The nation’s charter schools, including those in California, have made “slow and steady” progress over the past four years, with students in nearly a quarter of charters now outperforming their traditional school peers in reading and, on average, catching up to them in math, a group of Stanford researchers reported. The study also found that charter schools excel in teaching poor minority students and English language learners.

The study by the Center for Research on Education Outcomes, or CREDO, updates a 2009 report on charters in 15 states and the District of Columbia. That study’s major finding, which charter school critics have often cited to counter states’ and the Obama administration’s push for charter school expansion, was that 37 percent of charters produced academic results that were worse than traditional public schools in math, while only 17 percent performed significantly better, with no significant difference for the rest. There was little distinction in results between charters and traditional public schools in reading.

Charter schools now serve an estimated 2.3 million students in 6,000 schools – 4 percent of the nation’s school population. In California, nearly 8 percent of the state’s students – 471,000 – attend more than 1,000 charter schools.

Showing what the study called "steady" improvement from four years ago, students in 25 percent of charter schools did better than comparable traditional public schools in reading, compared with 19 percent that did worse. In math, 29 percent of charter schools did better and 31 percent did worse. Four years ago, only 17 percent of charter schools did better in math, and 37 percent did worse, with 46 percent about the same. Source: CREDO National Charter School Study (click to enlarge).

Showing what the study called “steady” improvement from four years ago, students in 25 percent of charter schools did better than comparable traditional public schools in reading, while 19 percent did worse. In math, 29 percent of charter schools did better and 31 percent did worse. Four years ago, only 17 percent of charter schools did better in math, and 37 percent did worse, with 46 percent about the same. Source: CREDO National Charter School Study.

The latest study analyzed state standardized tests results in math and reading in 25 states, Washington, D.C. and New York City. Among the major findings:

  • In reading, 25 percent of charters performed significantly better than district schools and 19 percent performed worse, with no significant difference for 56 percent of schools;
  • In math, 29 percent of charters performed significantly better and 31 percent performed worse, with no significant difference for 40 percent.

Students in charter schools states gained an average of eight days more per year in reading than their peers in other schools, the study said. That’s up from the 2009 study, which found that charter students in 16 states had seven fewer days of reading instruction each year, compared with students in other schools.  In math, where charter students’ scores equaled 22 fewer days of learning in 2009, the learning now is the same as with peers in traditional public schools, according to the study.

But there were great differences in results among states and among subgroups of students. Charter school students outperformed their traditional school peers in reading in 16 of the 27 states studied and underperformed students in traditional schools in eight states, with performance about even in three states. In math, charter schools in 12 states were stronger than traditional public schools and weaker in 13, with two having similar growth.

California was not among the 11 states where charter school students outperformed their peers in traditional public schools in both reading and math. In California, students in traditional public schools did better in math with seven extra learning days. But in reading, charter school students scored the equivalent of 22 extra learning days – nearly triple the national average of eight days’ improvement (page 61 of main report).

The CREDO analysis of the impact of charter schools found great differences among groups of students attending charter schools versus similar students in nearby traditional district schools, with scores on state standardized tests translated into  gains or losses in learning days. White students in charter schools did significantly worse – the equivalent of 50 days lost in math and 14 in reading, while English language learners in charters gained 36 days in reading and math. Source: CREDO's 2013 National Charter School Study.

The CREDO analysis of the impact of charter schools found great differences among groups of students attending charter schools versus similar students in nearby traditional district schools, with scores on state standardized tests translated into gains or losses in learning days. White students in charter schools did significantly worse – the equivalent of 50 days lost in math and 14 in reading, while English language learners in charters gained 36 days in reading and math. Source: CREDO’s 2013 National Charter School Study. (Click for clear image)

CREDO found that white students in traditional public schools substantially outperformed their peers in charter schools, as did Asian students in math; there was no difference for Hispanics overall. Poor students did better in charter schools (an extra 14 days in reading, 21 days in math) as did black students (14 extra days in math and reading) and English learners (36 extra days in both math and reading). Charter schools have been criticized for underenrolling special education students, especially students with moderate to severe handicaps. Special education students attending charter schools did better in math (14 extra learning days) than peers in district schools and about the same in reading, the study said.

