


What does this all mean?


  • First, it again illustrates how complex issues like achievement gaps are, and points out how many interrelated factors are at work.  
  • Second, it tells us that the achievement gap between white and black students is worse in schools with a higher percent of black students, even when you control for many of the interrelated factors such as poverty.


When discussing education policy, we tend to talk a lot about students “in general” and try to think about education divorced from issues like race and poverty.  But the reality is that education is not simple or transparent, and it is important to take all factors into consideration.  

Friday, September 25, 2015

Best Practices for Teacher Merit Pay Systems

The following post presents research or analyses from outside KASB and is presented for information purposes. KASB neither endorses nor refutes the conclusions or recommendations contained herein.

As teacher merit pay continues to receive media attention in Kansas, KASB is reviewing pertinent research on the topic.  


A report released in February by the Center for American Progress (CEP) discusses lessons learned by 10 school districts that implemented teacher merit pay systems.  CEP is “an independent nonpartisan policy institute that is dedicated to improving the lives of all Americans, through bold, progressive ideas, as well as strong leadership and concerted action.”


The report indicated that all districts considered the following elements in their merit pay systems:
  • Base salary
  • Teacher effectiveness
  • Speed of salary growth
  • Career pathway opportunities
  • Incentives for hard-to-staff schools and positions
  • Bonuses, rewards, and recognition
  • Opt-in timeframe


From these components, the following best practices emerged:
  1. Differentiate compensation based on roles and responsibilities.
  2. Set starting salaries to meet market demand.
  3. Align teacher compensation redesign with fair and proven teacher evaluation systems.
  4. Shift pay away from years of experience and advanced degree attainment.
  5. Use compensation incentives to attract highly effective teachers to hard-to-staff schools, districts, and subjects.
  6. Emphasize extra pay for effectiveness and career pathways instead of small bonuses.
  7. Accelerate the timeline to earning the maximum salary where possible.
  8. Allow teachers to opt-in to new compensation systems within a set timeframe.

For more details, see the full report.

Tuesday, September 22, 2015

Record Breaking Numbers!

In a recent blog post, the Kansas Policy Institute (KPI) reports that the “Kansas General Fund spending sets new records.” KPI’s claim is that the general fund spending is increasing at a rate higher than inflation, but the data presented is potentially misleading.  

The post presents the following table:




Here is the data presented in chart form:




KPI’s table has a column called “adjusted for inflation,” but that is not actually what it is.  


In statistics, when you say you are adjusting for something, it typically means you are controlling for, or removing, the effects of that something.  However, in KPI’s table, the “adjusted for inflation” amount is just the opposite; it is the effects of inflation with all other effects removed.


To calculate their “adjusted”column, KPI has taken the General Fund amount from 1995 and then added to it the annual inflation amount through 2017. So essentially they are saying that the General Fund should be exactly the same as it was in 1995 aside from the effects of inflation.


To explore this further, let’s first replace KPI’s “adjusted” line with a line representing the actual general fund amounts from each year with the effects of inflation removed. To do this, we take the actual dollar amount for each year and subtract from it the dollar amount times the cumulative inflation amount for that year.  This results in the yearly state general fund amount being expressed in 1995-equivalent dollars.    




The new line shows that dollar amounts have generally increased over time, but it also shows that the peak for state general fund amounts was in 2008-2009, and that the amounts have been fairly stable since 2012 or so. It shows the effects of everything but inflation, whereas KPI’s line shows us the effects of inflation alone.  


This is what you will usually see when people talk about adjusting for inflation. The chart above shows the state general fund amounts expressed in 1995 dollars, but it can also be done using the current year (2015) as the constant, as shown below:




Notice that when using 2015 equivalent dollars, the trend is essentially the same; the difference is where the actual and the adjusted numbers intersect.


Here is the table showing the data used for the two charts above:


Fiscal Year
Actual / Planned
Midwest Urban CPI
CPI % Change (1995)
CPI % Change (2015)
Adj. for Inflation (1995 Dollars)
Adj. for Inflation (2015 Dollars)
95
3.310
148.40
-  
(0.35)
3.310
5.510
96
3.439
153.00
0.03
(0.33)
3.336
5.513
97
3.538
156.70
0.06
(0.32)
3.353
5.511
98
3.799
159.30
0.07
(0.30)
3.556
5.701
99
4.196
162.70
0.10
(0.29)
3.877
6.006
00
4.368
168.30
0.13
(0.27)
3.924
6.025
01
4.43
172.80
0.16
(0.25)
3.886
5.964
02
4.466
174.90
0.18
(0.24)
3.875
5.942
03
4.138
178.30
0.20
(0.22)
3.471
5.522
04
4.317
182.60
0.23
(0.20)
3.554
5.583
05
4.69
188.40
0.27
(0.18)
3.798
5.798
06
5.139
193.00
0.30
(0.16)
4.144
6.121
07
5.608
198.12
0.34
(0.13)
4.499
6.451
08
6.102
205.38
0.38
(0.10)
4.831
6.746
09
6.064
204.06
0.38
(0.11)
4.822
6.744
10
5.268
208.05
0.40
(0.09)
3.938
5.840
11
5.667
214.74
0.45
(0.06)
4.187
6.056
12
6.098
219.10
0.48
(0.04)
4.521
6.368
13
6.135
222.17
0.50
(0.03)
4.490
6.321
14
5.983
225.43
0.52
(0.02)
4.265
6.080
15
6.251
228.99
0.54
-  
4.454
6.251
16
6.322
232.28
0.57
0.01
4.451
6.232
17
6.399
235.65
0.59
0.03
4.453
6.217


The point to everything above is this; KPI is providing a baseline figure indicating what the state general fund would be today if it had only increased based on inflation since 1995. That is a perfectly legitimate way to present the data, but it is misleading to call it an “adjustment for inflation.”  


