Saturday, December 17, 2016

Can High Poverty Schools Beat the Odds?

Part II: An Analysis of School Performance Related to Family Poverty.
 

Unintended negative consequences of Louisiana’s school grading system

Does the grading of schools based on student test scores produce higher student performance?

My analysis reveals that the evidence is building each successive year that such pressure produces more negative consequences than positive ones.

I suggest readers spend a few minutes reading the two recent articles by Danielle Dreilinger at nola.com here and here, to examine the rise of rampant cheating on state tests caused by the intense pressure to produce higher test scores.  In addition, here is another article about alleged cheating at SciTech Academy. These are exactly the kind of cheating by educators that resulted in educators going to jail in Atlanta Georgia.

Time after time, it has been found that charter schools in the New Orleans Recovery District have used various forms of cheating and test question teaching to artificially raise their school performance scores. Time after time school performance scores in the RSD have dropped like a rock the year after instances of cheating are exposed. The graduation rate of the RSD dropped by almost ten percentage points when the LDOE clamped down on the misreporting of dropouts as transfers.

Here is a study by Stanford University that shows that grading and closing schools in New Orleans neglects and underserves the students that are most at risk.

Yes, it is very clear that one of the major outcomes of grading schools and evaluation of educators using student test scores produces cheating! The articles above show that cheating includes educators changing student test answers and educators making copies of test questions so that the answers can be taught to students before they take the state tests.

The problem of cheating is compounded by the long-standing policy of the Department of Education that allows school districts or charter groups to investigate themselves when allegations of cheating arise. Such a policy probably makes it more hazardous for whistleblowers reporting cheating than for the propagators of cheating.

To better understand the pressure that produces cheating, let’s look more closely at the relation between school ratings and student poverty.

There are only 10 public school districts in Louisiana with less than 60% of their students qualifying for free or reduced price lunch.  (See the spreadsheet in Part I of this investigation) Free or Reduced Lunch (FRL) has become a standard measure of the degree of poverty in public schools nationwide. Louisiana is one of the highest poverty states in the nation. All public school systems in Louisiana except one, have more than 50% of their students on FRL. That one exception is the Zachary Community School system which has 45% of its students on FRL It just so happens that the Zachary Community School system is the highest performing school district in Louisiana. It is rated as an “A” school district.

Out of the 10 school districts with fewer than 60% FRL students, 9 are also rated as “A” school systems. Is this a coincidence? I don’t think so.

One of the highest poverty districts, which seems to have cracked the code for producing higher test scores just happens to be the district described in one of the Dreilinger articles linked above. The state inspector general has investigated that district for alleged cheating. The IG determined that cheating was a serious problem in the district, ("extensive violations of test security policy") but apparently took no corrective action.  The LDOE has allowed the highly questionable test results and the improved district grade to stand. Except for the Dreilinger article in nola.com, which is hundreds of miles away from the district implicated in cheating, no local news media has even bothered to inform the public in the district of the alleged cheating.

The State Department of Education has also made it easier for school systems to appear to have improved performance by the lowering of cut scores on state tests. There is also a state policy of curving school performance scores in the last two years that keeps scores artificially high. As a result, no school system was assigned an “F” grade in 2016.  At the same time the ranking of Louisiana compared to other states on the NAEP tests has dropped even lower.

It is obvious that rating and grading schools using primarily student test scores reveals mostly the level of poverty of the students attending each school. Does such a rating system really tell us something about the quality of instruction? What do you think?

There is also strong evidence that the grading of schools based on student test scores results in neglect of students with disabilities in some schools because such students have little effect on improving school performance scores. One administrator was quoted advising teachers not to waste their time on such students even though the school had contrived to receive extra funding for more students with disabilities.

