The Education Gadfly Weekly

Yesterday, Texas became the thirty-fourth state (in addition to the District of Columbia and Puerto Rico) to receive approval for its plan under the Every Student Succeeds Act. The rubber-stamping of these documents has become so routine that even the education trade press doesn’t bother to write more than a few paragraphs about each occasion.

But as with everything about Texas, this approval is a big, big deal. That’s because buried inside Texas’s plan is a hidden treasure for charter schools, one that has the potential to significantly increase the amount of federal money going into charter startups and replications.

On pages thirty-one and thirty-two of its plan, Texas describes how it intends to spend its 7 percent Title I set-aside for school improvement. Note in particular the items that I’ve bolded for emphasis.

A portion of the seven percent set aside will be distributed to LEAs with comprehensive or targeted schools via a series of competitive grant programs. These grant programs will require the applicants submit their district- and campus-level improvement plans, which will outline the use of evidence-based strategies. TEA will give priority points to LEA applications that ensure the identified campuses have the operational flexibility necessary to successfully implement plans. These grants may incentivize the following types of school improvement and transformation actions:

• Restarting the school in partnership with a high-quality school management organization or converting it to a charter school;

• Redesigning the school, including replacing the school leadership team with a new team, implementing a new instructional model, or related activities aimed at better serving the needs of the students;

• Replicating an existing successful school model into an identified school, including as a charter school;

• Closing the identified school and consolidating the students into a higher performing or new school, whether charter or district managed;

• Creating new schools, whether district or charter, to provide students in identified schools with new and better education options. TEA will ensure these new schools guarantee and prioritize access to students currently attending the identified school(s);

• Increasing access to effective teachers or leaders or adopting incentives to recruit and retain effective teachers and leaders;

• Building the instructional leadership capacity of school leadership teams to understand and implement evidence-based strategies such as data driven instruction;

• Building district capacity to analyze campus performance and make and execute strategic decisions about school improvement or transformation actions; or

• Grouping identified schools together in a zone or cluster and providing those schools with operational flexibility and additional school improvement supports.

What this means is that Texas can use its Title I resources to start new schools, rather than just work to turn around low-performing ones. Given the weak track record of turnarounds and the overwhelmingly positive results of urban charter schools, this makes a ton of sense. But some have argued that ESSA doesn’t allow the Title I set-aside to be used this way; Secretary DeVos and her team clearly disagree. And now that they gave the green light to Texas, any state in the country can follow suit.

The appropriations bill that President Trump signed last week provides $15.8 billion for Title I. Seven percent of $15.8 billion is $1.1 billion. To be sure, it’s highly unlikely that every state in the country would ever spend 100 percent of its Title I set-aside on new charter schools (especially states that don’t even allow charters). But might states spend up to $400 million of their set-aside funds on new charters? If so, given that the Charter Schools Program is now funded at $400 million, that would double the federal investment in these life-saving schools.

Thank you, Madame Secretary. And now, pro-charter governors, legislators, and state superintendents: Show us the money! 

If you have forty seconds, please indulge me for a brief brain experiment. Watch the first half minute of the embedded video of college students playing in a circle. Your task? Count how many times the students wearing white shirts pass the basketball.

If you guessed fifteen passes, you win!

But did you see the gorilla?!?

For two decades, in multiple countries and contexts, as part of repeated research studies, thousands of audiences from diverse backgrounds have watched this video for the first time. A stunning 50 percent become so distracted trying to count the passes that they completely miss something extraordinary. As the researchers, Christopher Chabris and Daniel Simons, note, “Halfway through the video, a female student wearing a full-body gorilla suit, walked into the scene, stopped in the middle of the players, thumped her chest, and then walked off, spending about nine seconds on-screen.”

Don’t believe me? Watch again.

In the world of neuroscience, this phenomenon of being oblivious to the obvious is called “inattentional blindness.” This occurs any time we as human beings fail to notice a fully visible but unexpected object because our attention was on another task, event, or activity.

