Give back to HGSE and support the next generation of passionate educators and innovative leaders.

Winter 2014

Dropping Out: Is Your First Grader at Risk?

Infographics by Killer InfographicsIllustrations by Killer Infographics

In January 2013, Joshua Starr, Ed.M.'98, Ed.D.'01, the superintendent of Montgomery County [Md.] Public Schools, was walking through New York City's Central Park when the idea hit him.

"I was actually walking with another Harvard grad, Brian Osborne, and I was talking about the Seven Keys," he says, referring to an approach that the county is known for, which uses data to highlight a path to college persistence. Osborne, Ed.M.'01, Ed.D.'11, is superintendent of schools in South Orange-Maplewood, N.J. "I thought, wait a second. We've got to do this for kids who've dropped out. We actually don't know what's happened with them. Montgomery County has always been organized around the experience of successful kids. I was thinking, what about the opposite? I want to understand what the experience is of our kids at the youngest grades and the kids who are most vulnerable."

So Starr directed his staff to prepare a report on the question, and they looked at the experience of the class of 2011, dropouts and non-dropouts, from all of the county's 26 high schools, to see what they could learn. They looked in particular at behavior, attendance, and aspects of academic performance.

Initially, the researchers came back with results as far back as middle school, but Starr says he sent them back to look further. "The first draft they gave me was down to sixth grade, and I said I wanted to go down as early as possible. Frankly I'd like to get down to preK."

The final report, titled "Just the Right Mix: Identifying Potential Dropouts in Montgomery County Public Schools Using an Early Warning Indicators Approach," goes back to first grade and identifies "cut points" that are related to an increased likelihood of students dropping out of school. The findings were compared with the class of 2012.

The earliest findings: First-graders are considered to be at risk of disengaging from school (and potentially becoming fully disengaged and dropping out in the future) if by the third marking period they're absent from school nine or more times, are below grade level in reading and/or mathematics, and/ or have a calculated grade point average below 1.2 in the third marking period.

The notion that children who may not even be able to tie their shoes, nor are entirely clear what it means to share, can already be showing signs of disengagement may seem hard to comprehend, yet for those familiar with the literature, intervening early makes sense. The earlier, the better, in fact.

"There's a sense of surprise for first grade," says Nicholas Morgan, executive director of the Ed School-based Strategic Data Project, housed at the Center for Education Policy Research at Harvard University. "One of the biggest things that's talked about nationally is third-grade reading. If a child is off track in third-grade reading, then the chances of recouping that are much more challenging. I think you still see a lot of school districts nationally thinking about high school graduation as a high school problem. The problem is when you think about it that way, you simply don't have enough days to catch up."

Starr and other Montgomery County officials are adamant that being designated as high risk for dropping out is not intended to be used to label or marginalize kids, but rather to make their issues less opaque and more personal and to serve as a starting point for possible intervention.

"I have two mantras about data, and I'm a former accountability guy," Starr says. "One is, at best, data helps you ask better questions. The other is use the data to name names. So if a first-grader's on track to be absent nine times, who is that kid? What is his story? What else do we need to know about him? Use the early warning indicators to identify kids and families. But you're not going to draw conclusions unless you get in there and work with the family and the kid and figure out solutions. So it's a starting point."

In June, 10 Montgomery County Schools were chosen to do just that — to become a kind of incubator network that would begin to refine the data and test possible ways to help at-risk kids. Known as the Interventions Network and featuring their own acronym, EPIC (for Early warning indicators, Personalized learning plan, Implementation, and Culture of high expectations), the schools applied for and were chosen, according to Kimberly Statham, deputy superintendent for teaching, learning, and programs, based on how interested and engaged the faculty, PTA, and parents were in working to keep students from disengaging and dropping out. The end result of the incubating work, as well as continued work on the data, will be to test and build an early warning indicator system founded on the Montgomery-specific experience.

For Starr, responding to the earliest early warning indicators may require more players than just schools. "If a first-grader has nine absences, that is not a six-year-old's problem; that's a family issue," he says. "We've got to develop systems that coordinate collaborative actions amongst the various agencies that touch kids' and families' lives. If a family is not able to get their kids to school, that family may be working with other agencies: Health and Human Services, the police department, juvenile justice, whatever it may be. And we've got to organize ourselves to work with those groups."

Statham says this coordination will be part of the Interventions Network's responsibilities.

"We have our pupil personnel workers (PPWs), whom we would ask to keep an eye on this family," she says. "If that PPW finds out that the family is struggling, mom or dad lost a job and they're short on resources, the Office of Community Engagement and Partnerships can provide wraparound services. Connect them to our agency partners in the system. This is not just academic support for students in these 10 schools. We definitely are taking a more holistic view. And so with each child having a personalized learning plan, it will customize the support that we provide."

Statham says the short-term goal is to develop personalized learning plans for students in the Intervention Network schools who are showing signs of needing extra support, but that, long term, all children will have one.

Starr sees one key advantage of personalizing the data, of "naming names," is potentially moving administrators beyond demographic assumptions in assessing kids' issues.

"It's not just about, here are black kids, here are Latino kids, what are we doing overall? It's, wait a second, Josh has been absent six times this month, what's going on? Let's do something. Not, Josh is a black kid, so we've got to do something different. It creates a degree of accountability and opportunity to say there are real-time indicators and real actions we can take every day to reverse the potential that some of our kids aren't going to graduate."

