Thursday, June 26, 2014

The Most Powerful Predictor of Student Performance

The public school accountability system in Louisiana as in most other states is based primarily upon student performance on standardized testing. Louisiana has gone so far as to give schools and school systems a letter grade for performance based primarily on student test performance. Based upon these ratings, pronouncements are often made to the public and to lawmakers about the quality of the schools and teacher effectiveness in individual school systems. Here is an example of education officials blaming teachers:
In the Louisiana application for the ESEA waiver of No Child Left Behind, Superintendent John White criticized the results of past teacher evaluations in Louisiana that yielded a high percentage of proficient teacher ratings even though more than one third of Louisiana students were not achieving at grade level on annual state assessments.

Is it correct to assume that the quality of teaching in a particular school system can be equated to the performance of its students on the state standardized testing system, or could there be other factors that affect student performance other than the quality of instruction? Do the letter grades given to our individual schools and school systems really reflect the quality of instruction in those systems?

The following analysis and table compares the student performance in each of the 74 public school systems in Louisiana with an accepted measure of student poverty (the percentage of students who qualify for free or reduced lunch). Using the latest LDOE analysis of state testing, school systems are ranked according to overall student performance along with the percentage of at risk students in each school system as determined by the free or reduced lunch data.

The table below demonstrates that the lowest three performing school systems: The Louisiana Recovery District, the Baton Rouge Recovery District and the St. Helena school district also had the highest percentage of students in poverty in the state (approximately 97.5% of students in those three systems were on free or reduced lunch). In addition, the rankings of all the other school systems indicate a clear relationship between average student performance and the rate of student poverty. Most of the school systems fall at approximately the same position on the academic achievement ranking as they fall on average family income ranking. As the level of poverty goes up, the level of student achievement goes down. It is an almost perfect mathematical relationship. In fact, the three school systems with the highest performance in the state have basically the lowest level of student poverty.

Louisiana is second only to Mississippi in the percentage of its students eligible for free and reduced lunch because of low family income. The median percentage of students in all school districts in Louisiana classified as at-risk because of low family income is 69.7%. The school system with the lowest percentage of students in free or reduced lunch is at 42.7%. The number one school system in student performance (Zachary Communty Schools) is second from the bottom of all school systems in percentage of its students on free or reduced lunch. Of the top 10 performing school systems, all but one fall near the 50% level in the percentage of their students on free or reduced lunch. For Louisiana that equates to relatively low poverty!

Would it be logical to conclude (as Louisiana has implied by use of the present school grading system) that school systems with the poorest children almost always have the least effective teachers? Or could it be that student poverty is more important than the effectiveness of teaching in determining average student performance? The table below shows an incredibly strong relationship between the level of student poverty in a school system and the level of student achievement. Is it logical therefore to conclude that the consistent difference in student performance in school systems with a difference of as little as 5 to 10 percentage points in poverty is probably caused more by the influence of poverty than by the competence of their teachers?

When one digs into the data at the local school level, one finds that there is a single factor that can trump poverty as the dominant factor in student performance. The data shows that schools such as magnet schools that select their students based upon past high academic achievement, perform significantly better regardless of the level of student poverty. For example, Forest Heights Academy of Excellence in East Baton Rouge Parish has a free or reduced lunch population of 77.4%, yet has achieved an “A”rating on the state accountability system. If this were not a selective academic school, the poverty rate would predict a school grade of “C” or less. But this outcome is a no-brain-er since the students who are allowed to attend were already high performers. So the fact that a student comes from a poor family does not automatically condemn that student to low academic performance. The lower academic performance of high poverty populations is simply an average result for such populations even though there is still a wide range of performances for individual students.

The KIPP schools in the New Orleans Recovery District have also managed to DE-select of counsel out most of the poorest kids that may have wanted to attend, and now have a free or reduced lunch percentage of 54%. Most schools in Louisiana with that low a percentage of high poverty students achieve an “A” rating on the Louisiana school grading scale, yet these schools have only managed a “B” rating. But the point is that their higher rating than most schools in the RSD is most probably much more related to their lower poverty student enrollment than to their teaching methods.

The same deviation caused by selective admission is demonstrated by the one public school system in the state which selects the majority of its students based on academic performance. The Orleans Parish School Board system (OPSB) still manages the schools that were performing above average when the state Recovery District took over all schools in that parish that were performing below the state average right after hurricane Katrina. Several of the schools in the OPSB system were highly selective schools before Katrina and continue to be highly selective. So even though the poverty rate for the OPSB system now stands at 67%, the school system ranks 6th in the state in academic performance. However, almost all the other school systems in the state that are not selective perform exactly as would be predicted if the rate of poverty were the only factor affecting school performance.

Based on this data and similar data collected over the years, this researcher and many others have concluded that the present school rating and grading system used in Louisiana and in several other states is extremely misleading and can have damaging results. In Louisiana, the school rating system has resulted in the state takeover of many schools. As this blog has pointed out as recently as June 16 and June 19, the state takeover and conversion of schools to charters has not improved instruction. All of the schools except one taken over in East Baton Rouge, Pointe Coupee, St Helena and Caddo are still rated as failing. These schools and many of the schools in the New Orleans Recovery District have been subjected to the disruption of multiple management changes. Some schools are closed and their students and parents required to fend for themselves in finding a new school when none of the remaining choices are highly rated. A few schools in the RSD (such as the KIPP schools) have been successful in selecting the more privileged students and in removing low performers so as to be rated relatively higher. Such practices do not empower parents nor do they serve the most at-risk students.