Inattentional blindness is an important concept to keep in mind as we all await the April 10 release of the 2017 National Assessment of Educational Progress (NAEP) for reading and mathematics for fourth and eighth grades. The U.S. Department of Education releases results bi-annually to help inform decisions about how to improve the education system in our country.

If prior results are any guide, reading achievement results will likely remain stagnant. Nevertheless, there will be reams of analysis of certain subgroups, especially highlighting the stubborn achievement gaps within the mesmerizing categories of race and income percentile. For example, as a NAEP teaser, Fordham Institute produced thirty-five charts of analysis, the vast majority of which focused on racial differences in outcomes for African American, Hispanic, and white students.

Yet you’re unlikely to see NAEP analyses of achievement differences by family structure, even though we know the stability of the family within which children are raised matters monumentally to their educational outcomes.

Take a recent study from the Institute for Women’s Policy Research, which finds that the number of single mothers in college has doubled over the last decade to nearly 2.1 million. Despite their best efforts to create a better future for themselves and their children, only 28 percent of single mothers who entered college between 2003 and 2009 earned a degree or certificate within six years—an outcome that adversely affects both mother and child. For college-going women without children, it was 57 percent.

Indeed, evidence on the adverse effects of fragile families on children is widely accepted. Even studies cited by those who believe promotion of marriage is an ineffective strategy to wage in the war against poverty correctly assert that “decades of research show that children raised in single-parent homes fare worse on a wide range of outcomes (e.g., poverty, educational attainment, nonmarital and teen childbearing) than children raised by two biological parents. The poverty rates of single parent households are particularly striking.”

Consider too recent words from the University of North Carolina at Chapel Hill. In announcing a grant to study the link between poverty and the high-risk impact on processes that facilitate learning, the school declared that “Early-life experiences that occur between the prenatal period through age 2 have major impacts on the trajectories of children's early cognitive development, including the neural substrates that support these abilities. Traditional investments toward children in the one-to-two years prior to their enrollment in kindergarten may be too late.”

And according to the CDC, there were 1,016,040 births to women aged twenty-four and under in 2016. Seventy-one percent of these births were to unmarried women, and 42 percent were to women who were having between their second and eighth child. There are always exceptions, but the magnitude in the number of multiple, non-marital births to typically unprepared young women and men, and the decades-long rise in the rate of single-parenthood, create a much greater risk of unstable home environments and fragile families that frequently correlate with child poverty, chronic student absenteeism, and the kind of toxic stress that begins in utero and impairs long term academic advancement of children.

***

Alongside the spirited debate about the causality between single-parent households and child poverty, there is new evidence of the impact of absent fathers, particularly on boys. After studying more than 1 million children born in Florida between 1992 and 2002, MIT researcher David Autor found that “Relative to their sisters, boys born to low-education and unmarried mothers, raised in low-income neighborhoods, and enrolled at poor-quality public schools have a higher incidence of truancy and behavioral problems throughout elementary and middle school, exhibit higher rates of behavioral and cognitive disability, perform worse on standardized tests, are less likely to graduate high school, and are more likely to commit serious crimes as juveniles.”

And Raj Chetty and Nathaniel Hendren just released a comprehensive study titled “Race and Economic Opportunity in the United States” to understand why “racial disparities in income and other outcomes are among the most visible and persistent features of American society.” Some have framed the findings as evidence of the punishing reach of racism for black boys, but the research itself suggests a more nuanced, inclusive recognition of family structure as a key contributor to racial disparities. The study states that “higher rates of father presence among low-income black households are associated with better outcomes for black boys...Black father presence at the neighborhood level predicts black boys' outcomes irrespective of whether their own father is present or not, suggesting that what matters is...community-level factors associated with the presence of fathers, such as role-model effects or changes in social norms.”

In other words, fathers matter.

***

So what does all this mean for NAEP and understanding the factors that truly drive education outcomes? To be clear, NAEP data are a vital and invaluable resource. The test remains the largest continuing and nationally representative assessment of what our nation’s students know and can do. It has served as a national yardstick of student achievement since 1969.