A further breakdown of subgroups revealed that poor Hispanic students did better in charter schools while non-poor Hispanic students did better in district schools. Black students in poverty substantially outperformed their peers in district schools (36 extra days of learning in math, 39 in reading), while results were indistinguishable for non-poor black students.

Breaking down the results of student subgroups further, CREDO found that students in black and Hispanic poor students and especially poor English learners made substantial gains compared with like students in traditional public schools. Test results were translated into learning days gained and lost (click to enlarge).

CREDO found that black and Hispanic poor students and especially poor English learners made substantial gains compared with like students in traditional public schools. Test results were translated into learning days gained and lost. (Click for clear image)

CREDO’s analysis of student subgroups in California and other states could show the impact of high-performing charter management organizations like Aspire Public Schools, Rocketship Education and Los Angeles-based Alliance for College Ready Schools, which target low-income, minority children. But it won’t be available until late summer or fall, according to CREDO.

In a statement Tuesday, Jed Wallace, president and CEO of the California Charter Schools Association, said, “To understand the story of charter performance, it is important to move beyond a discussion of averages and look at how charters are performing with key subgroups of students. … CCSA’s own research highlights that historically disadvantaged populations are achieving at high levels in California’s charter schools.” (See a previous article on the state Charter Schools Association’s 2012 detailed analysis of student performance.)

While the report concluded that “moving the needle for 2 million students (the number in charter schools) is no small feat,” it also stated: “The charter sector is getting better on average, but not because existing schools are getting dramatically better; it is mainly driven by opening higher-performing schools and by closing those that underperform.”

Between the first and latest reports, 8 percent of charter schools – assumed to be underperformers – were closed.

“The results reveal that the charter school sector is getting better on average and that charter schools are benefiting low-income, disadvantaged and special-education students,” said Margaret Raymond, director of CREDO. “As welcome as these changes are, more work needs to be done to ensure that all charter schools provide their students high-quality education.”

The report suggested that states do more to close down poorly performing charters and identify models that are likely to succeed before granting charters. It praised advocates, among them the California Charter Schools Association, that have taken the lead in encouraging regulators and lawmakers to revoke charters of non-performers. “We recently refined our accountability initiative, enhancing minimum performance expectations” for schools whose charters are up for renewal, Wallace said in his statement.

But other advocates, including Eric Premack, who directs the Sacramento-based Charter Schools Development Center, have warned that shutting down schools based on test scores would discourage some charters from targeting students at risk of dropping out and would be too narrow a measure.

CREDO’s methodology examined the records of 1.5 million charter school students, covering 95 percent of students in the states studied, and matched each with the results of a “virtual twin” – a student with similar demographics and test scores prior to enrollment in a nearby district school.

Filed under: Charter Schools, Featured, Reporting & Analysis · Tags:

Comment Policy

EdSource encourages a robust debate on education issues and welcomes comments from our readers. The level of thoughtfulness of our community of readers is rare among online news sites. To preserve a civil dialogue, writers should avoid personal, gratuitous attacks and invective. Comments should be relevant to the subject of the article responded to. EdSource retains the right not to publish inappropriate and non-germaine comments.

EdSource encourages commenters to use their real names. Commenters who do decide to use a pseudonym should use it consistently.

Leave a Reply

Your email address will not be published. Required fields are marked *

 characters available

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

  1. Eric Premack says:

    In addition to a large number of serious substantive and methodological problems, the CREDO study and many similar efforts all suffer from a profound and fatal flaw: They assume that the goal of chartering is to improve standardized test scores.

    Bad assumption.