But the most damaging effect of the present school grading system in my opinion, is to make schools serving the neediest students unattractive to the most qualified and possibly the most effective teachers. What teacher in Louisiana wants to teach in a school that is constantly berated and scorned for having low performance. What highly effective teacher wants to risk her/his reputation and career by committing to serve the most needy students? Fortunately there are a few teachers that are willing to do this thankless job, but there will never be enough good teachers to fill this critical need as long as Louisiana falsely labels schools and teachers as failures.
2014 Percentile Ranks for Percent of Students Basic and Above (LEAP and iLEAP state testing) and 2013 letter grade
 






 





Site Name % of students on free or reduced lunch Letter grade assigned for 2013 Ranking for % proficient for 2014 2014 Basic or Above (Percentile Rank)
CAMERON PARISH 42.7% B 19 79
ZACHARY COM. SCHOOL DIS. 44.3% A 1 100
ST. TAMMANY PARISH 48.0% A 9 89
BOSSIER PARISH 48.1% B 13 80
ASCENSION PARISH 48.5% A 10 87
LIVINGSTON PARISH 50.0% B 8 89
CENTRAL COM. SCHOOL DIS. 50.6% A 2 97
WEST FELICIANA PARISH 52.3% A 7 89
ST. CHARLES PARISH 53.0% A 4 95
BEAUREGARD PARISH 53.6% B 20 71
PLAQUEMINES PARISH 54.5% B 3 97
LASALLE PARISH 57.2% B 15 75
VERMILION PARISH 57.5% B 21 75
VERNON PARISH 58.1% A 5 94
JEFFERSON DAVIS PARISH 58.4% B 14 80
OUACHITA PARISH 59.4% B 16 80
CALCASIEU PARISH 59.4% B 17 75
LAFOURCHE PARISH 60.6% B 25 65
LINCOLN PARISH 61.1% B 28 49
LAFAYETTE PARISH 61.4% B 32 56
ALLEN PARISH 62.5% B 12 84
WEBSTER PARISH 62.9% C 45 38
DESOTO PARISH 63.3% B 31 69
JACKSON PARISH 64.5% B 42 35
ASSUMPTION PARISH 66.1% B 38 49
CADDO PARISH 66.5% C 54 21
ACADIA PARISH 66.7% B 37 56
CALDWELL PARISH 66.8% C 35 49
ORLEANS (WITHOUT RSD) 67.0% A 6 89
GRANT PARISH 67.2% B 23 65
TERREBONNE PARISH 68.1% B 27 65
BIENVILLE PARISH 68.1% C 52 45
ST. JAMES PARISH 68.7% B 26 60
SABINE PARISH 69.7% C 24 60
RAPIDES PARISH 69.7% C 36 56
IBERIA PARISH 70.2% B 33 49
WEST BATON ROUGE PAR. 70.3% B 30 63
WINN PARISH 71.2% B 34 38
CATAHOULA PARISH 72.8% C 39 45
NATCHITOCHES PARISH 73.1% C 59 19
ST. MARY PARISH 73.9% B 29 63
ST. MARTIN PARISH 75.2% B 43 43
WEST CARROLL PARISH 76.3% C 18 71
EVANGELINE PARISH 76.3% C 22 71
JEFFERSON PARISH 77.1% B 40 49
TANGIPAHOA PARISH 77.4% C 55 30
CONCORDIA PARISH 77.7% C 47 30
ST. LANDRY PARISH 78.0% D 51 21
UNION PARISH 79.0% C 65 15
CITY OF MONROE SCHOOL DIS. 79.7% D 46 21
CLAIBORNE PARISH 80.1% D 68 8
EAST BATON ROUGE W/O RSD 81.9% C 49 30
CITY OF BAKER SCHOOL DIS. 82.4% D 70 5
FRANKLIN PARISH 82.6% C 56 21
RICHLAND PARISH 83.2% C 62 16
MOREHOUSE PARISH 83.5% D 63 10
EAST FELICIANA PARISH 83.7% C 58 21
ST. BERNARD PARISH 84.0% A 11 86
EAST CARROLL PARISH 84.1% C 66 19
AVOYELLES PARISH 84.2% C 64 13
WASHINGTON PARISH 84.3% B 44 45
IBERVILLE PARISH 84.6% C 41 38
POINTE COUPEE PARISH 85.0% C 60 30
RSD—NEW ORLEANS 85.2% C 61 17
TENSAS PARISH 86.4% D 69 9
RED RIVER PARISH 87.8% C 57 10
MADISON PARISH 88.1% F 71 4
CITY OF BOGALUSA SC. DIS. 94.7% D 67 5
ST. JOHN THE BAPTIST PAR. 97.5% C 50 38
RSD—BATON ROUGE 97.5% F 73 2
ST. HELENA PARISH 97.5% F 72 1
RSD—LOUISIANA 97.5% F 74 0

4 comments:

Kimmy G said...

There are those who say because St. Bernard ranks so high in poverty but got an A rating...this statistical anomaly can happen in other poverty stricken districts. However, that completely ignores the laws of statistics that the anomaly can become the new standard.

Michael Deshotels said...

Kimmy, that is a very astute observation. I also noticed that St. Bernard is outperforming districts with similar poverty. But the fact that there is one outlier does not negate the conclusion, nor does it mean that the result can be easily replicated. Also, we are not saying that other factors cannot affect school performance. We are simply saying that poverty is the dominant factor in almost all cases.

Anonymous said...

Great article!

Bridget said...

This is the kind of reporting we should be seeing and hearing about in our state and national media. Thanks Mike.
Also, I would venture to guess that there may be more to the story on St. Bernard parish poverty level reporting than meets the eye. When one digs deeper, there is usually some sort of explanation for the outliers. Maybe someone from that parish could shed more light on it??? For example, in my parish there is a fishing community that is high poverty, but they earn most of their living through cash sales of ther catch, therefore under reporting of income is probably going on there. Just a thought.