NAEP includes information on special student demographic groups, such as race, ethnicity, socioeconomic status, gender, disability, and limited English proficiency. But reporting by these main measurement groupings—though critical—has the unintended effect of fueling the fixation on those categories as the primary bases for comparison. This is particularly true in the case of race, where the differences in outcomes (e.g., the Hispanic-white achievement gap) are perceived to be caused by race-related reasons (e.g., racism), which often elicit only race-related interventions, as opposed to other factors complementing those forces. What gets measured gets managed. And what does not get measured gets ignored.

Enter family structure in a full-body gorilla suit, undetected due to obsessive counting in each of the other NAEP demographic groupings.

Information on the stability of family structure is hard to track, especially if self-reported. Nevertheless, I challenge the technical experts at the National Center for Education Statistics (NCES) to incorporate within NAEP groupings some kind of proxy measure for family structure or stability that already exists and for which the government has already collected data. What about age of parents? Or marital status of parents at time of birth? Perhaps there is a way to make use of (anonymized) tax records, as has been done with the Equality of Opportunity Project.

Unlike gender or race, family structure is not a static measure that is easily discernible. It changes over time. Take for example the powerful analysis done by Maria Cancian and Daniel R. Meyer in their study titled “Implications of Complex Families on Poverty and Child Support Policy.” They monitored 7,169 first-born children of unmarried mothers in Wisconsin between 1997 and 2007. See figure 1, from the study.

Figure 1. Most children born to unmarried parents will be part of complex families

Source: Maria Cancian and Daniel R. Meyer, “The Implications of Complex Families for Poverty and Child Support Policy,” University of Wisconsin-Madison Institute for Research on Poverty Webinar (September 19, 2012).

Over the decade, ten-year-old children of unmarried parents lived in increasingly complex family structure, as both the biological father and mother had more children in and outside of their relationship.

One suggestion from noted psychologist and researcher Nicholas Zill is that parents of a representative subsample of NAEP-tested students could be contacted by internet, phone, or mail to gather information on parent education, family income, welfare receipt, marital status, etc. These data could then enrich the NAEP tabulations to better estimate family structure and stability. This would add to the cost of the program, of course, but it would also strengthen the utility of the enterprise and perhaps shed new light on why educational progress has slowed.

As the data cited above indicate, family structure and stability are undeniably significant influencers on a host of child outcomes, including academic achievement. NAEP researchers would do our nation a great service if they developed a common metric that could estimate the effects of these forces on educational progress.

***

Virtually 100 percent of the research subjects who failed to witness the invisible gorilla upon first viewing immediately saw it the second time around. They usually express disbelief that they initially missed it. But once confronted with the truth, the viewers could no longer ignore the reality right before their eyes.

Leaders in education reform and the researchers at NCES have both the capacity and the responsibility to elevate family structure as a critical prism through which we evaluate our country’s educational progress, on par with race, socioeconomic status, and the other key groupings.

We do not yet know whether 2018 will bring more disappointing, unchanged NAEP scores, particularly in reading. But we can predict with great certainty that education reformers and policymakers will engage in handwringing about outcome disparities between races, classes, and genders. And that they’ll ignore a force at least as fundamental to human development. So let 2020 be an opportunity to create a new baseline to measure progress by a critical indicator that has been hidden in plain sight for decades: family structure.

If we really want to cure our blindness and understand why our children are not making the progress we seek, we must make this essential and predictive measure invisible no more.

This post is the fourth in a series of commentaries leading up to the release of new NAEP results on April 10. The first post discussed the value of the NAEP; the second looked at recent national trends; and the third examined state-by-state trends.

For the past fifteen years, a set of large urban districts have agreed to participate in what’s still called the “Trial Urban District Assessment,” or TUDA, as part of the National Assessment of Educational Progress. New TUDA scores will come out next month, just as they will for the nation and the states.

To help us prepare, Fordham’s research interns and I dug into NAEP data to see which district-level trends are worth watching. As I’ve argued before, we don’t want to over-interpret short-term changes, so it’s better to look at trends that span four years or more, i.e., trends that are based on at least three test iterations. Here’s a look, then, at statistically significant changes from 2011–15 for every participating district.