I must argue forcefully that the grading of high poverty schools places an unfair stigma or assumption of blame on the teachers and administrators of such schools. The general public automatically assumes that students in so called “failing” schools are not getting good instruction and that students in “A” schools are getting the best instruction. Surprisingly, Herb Basset has reviewed data that indicates mixed results, or at best, very slight improvement for students transferring to higher grade schools.

Another unintended consequence of the grading of schools using student test scores frustrates the mission of schools designed to address the needs of handicapped or troubled students. The schools for the deaf and visually impaired and a charter school for dyslexia are all rated F by our system of rating schools. This is outrageous considering the valuable services provided by these schools. In addition, practically all the alternative schools addressing the needs of suspended students and potential dropouts are rated F. The rating system has no relation to the purpose of those schools and it serves only to smear the reputations of their highly dedicated educators.

I believe that school grades tell us almost nothing about the quality of instruction. Poverty factors seem to be the dominant force in determining a school’s grade. So why do we still insist on stigmatizing the teachers and administrators that serve students struggling with the negative effects of poverty?

I happen to live in the Zachary Community School system. All my children and most of my grandchildren have attended the Zachary system, which continues to receive the highest ratings in the state. I have first hand knowledge that it is indeed an excellent school system!

I personally know many of the teachers and administrators in the Zachary system and can attest to the fact that they are superior educators and we are lucky to have them in my community.

But another negative unintended consequence of the school rating system, in Louisiana is that it makes it easier for top systems like Zachary to continue to attract the very best educators from any of the systems that are rated near the bottom. What possible benefit is there for a top educator to go to or remain in one of the “D” rated systems? So the rating system automatically drives top educators away from the students that need them most. This is the ultimate insult to our students and educators caused by the school grading system.



Thursday, December 15, 2016

The Secret Behind the Best School Performance Scores

Part I: A look at the data relating school performance scores and school letter grades to the rate of family poverty in Louisiana public schools

It took 3 successful public records lawsuits against the Louisiana Department of Education and the successful defense of a 4th lawsuit against me by the LDOE to get the accurate data that can be used to compare school performance scores to family poverty in each public school district. Now any person interested in the possible effect of poverty on school performance can view the results of this comparison.

The following chart was created by my son Donny, who is a graduate of the LSU school of business and who holds an MBA degree from Mississippi State. Donny majored in Quantitative Business Analysis and does statistical analysis for one of the largest companies in Baton Rouge. It turns out his skills can be applied to analyze our Louisiana public education data.

The graph below compares poverty of families of children enrolled in each school district in 2016 to average School Performance Scores (SPS). Poverty is determined by the percentage of families qualifying for free or reduced lunch or a comparable measure of wealth. The correlation coefficient comparing school performance to poverty is -0.826. The negative sign indicates an inverse or opposite correlation of school performance to the percentage of families that are economically disadvantaged. This relation is very statistically strong.


The  Excel chart below is a ranking of all public school districts in Louisiana based on poverty. The ranking starts with the school systems serving the smallest percentage of students living in poverty. This ranking is the data that was used to produce the graph shown above. Notice that even though the ranking is not a ranking by school performance scores, it ends up being almost the same. That's because of the strong correlation of family poverty based on free or reduced priced lunch to school performance. (Note: The gap in the chart is simply a page break.)
The school districts near the top of the chart are those with the smallest percentage of families struggling with poverty. Note that they are also the school districts with the top letter grades. These rankings demonstrate an uncanny relation between school performance and poverty in Louisiana. The smaller percentage of families living in poverty is the secret behind the best school performance scores!

I believe this data calls into question Louisiana's policy of grading public schools without considering the negative effects of poverty on SPS and therefore school grades. Apparently all we are doing is grading our schools based on the poverty of their students. What is the purpose of a system that classifies community schools as failing just because they happen to serve a large percentage of students that are struggling with poverty?

Think about it. Is this a fair way of rating our schools?

Part II will look at specific cases of school performance and examine the unintended consequences of our obsession with rating and grading schools using test score results.

Michael Deshotels