Tuesday, July 22, 2014

Another round of the miracle-gains game

While reporting a recent article on Sugar Creek Charter School's plans to add high school, I was dismayed to see the test-score reporting by UNC Charlotte's Urban Education Collaborative.  A 37-page report from the collaborative,  which is part of the College of Education,  bases its claims for  "extraordinary outcomes in public education"  on the school's proficiency gains between 2008 and 2012.

Source: SchoolwiseCharlotte.org
Not surprisingly, you don't see that number plunge to 40 percent in 2013.  And if the researchers had included that shocker,  they'd certainly have explained that North Carolina introduced new exams with a higher bar for passing.  They'd have noted that most schools across the state saw pass rates plummet,  with the biggest drop among schools such as Sugar Creek that serve mostly low-income and minority students.

Yet nowhere do these researchers, who are part of a partnership with Sugar Creek known as Schoolwise,  explain that scores also plunged statewide in 2008,  when North Carolina introduced a tougher reading exam. And that they rose sharply in 2009,  when the state started giving students a second chance to pass.  The curve depicted for Sugar Creek is common to most N.C. schools  --  again,  with the biggest plunge-and-rise among schools serving kids who traditionally struggle to reach grade level.

I've called Charlotte-Mecklenburg Schools out on playing the same game in the past.  I shudder to think how many national experts believe schools and programs across our state are successful based on big gains since 2008.  Charts like this are a great marketing tool,  if not exactly a testament to integrity in reporting.

When testing changes,  year-to-year comparisons carry little meaning.  At that point,  the best bet is to see how a given school,  district or group of schools compares with similar students.  As I noted in my article,  such comparisons indicate Sugar Creek is doing well compared with state and CMS averages,  though the latest numbers are unlikely to inspire breathless praise. (For careful readers,  the numbers I cite,  from N.C. school report cards,  represent the percent of students who passed both reading and math tests;  that's different from the composite score based on reading,  math and science exams.  Both are legitimate ways to measure proficiency.)

CMS used to do a good job of this when officials evaluated programs such as strategic staffing.  The studies were sometimes buried online,  but they existed.  Unfortunately,  a reader recently pointed out to me that the CMS research link,  which I'd kept in the rail at the right of this blog,  is now dead.  If there's a new one I can't find it.

Chance Lewis,  director of the Urban Education Collaborative,  says he's working on just such a comparison for Sugar Creek,  which is the collaborative's partner.  He and I agree that the challenge is figuring out the fairest comparison for the charter school,  which serves grades K-8.  Do you look at CMS neighborhood schools or at magnets?  Focus only on other K-8 schools,  or on elementaries and middle schools?  Do categories such as "African American" and "economically disadvantaged" give a true apples-to-apples look?

Results for 2014 are due out later this summer.  We already know they'll be up,  because the state changed the scoring system to allow more students to pass.  With all the uncertainty about Common Core,  it's hard to tell what we'll get in coming years.  Here's my forecast:  By 2017,  we're going to see lots of charts showing that schools have made amazing gains since 2013.


Anonymous said...

How would you like your educational statistical baloney sliced? Would that be number 1, 2, 3, or a Painter's Steak thickness? Wedges? Cubed? Julienne? Casing Removed? The obfuscation never stops.

Unknown said...


I suppose that if we were to see a repeat of forgetfulness, NC school administrators will also not want to point-out that “awesome” achievement was accomplished while teacher pay was stagnant.

Otherwise, what better logic could there be for not raising teacher pay?

Bolyn McClung

Anonymous said...

I'm shocked, shocked I tell you, that statistics would be manipulated to present a shiny rosy picture about charter schools while also boosting UNCC's credibility. But I'm certain the pro-charter crowd will have the perfect reply to make it all better.

Anonymous said...

Scores go up the second year for new tests for one simple reason, the same test is given for several years.....at least it was with the old state eocs. Imagine if the SAT were the same for five years in a row how much higher the scores would be. One way or another information trickles out about what is on the tests. Students talk and blog about it. Every school has a thing called read alouds in which teachers read the tests to students with disabilities. So teachers are sworn to not look at the tests are slowly, reading the tests legally! I will barely mention that deceiptful teachers can photograph the entire test in seconds with a camera. Failure to make new tests yearly guarantees higher results, period.

Anonymous said...

regardless, unless things have changed, a teacher's scores THIS YEAR are measured against his/her scores LAST YEAR for growth, but that ends up comparing 2 different groups of kids, not the growth of ONE set of kids. I would assume schools are measured the same way. It's all fog and mirrors.

Wiley Coyote said...

The really sad thing in all this?

Many in the public will believe this bullspeak.

As Bolyn, myself and others have pointed out, all these miraculous gains have been made with the same teachers we have now who haven't had raises in years. Yet the Obama administration is pushing to have "high quality teachers" in all urban schools?

My God, if we had those high quality teachers in our urban schools? The gains would top 100% and keep going to infinity based on the criteria being used to track gains in this report.

Interesting how such wonderful gains have been tracked, but still no data from the 14+ year $20 million dollar per year boondoggle called Bright Beginnings.

You can't make this stuff up.

Oh, wait. Yes you can. Duh, what was I thinking?

Anonymous said...

May wonders never cease...

While still contemplating opportunities, I will most likely teach 5th grade Singapore math this year at a successful charter school that made large "gains" last year after switching to this method of teaching.