    A large proportion of charter schools care little about standardized testing–or only enough to pass muster when it comes to charter renewal–and no more. A large number of charter schools are founded by groups who are specifically seeking to escape the test-crazed focus that plagues many conventional public schools. Many have distinct missions that focus on instructional goals that aren’t touched by state testing, including fine/performing arts, foreign languages, technology, career/vocational prep, etc.–along with deeper forms of English/language arts, math, social studies, and science that cannot be assessed with a #2 pencil and a bubble form.

    Research of this sort does virtually nothing to illuminate the charter debate.

    1. Gary Ravani says:

      How ‘true,” Eric. Take one of the better know charter organizations “Aspire,” for example. Certainly no one has heard of them hyping test scores.

    2. Rick Pratt says:

      Eric, join the crowd. Most people seem to think that the goal of public education in general–not just of charter schools–is to improve standardized test scores. The test-crazed focus of which you speak is imposed on schools from the federal government on down, and most educators would love to get away from it. And if a public school does not improve test scores, then one of the “fixes” is to convert it to a charter, which, as you note, may “care little about standardized testing.” Go figure!

      1. Manuel says:

        Rick, how can you blame them? It says so right in the EdCode:

        52050.5.(k)(6)Utilize the state infrastructure to support schools, school
        districts, and county offices of education in their efforts to improve
        pupil achievement and progress.

        The only problem is that if a school does not care about its scores, then its API will be low and will then be moved into “Program Improvement” status. After a while, it could have its charter revoked or, horror, have the parent trigger pulled on it. So, yeah, I’d say that a charter school can dismiss its scores at its own peril.

    3. el says:

      Great point.

      But it goes to show that if they don’t recruit high-scoring kids, they won’t get to keep those programs, no matter how valued or valuable they might be. So what if the kids are speaking fluent Japanese.

    4. Manuel says:

      That might be so in many charters, but the ones I hear about in Los Angeles are about something else.

      The LAUSD schools, mostly high schools, I believe, that have gone conversion independent charter did so because they wanted to stop financing the high overhead of LAUSD (the latest published budget has $2.8 billion going to school sites, $1.5 billion to Special Ed costs, and $2.2 billion going to “overhead”), and the ones who have gone conversion “affiliated” to get autonomy plus $380/kid from the block grant.

      The schools that are independent startups are, in its majority, run as increasing scores machines, at least according to what I read in their web pages and what I occasionally hear from people that work in them. More tellingly, many of these schools are run in poor areas and have low enrollments.

      This brings another problem: the achievement gap is not due to ethnicity but to poverty (and, interestingly, it happens in other countries, too, so it is not “cultural” as we say here in the US). It is therefore not surprising to me that charters in poor areas cannot do significantly better than the local public schools. As for what the schools (and this also includes public ones) who have consistently high scores and yet serve poor students, I would hope that serious study is done to understand their best practices or to find out how they are gaming the system. (This recently came to light at LAUSD’s Hart Elementary, where teachers do one hour a day test prepping from October until the CSTs are administered. These kids are not learning. They are being trained to pass the test.)

  2. Rick Pratt says:

    John, I have to agree with el’s comment. This article is a good case study in how the public continues to be mislead about the quality of charter school vs. non-charter schools. The headline will be interpreted by the casual reader to mean “MOST charters outperform school districts in reading.” Many readers don’t get beyond the headline, which, with equal accuracy, could have read, “Only a quarter of charter schools outperform school districts in reading.” Also, take a look at the bullets describing the reading and math results. In both cases, some charters are described as performing “significantly” better, while others are described as being simply worse (but not significantly worse). There is a very subtle bias at work here, because the average person thinks that the word, “significant” describes the size of the difference and not the statistical probability that the differences are due to random chance. Hence, the takeaway for the typical reader is that the charter schools that are better are better by a wide margin, while those are are worse are only slightly worse.

    1. John Fensterwald says:

      Rick: When I worked at the Merc, I could deflect criticism of headlines by blaming a copy editor. I can’t do that now, so am always interested in readers’ responses to them.

      I trust that the readers of EdSource Today will go beyond the headlines — otherwise, why come to a policy site? — and not infer that “more” means “most” since the article goes to great length explaining the details of the CREDO findings. It is true, however, that some readers may see a headline on other sites, like Rough and Tumble, and not read the story beyond a sentence or two.