Cells that are empty indicate that there were no statistically significant changes, and the numbers represent statistically significant scale score changes from 2011 to 2015. Since districts, rather than cities, volunteer to participate in TUDA, most data only include students enrolled in traditional public schools. However, results for Atlanta, Baltimore, and Chicago include charter students in both years, and Los Angeles and Miami-Dade include charter students in 2015.

Table 1: Statistically significant changes on NAEP, 2011–15

As was the case with the state-by-state trends, the urban trends in reading generally look more positive than those in math. Four cities saw statistically significant improvements in reading in at least one grade level (D.C. and Miami saw them in both; Chicago and Los Angeles in one grade each), and no city went backwards. In math the picture is more mixed, with four cities showing progress in one grade (Chicago, Cleveland, D.C., and Miami), but six cities seeing declines (Albuquerque and Baltimore saw them in both grade levels; Hillsborough County, New York, Philadelphia, and San Diego in one grade each); in one city, Dallas, it was mixed.

It’s clear why Miami’s Alberto Carvalho is a hot property, with his gains in three of the four categories. Chicago’s fourth graders saw a big jump. And as we’ve discussed before, the District of Columbia continues to make remarkable progress.

As for the laggards, Baltimore’s scores are probably down because Maryland is no longer excluding so many students with disabilities from NAEP. But I can’t begin to explain Albuquerque. Maybe Hanna Skandera can.

As with national and state data, we should be mindful that demographic changes can affect achievement trends. So if we want to understand which policies and practices might be helping or hurting, we need to find a way to deal with these demographic trends. Our approach is to analyze results for each of three major racial groups, in addition to the overall student population. So let’s take a look at that, first for reading and then for math.

Table 2: Statistically significant changes in reading scale scores, 2011–15

This approach allows a few more districts to enter the winner’s circle, namely Dallas, New York City, and San Diego. The first two saw improvements for their Hispanic students; the latter made gains for its white and black students. We can also see which racial groups were driving improvements overall. In Chicago, it was white and black students (but not Hispanics). In D.C. and Miami it was Hispanic and black students (but not whites). And in L.A. it was just Hispanic students.

Houston saw the only decline—in fourth grade reading for its Hispanic students.

Now let’s look at math.

Table 3: Statistically significant changes in math scale scores, 2011–15

In this case, disaggregating by race allows us to identify some additional laggards: Austin, Houston, and Jefferson County, all of which saw statistically significant declines for their Hispanic students. On the flip side, Atlanta and New York City each posted gains for at least one subgroup—white students in the case of the Atlanta and Hispanic students in the case of New York.

There are a few other interesting findings to note. First let’s examine Chicago. Its white students made significant gains in both fourth and eighth grades; perhaps gentrification is a possible explanation. On the other hand, the District of Columbia, often seen as the capital of gentrification, didn’t post any statistically significant progress for its white students. (Of course, D.C. has so few white students that they would have to make enormous gains in order for it to register statistically.) Then there’s Miami-Dade, which pulled off the hat trick of posting improvements across all three major racial groups. Impressive!

One more wonky comment: Chicago has received a ton of attention for significantly outpacing other urban districts, as reported by the New York Times. That was based on an analysis of its NAEP results by Sean Reardon, who subtracted its 2013 eighth grade scores from its 2009 fourth grade scores—thus tracking the same cohort of students over time. But note what happened in 2015: Its fourth grade scores went way up. That’s going to make it a lot harder for its “cohort gains” to look as impressive going forward.

***

Just as Indiana and Tennessee are the states with momentum going into the 2017 NAEP release, Chicago, D.C., and Miami are the districts to watch and learn from. Anyone up for a “research trip” to South Beach? We’d better go soon, in case April 10 brings disappointing news.

On this week's podcast, Carissa Moffat Miller, the new executive director of the Council of Chief State School Officers, joins Mike Petrilli and Alyssa Schwenk to discuss CCSSO’s campaign to highlight innovative state education policies. On the Research Minute, Amber Northern examines the effects of the National Heritage Academies chain of for-profit charter schools.