The question we should ask is this: Is it reasonable to expect the amount we are spending today to be exactly the same as what we were spending in 1995? Because that is what KPI’s post implies; that we should expect the rate of state general fund increase to match the rate of inflation. Have there been any other changes in Kansas over the past 20 years that would justify spending more now than we did then?


When looking at education-specific dollars, we frequently talk about “per student” amounts to control for the effects of the student population. What if we did the same with this information?  Should we not take into consideration the population of Kansas when we look at the state general fund amounts?


The next table and chart show the information KPI presented when adjusted for the population in Kansas:


 


Fiscal Year
Population
Per Person Actual/Planned
Per Person Adj. for Inflation (KPI's Way)
Variance
95
  2,565,000
1,290.45
1,290.45
-  
96
  2,615,000
1,315.11
1,326.32
(11.21)
97
  2,635,000
1,342.69
1,367.25
(24.56)
98
  2,661,000
1,427.66
1,392.20
35.46
99
  2,678,000
1,566.84
1,415.98
150.86
00
  2,693,681
1,621.57
1,458.48
163.09
01
  2,702,162
1,639.43
1,508.38
131.05
02
  2,713,535
1,645.82
1,527.49
118.34
03
  2,723,004
1,519.65
1,558.28
(38.64)
04
  2,734,373
1,578.79
1,587.13
(8.34)
05
  2,745,299
1,708.37
1,631.58
76.80
06
  2,762,931
1,859.98
1,686.94
173.04
07
  2,783,785
2,014.52
1,719.30
295.23
08
  2,808,076
2,173.02
1,781.68
391.34
09
  2,832,704
2,140.71
1,800.00
340.71
10
  2,858,837
1,842.71
1,819.49
23.21
11
  2,870,386
1,974.30
1,860.43
113.87
12
  2,885,905
2,113.03
1,913.84
199.19
13
  2,893,957
2,119.93
1,946.98
172.96
14
  2,904,000
2,060.26
1,972.32
87.94
15
  2,939,000
2,126.91
1,979.34
147.58
16
  2,912,319
2,170.78
2,001.95
168.83
17
  2,918,440
2,192.61
2,020.66
171.95


When looking at this data on a per-person amount, you see that the actual amounts are much closer to the inflation line, and actually dip below it in 2003.  


When you look at the actual “adjusted for inflation” amounts on a per-person basis, you see this:




Fiscal Year
Population
Per Person Actual/Planned
Per Person Adj. for Inflation (1995 Dollars)
Variance
95
  2,565,000
1,290.45
1,290.45
-  
96
  2,615,000
1,315.11
1,274.34
40.76
97
  2,635,000
1,342.69
1,267.60
    75.10
98
  2,661,000
1,427.66
1,322.80
 104.86
99
  2,678,000
1,566.84
1,415.86
 150.98
00
  2,693,681
1,621.57
1,404.12
 217.45
01
  2,702,162
1,639.43
1,369.87
 269.56
02
  2,713,535
1,645.82
1,351.93
 293.90
03
  2,723,004
1,519.65
1,213.46
 306.18
04
  2,734,373
1,578.79
1,214.94
 363.85
05
  2,745,299
1,708.37
1,247.90
 460.48
06
  2,762,931
1,859.98
1,300.98
 559.00
07
  2,783,785
2,014.52
1,339.54
 674.99
08
  2,808,076
2,173.02
1,338.63
 834.39
09
  2,832,704
2,140.71
1,337.74
 802.97
10
  2,858,837
1,842.71
1,102.07
 740.63
11
  2,870,386
1,974.30
1,091.68
 882.62
12
  2,885,905
2,113.03
1,106.35
  1,006.68
13
  2,893,957
2,119.93
1,066.11
  1,053.82
14
  2,904,000
2,060.26
990.91
  1,069.35
15
  2,939,000
2,126.91
971.94
  1,154.98
16
  2,912,319
2,170.78
943.78
  1,227.00
17
  2,918,440
2,192.61
903.47
  1,289.14


So, when you take the state general fund amount, adjust for inflation, and adjust for the population, you discover that we are actually spending less per-person than we were in 1995.  


And keep in mind that this doesn’t take into account changes in the percent of children and adults in poverty or any other demographic changes that could have resulted in a population in need of more assistance.  


To summarize, what does all of this tell us?


  • First, it tells us that in terms of actual dollar amounts, KPI is right; the state general fund amounts for 2015-2017 will be record breakers.
  • Second, when you accurately adjust the dollar amounts for inflation, you see that the peak spending period in Kansas between 1995 and 2017 was in 2008-2009, and that the amount in 2015 is about what it was in 2006-2007.
  • Third, when you adjust for the population and inflation, the per capita amount for the state general fund in 2015 is actually lower than it was in 1995, and the trend would suggest this per capita decline will continue through 2017.  


But most importantly, this once again illustrates the importance of considering what data has been presented and how it is presented when determining how much credence to put into any kind of research, article, or blog post involving data such as this.