      I and others have cited the main finding of 2009 CREDO study so often I didn’t need to look up the numbers: 17 percent of charters produced better results, 37 percent worse results and 46 percent about the same. Had I been doing a post on the 2009 study four years ago, the headline would have been “More district schools outperform charters, including in California, in reading and math” — and then responded to charter school advocates who might have noted that the majority of charters — 63 percent — either performed as well or better than district schools.

      1. el says:

        I appreciate your response here. I think though that thinking that your headlines are not important is a mistake. You hope people will share your work, to get it more widely read, yes? In those cases your headline makes all the difference in whether it is shareable, how people react when they see the shared headline, and then whether they’ll elect to read the article.

        There is in fact a whole discipline and a great deal of discussion out there on how to write great headlines. You probably know this already. But the advice changes with some frequency.

        I’m going to say this honestly knowing that it’s not all that nice: to me, if I did not know you and your work and have a strong sense of your strengths as a journalist, I would assume this was a press release from CREDO that you just put your name on. I say that not because I think it is true, but because I really do respect your work.

        One thing you might consider adding to your reports on these studies is to ask yourself some critical questions about methodology and try to give readers a taste of where any holes in information might be. Don’t take their word for it that the study shows what they say it shows. For example, while they use “days of learning” as a substitute for “growth” as expressed in standard deviations to make it “more accessible” – the implication that those kids would have learned that topic if only they had 5 more days in the school year is probably not accurate.

        And you can probably create some boilerplate cut and paste URLs for the issue of comparing all kids in free & reduced lunch or special ed as if they were the same (frequently we see that schools marked as better performing have lots of reduced lunch and schools marked as poorer performing have mostly free lunch).

        A better, clearer headline might have been “Charter schools have improved performance since 2009.”

        Thanks for all you do.

        1. John Fensterwald says:

          Thanks for your thoughts, el.

          Since you are a faithful, thorough and thoughtful reader of the site, I take your comments seriously. Which is not to say I would change the headline but I hear you.

          I wish more readers would discuss the methodology. I will say that translating the results into learning days makes reporters’ jobs easier than having to discuss standard deviations.

          The study was peer-reviewed, and, in my eyes, has the credibility from having used similar methodology four years ago. I didn’t hear the same skepticism of lost and gained learning days then — at least from charter critics — because charters didn’t end up looking good in that study. Critics have been citing that CREDO study as if it were the Gospel.

          CREDO was frank in assessing the reasons for the higher charter scores — much of it because of closure of the least performing schools. Eric Premack raises valid points when he questions the use of test scores to determine the life and death of schools, but in California, those are the rules that the Legislature has set. In focusing on CST results, charters play by the cards they have been dealt. And some of the ones in the Bay Area that I watch — Aspire, Rocketship, Summit — have done exceedingly well by those measures.

          1. el says:

            You’re totally right that in general it’s a decently done study for what it purports to measure, and it’s likely as well done as the 2009. I look at their results and I feel pretty comfortable with them.

            Where we might differ is that I don’t think this study is particularly rosy for charter schools if your measure is improving CST scores. Better than they were – probably so.

            Some of the things I mentioned about poking holes are something any study will have in any field and I meant them to be more general rather than specific to this one study. It’s just the nature of research that it rests on your assumptions and that you can’t grok the study and what it means without digging into them a little. That’s a general critique of all journalists BTW. :-) My apology for not making that clearer.

            I would love to read more frank coverage of Aspire, Rocketship, etc to get a sense of what they’re doing, what is working for them, and what is not. I ask this not for grand policy reasons but because I am genuinely interested in bringing useful innovations into my district. I want to know where the quicksand was and where things really came together and sang. For example, I love this discussion of a pilot using Khan Academy: – because it had such wonderful nitty-gritty about things like how the Chromebooks worked for them and how the students unexpectedly researched their answers rather than risk breaking a “streak”. That’s what I want to know most about those charters – not their test scores.