Amber’s Research Minute

Susan Dynarski et al., “Estimating the Effects of a Large For-Profit Charter School Operator,” The National Bureau of Economic Research (March 2018).

STEM education is, by design, integrative. It strives to emulate the real-world work of engineers within a teaching environment. Traditional science and math concepts merge with hands-on design-and-build work using technology, often through “design challenges.” Team dynamics, learning by failure and revision, and analytical thinking all factor in as well. It’s a big lift, but such efforts are vital for schools to attempt as demand for STEM—from parents, employers, the military, and colleges—increases. Traditional education models may not readily adapt to the hands-on demands of STEM, nor can many practitioners turn on a dime to accommodate a tech-heavy pedagogy. A new report from Michigan Technological University sheds light on some of these complexities that teachers face bringing STEM education into their practice.

Authors Emily Dare, Joshua Ellis, and Gillian Roehrig use observation and interview data to assess the first-time STEM integration efforts of teachers in nine physical science classrooms in different, unnamed middle schools in the United States. The researchers posit that a lack of consensus over best practices and a lack of professional development contribute to the difficulties. Both classroom observation and teacher reflection data for these nine case studies of teachers attempting STEM integration with little or no previous training data showed fundamental difficulties in full integration efforts.

The researchers’ primary variable was the amount of time teachers spent in each daily lesson on science, math, and engineering, both singularly and in integration with others. More integration was considered desirable, but the data showed far fewer instructional minutes with two or more disciplines fully integrated than researchers expected. Follow-up interviews with teachers focused on three specific factors—the style of integration chosen, how science and math concepts were integrated with engineering within the style chosen, and levels of student engagement and motivation.

STEM-integration can take on several styles in classroom practice: Engineering work can be used as an “add-on” to reinforce concepts—say, by having students design and build a roller coaster model to cap off a traditional unit on velocity—or it can be used as a primary means of imparting science and math concepts, or the balance can fall somewhere in between. The nine teachers in the case studies opted for variations of each style, determined largely by their reported comfort and familiarity with hands-on engineering projects. However, the observed time spent on integration was sometimes at odds with the teachers’ interview responses.

The dominant pattern observed, even among teachers reporting an effort at full integration, was discrete blocks of science instruction followed by engineering work, although these blocks varied from minutes to hours to days across classrooms. The smaller the discrete blocks, the more integrated the classroom, but that’s a far cry from proper STEM integration as defined by the researchers. It makes some sense that these science teachers would be focused on imparting the core knowledge of their traditional discipline—it is their job, of course, and many states have mandated science exams in middle school as well. However, interviews reinforced the notion that these teachers experienced discomfort with their own skill levels around both math and engineering.

Student engagement appeared to be a key factor for teachers in deciding their level and patterns of integration, both in designing and executing lessons. Teachers reported trepidation that students would not grasp science concepts through the engineering work, so they often defaulted to teaching science first, engineering last—which is again unsurprising, given the teachers’ job descriptions. But even those who were observed engaging in the most integration often reported that they believed their students didn’t grasp the science concepts well enough. Those teachers would typically do science first, engineering second, and then backtrack—returning to traditional science teaching so quickly that it shortchanged the hands-on engineering work or ended it entirely. This could also contribute to the even smaller amounts of instruction time dedicated to mathematics, either alone or in integration. Many teachers also reported feeling a mental clock ticking away, urging them to “finish” and move on. All of this is problematic because most teachers reported their students were eager to do the hands-on work and enjoyed it while it was underway.

Real-world work is a defining feature of STEM education—from labs to field research to problem-solving challenges in design—but these case studies suggest that single-subject teachers may struggle to integrate the various disciplines successfully. Concerns related to time, student engagement, and teacher comfort and confidence all impinge on possible success. Perhaps it is even impossible to provide high-quality STEM education in a single classroom. But there is some hope. One of the more successful cases in this report notes the assistance given by a second teacher—a math teacher—to a science instructor, so maybe fully-STEM-integrated schools or programs are the most effective providers. As demand increases, more work is needed to make sure that teachers and schools are confident in providing high-quality STEM education.