I look forward to comparing notes with Shamash because it's important to know how to calculate the area of a composite figure before executing a respectable tour jete :)


Anonymous said...

I've always been very skeptical of miraculous increases in test scores from one year to the next, whether the scores be for charter or public schools. And I've always been surprised how many people keep looking for the silver bullet that will provide these overnight increases. Kids who are very poor readers need a long, steady, focused course of instruction to become proficient readers. Maybe gimmicks that teach to the test (or manipulate the outcomes) can result in year to year higher test scores for these students. But are they really learning to read in a meaningful way that that will enable them to become successful adults? I doubt it. I spent a year tutoring third graders at a high poverty school where most of the kids stumbled through the reading of a basic passage. Yet what was the teacher required to do--drill, drill, drill on finding the "theme", the "plot", the mood, etc, which is what the test most likely emphasized. These kids needed basic reading instruction and practice so that they could come to enjoy reading. If that would happen I believe test scores would gradually rise, and that gradual rise would be believable.

Anonymous said...

This article would be more informative if it pointed out that Sugar Creek's numbers are still above district and state averages. A quick search on the DPI website will also reveal scores of traditional public scores in the same zip code. I will remind you that Sugar Creek has a similar socio-economic breakdown.
Druid Hills- Math 13.6% proficient, Reading 10.5% proficient
Turning point-less than 5% proficient in both categories
Walter G. Byers 14% proficient in math, 16.9% in reading
Highland Ren. 26% in math and 40% in Reading

One can also measure the percentage of fall from one year to the next. Sugar Creek's fall was not nearly as drastic as other schools inside of Mecklenburg County and across the state. It also makes perfect sense to only include statistics that are comparable. Next year the graph should include numbers for 2013 and 2014.

Anonymous said...

12:47, you are correct in saying that comparison is important but incorrect in thinking I didn't include it. I put that comparison in the news article, which is linked from the first blog sentence, and discuss it in the fifth paragraph of the blog post. I am not disputing that Sugar Creek has had success, just questioning the use of numbers that exaggerate that success. I've raised the same issue with CMS and national reports that use 2008 as a starting point without context.

Anonymous said...

I don't mean to be disrepsectful to anyone either at Sugar Creek or any of the other schools mentioned in this article. I suspect Sugar Creek has a distinct advantage over all the local schools mentioned. The mere fact that those children are attending a charter indicates to me, those parents are probably more involved/concerned with their child's education. Involved enough to explore a different option for their child and are probably more involved at home when reading and doing homework. I think we can all agree that parental involvement is vital to a child's success.

Anonymous said...

Teachers are less important than:

Coal Ash

STRIKE !!!!!!!!!!!!!!!!!

Cornelia said...

And in related news, a study just released by the University of Arkansas shows that charter schools are more cost effective than public schools, primarily because charters don't have the bloated central staffs, a shining example of which is CMS.

Anonymous said...

If you think this is bad, wait until someone uncovers the cheating and misrepresentation of the data by project LIFT!

Anonymous said...

Of course Charters are more efficient and cost effective in regards to all resources, transportation, lunch, etc...
CMS is a mess.

Anonymous said...


I couldn't agree with you more.

And I repeat, a school is nothing more than a reflection of the community of people it serves.

In my perfect world, we'd prioritize educating and holding up young unmarried women with children instead of incessantly blaming the educational establishment for failing to provide "effective" teachers. Enough already.


Anonymous said...

Cornelia, you might want to finish the NPR story which indeed said that but also included comments on methodology that Ann mentioned above from Charter school advocates who questioned the entire testing scenario and the validity of the Arkansas conclusions.

Wiley Coyote said...

Clean slate for ninth-grade athletes gets CMS approval...

Way to dumb it down BOE....

Staying true to your course.

Anonymous said...

Ann...here's the link to the CMS Research & Evaluation reports:

Also, just because Sugar Creek charter is in/near the same zip code does not necessarily mean that the students who attend that school are similar to those that attend CMS schools in that area.

Rigorous research/evaluation work, short of a randomized trial which isn't feasible, might involve some form of matching to create a synthetic comparison group comprised of CMS students that are similar to the Sugar Creek students...with longitudinal tracking.

Anonymous said...

with regards to charters being more cost effective, I can understand why. Traditional public schools systems requirement far too many advanced degrees. A 4 year degree should suffice.

Secondly Alicia, while you agreed with my post from earlier today, I completely agree with your comment in that schools are a reflection of the community in which they serve.

To confirm this theory, all one has to do is look at the high performing schools versus the low performing schools and then look at who attends those schools. We don't need any clever graphics to understand this.

Anonymous said...


Our state legislator agrees with you that advanced degrees in education shouldn't be rewarded after removing financial incentives this year for holding a master's degree. I believe our state may reinstate master's pay at some point but only if the degree aligns with the subject taught meaning I'd be paid more to teach dance than fifth grade math and ensuring there's no point in killing myself to earn dual certification in science. Of course, I can't remember a recent Charlotte school superintendent that didn't hold a doctorate in education.

On the flip side, charter schools can hire a certain percentage of unlicensed teachers (who aren't required to take courses like "Diversity in Education") and freely adjust salaries based on a number of factors. However, in NC a teacher with or without a master's degree is still financially better off teaching at a charter school or a regular public school in SC. This past April, a charter school in Rock Hill SC (with a low FRL population) sent out a recruiter to solicit new elementary education graduates in the Charlotte area with a starting salary incentive close to $40,000. Yep.