      2. navigio says:

        Well, you did get the opportunity to write a headline for credo’s last study, though it was admittedly only 3 years ago, and it was actually for a federal study, but one that you said had come up with the same finding as credo’s. the headline was:

        “Low-income charter students outperform peers”

        Personally I ignore headlines because creating them is inherently biased.

        1. John Fensterwald says:

          Very attentive, navigio. I am pleased someone is using the library at TOP-Ed. :)

        2. CarolineSF says:

          I believe that the SVEF blog was openly pro-”reform,” while EdSource Today aims to be impartial journalism.

  3. CarolineSF says:

    How does this impact charter schools’ achievement?

    Special Report: Class Struggle – How charter schools get students they want
    Getting in can be grueling

    From the article:
    Charters are public schools, funded by taxpayers and widely promoted as open to all. But Reuters has found that across the United States, charters aggressively screen student applicants, assessing their academic records, parental support, disciplinary history, motivation, special needs and even their citizenship, sometimes in violation of state and federal law.

    “I didn’t get the sense that was what charter schools were all about – we’ll pick the students who are the most motivated? Who are going to make our test scores look good?” said Michelle Newman, whose 8-year-old son lost his seat in an Ohio charter school last fall after he did poorly on an admissions test. “It left a bad taste in my mouth.”

  4. navigio says:

    Give it up credo. For the good of our kids. The only reason the charter model will ever be any better at educating kids will be when charters have successfully destroyed the traditional public education system. And that will be nothing to celebrate.

    1. CarolineSF says:

      CREDO, as part of the Hoover Institution, is inherently in support of privatizing public education.

      1. John Fensterwald says:

        Caroline: Other than by marriage (Margaret Raymond, director of CREDO, is married to Eric Hanushek, a Hoover researcher), I see no ties between CREDO and Hoover. CREDO is affiliated with Stanford.

        1. CarolineSF says:

          Margaret E. Raymond is a research fellow at the Hoover Institution.

          At Hoover, Raymond serves as director of the Center for Research on Education Outcomes (CREDO), which analyzes education reform efforts around the country. CREDO’s mission is to improve the quantity and quality of evidence about the impacts of education innovations on student achievement in public K–12 education.

        2. CarolineSF says:

          And it’s listed here under the Hoover Institution’s “projects and programs.”

          1. John Fensterwald says:

            Thanks for the info, Caroline. I had missed that.

            1. CarolineSF says:

              That information appears to be downplayed, put it that way. It isn’t indicated on CREDO’s website, as far as I can tell.

  5. el says:

    When I look at the numbers in this article, I think the headline is a bit misleading. I see that more charter schools outperform public schools than did so in 2009. That’s great. But I don’t see that on average, that charter schools outperform public schools – on average, most are the same, around a quarter are better, and a fifth are worse. IE, the averages are about the same, but you see more variance in the charter schools.

    The whole feel of the language of the article is “yay charter schools” but honestly I don’t think the words add up to create a picture that they’re performing all that differently from the public schools. That’s kind of remarkable when you consider how different their approaches are and the advantages they’re supposed to have.

    In general, while charter schools do have special education students and low income students, most of the studies I’ve seen suggest that their populations of those more challenging students are measurably different from those in the general public school population.

    What I think is clearly the case is that pulling willing kids into environments where the majority of the kids are academically oriented is helpful. That’s something we should look at and celebrate. But, we need to be mindful that that leaves the original schools with a much more challenging and different problem to solve, if they’re left with a concentration of kids who aren’t on task and don’t have supportive families and are extremely stressed at home.

    1. Paul Muench says:

      Yeah, it would probably be clearer to title this report by the implications drawn in the executive summary. Something like, “Charter granting process improves even as the number of charter schools increase.” The implicit assumption is that charter schools improve education by being a school system apart rather than as a means of spreading innovation to all schools. Does anyone see that the study has any imterest in that aspect of charter schools?