SOURCE: Emily A. Dare, Joshua A. Ellis, and Gillian H. Roehrig, “Understanding science teachers’ implementations of integrated STEM curricular units through a phenomenological multiple case study,” International Journal of STEM Education (February 2018).

Personalized learning (PL) is becoming cause célèbre in education circles, drawing support from important outside influencers like RAND and the Chan Zuckerberg Initiative. It has the potential to revolutionize classroom practice. And a new analysis from KnowledgeWorks indicates that PL has strong support across the country, at least according to state ESSA plans. Researchers looked at all the plans submitted to the U.S. Department of Education—some of which have been approved, and some of which are still under consideration—to identify nationwide trends related to PL.

With the definition of PL an unsettled matter, KnowledgeWorks sets forth their own criteria: instruction aligned to rigorous academic standards and social-emotional skills students need to be ready for what’s next after high school; customized instruction allowing each student to design personalized learning experiences aligned to his or her interests; varied pacing of instruction based on individual student needs, which allows students to accelerate or take additional time based on their level of mastery; real-time differentiation of instruction, supports, and interventions based on data from formative assessments and student feedback; and access to clear, transferable learning objectives and assessment results so students and families understand what is expected for mastery and advancement.

KnowledgeWorks finds that most states evince one or more of the above criteria in their ESSA plans, and fully one-third of them have personalization at the core of their vision statements for K–12 education. Support for competency-based practices; a focus on readiness, social-emotional, or employability skills; and provision for multiple pathways to achieve and demonstrate readiness are the most prevalent features across state plans.

Most of the support for PL shows up in the accountability portion of state plans. Thirty-seven plans, for example, include an extended-time graduation-rate indicator that gives schools credit for successfully boosting students who need extra time to earn their diploma. Thirty-five states incorporate multiple pathways for demonstrating student readiness based on a pupils’ trajectory into career or college or military service. Twenty-five feature proficiency indicators that include growth measures, which better represent students’ true academic status. And nineteen states establish informational dashboards for families to evaluate schools and to keep track of students’ progress toward readiness via multiple indicators.

Yet personalized learning also shows up in ESSA plan sections that concern improving low-performing schools, supporting teachers, using Title IV funds, and personalizing education plans for all students. Eleven plans, for example, prioritize PL strategies as the first line of support for struggling schools. These strategies include real-time data tracking of students’ growth toward mastery, community and family engagement, and a whole-child focus that comprises wraparound services and efforts to boost social-emotional learning opportunities. In terms of teacher supports, more than fourteen states offer personalized professional development opportunities, flexible micro-credentialing, and specific training on in-class PL strategies. Twenty-one states devote portions of their Title IV dollars to social-emotional learning and whole child supports. And nineteen states ensure that all students will have personalized education plans that align to their academic needs, interests, and goals.

The news is not, however, all good for those looking for a nationwide movement toward high quality personalized learning. As KnowledgeWorks notes, the flexibility afforded by ESSA did not seem to encourage a trend toward new and better assessments. “Several states are advancing new performance-based science assessments,” they wrote, “a few are exploring local assessments, and one state is committed to more granular score reporting.” But the status quo reigns for the most part. KnowledgeWorks promises continued vigilance on this front, looking for a move toward innovative assessments with a focus less on standardization and more on personalization along with more useful and timely reporting of results. Assessments that inform students and families and guide practice must be part of successful PL. KnowledgeWorks gives props to the community engagement process required for ESSA plan development for bringing personalized learning to the fore.

It is clear that parents, policymakers, and the public are looking for education that produces well-qualified graduates across the board—young people ready and able to take on whatever challenges they choose. The proof of the pudding is in the tasting, they say. The recipe is largely here. And it’s time to get cooking.

SOURCE: “Personalized Learning and the Every Student Succeeds Act: Mapping Emerging Trends for Personalized Learning in State ESSA Plans,” KnowledgeWorks (March 2018).