The idea of taking assertive action in the face of early warning indicators appeals to Janette Gilman, president of the Montgomery County Council of PTAs.

"If I take a test and find out I'm prone to early Alzheimer's, what do I do? There's no intervention," Gilman says. "But in this case, if the data is good and it shows a strong correlation — it's not predictive, but it's an indicator — if I step in and try to do the best for that child, then at the end of the 12 years of school, at least the system and the families can feel like we did the best we could to let every child achieve."

Montgomery County is one of a number of places across the country where early warning indicator systems are either under development or already in place. Other efforts are happening in places like Denver and Massachusetts, as well as several districts in Ohio.

In Denver, the system, which is not yet approved, grew out of efforts to develop a success-tracking program called the Gateway, modeled on Montgomery's Seven Keys system. Researchers say the early warning indicator system emerged as a way to measure whether students were on track to reach stages along the way to the project's goal: measuring true college readiness, meaning that students would not need a remedial course when they arrived at college.

"The final product is going to be [that] at any point in time from K–12, we know what is the arc of success that a kid is having and whether the kid is on track," says Chung Pham, senior strategic projects analyst in assessment, research, and evaluation in the Denver Public Schools through his fellowship at the Strategic Data Project. "Having a high arc or not, is the kid on or off track? And then what are we going to do to help that kid get back on track?"

Pham and fellow Strategic Data Fellow Tracy Diel Keenan, also a senior strategic projects analyst in assessment, research, and evaluation in the Denver Public Schools, say the Denver program could potentially offer educators real-time information — a number that includes attendance and behavior mixed in with academic information.

"We have dashboards — principal and teacher portals — so various people could have access to the data," Diel Keenan says. "Attendance is obviously the main thing that would change daily, as well as, potentially, behavior."

In Massachusetts, a statewide system divides students into three grade chunks (from first grade all the way up to 12th) and focuses on children's likelihood of achieving progressive academic targets — all of which are separately considered indicators for high school graduation but also have more immediate relevance. The system replaces a previous arrangement that only focused on incoming ninth-graders and their likelihood of graduating from high school.

"Folks that were using it told us they felt giving just the one-time snapshot of the incoming grade-nine students wasn't enough," says Jenny Curtin, coordinator for high school graduation initiatives in the Massachusetts Department of Elementary and Secondary Education. "They wanted that information for earlier grade levels. They were really thinking about middle school, and they also wanted early warning indicator information for students throughout their high school experience. They wanted to know if they were making progress with students and to track them over time."

Under the new system, grades one through three are measured against a goal of reading by the end of third grade; grades four through six on proficient or advanced performance on the English and math portions of a state test indicating middle school readiness; seven, eight, and nine on high school readiness with passing all ninth-grade; grades 10, 11, and 12 focus on the goal of high school graduation.

The new system, which is in its second year, estimates the likelihood of achieving goals based on historical data going back to 2005 but assesses current individual students' risk levels based on attendance and other factors from the prior year. The new system, which includes a partnership with American Institutes for Research (AIR), took a year to develop.

The point of the clustering arrangement is to empower educators.

"If I were a first-grade teacher or principal of an elementary school, finding out that my students are at risk for not graduating from high school — it's important, but it's not very tangible or actionable or relevant to where I'm at today," Curtin says. "Saying that they're at risk for not reading by the end of third grade is really meaningful to the educators working in the building."

Camia Hoard, Ed.M.'99, assesses early childhood and elementary teachers in the District of Columbia Public Schools as a master teacher, having spent 16 years in the classroom teaching kindergarteners through fifth-graders. She feels that the focus on what goes on in the building should be more acute.

"I think that we have a lot of systems that train teachers to control behavior as opposed to changing it," she says. "What I would like to see happen from the perspective of my teaching experience is to take a look at what things we are doing as teachers that are pushing kids out."

For example, Hoard cites what she says is a commonly used repressive color system in the classroom with which younger children are labeled for behavior — red means a call home, green is a good job.

"What's alarming about something like that is, one, how widely used it is," she says. "And two, how we recognize who our red kids are immediately. 'Little Johnny's out of the line. Change your color to yellow.' It's not even nine o'clock yet. Instead of saying, 'This system doesn't really work for Johnny.'"

In order to move beyond systems like these, uncomfortable conversations are necessary, Hoard says.

"I think back on my teaching career and I wonder, what made me change that? And it was probably very simply a conversation with somebody who said, 'Where's the discussion after you change your color? You've done something wrong, now you're on yellow; you've done something wrong again, now you're on red; now we tell your mom.' As opposed to, 'Let's have a conference about why walking in a line is important or safety issues around body space,' or 'What can Ms. Hoard do to help you?'"

Hoard says similar care has to be taken in approaching parents. "What I worry about is we've got teachers who are sitting back and saying, 'Well, he doesn't come to school on time, so what am I supposed to do?' Without having reached out as that number one connection. Because when other people step in, that could possibly cause resistance in families."

Approached correctly, a parent can be made to feel he or she is helping an exceptional child reach her potential, rather than being called to account.

"Your kid's really smart," Hoard says she once told a parent whose child was regularly late to school. "Really, really intelligent and showing me higher-level thinking skills. But he misses so much school that he's not performing, and I'm seeing [questionable] behavior and all these things because we don't get started on the right track in the morning."

In Massachusetts, how districts conduct conversations about the early warning indicators is considered a key part of the process.

"We have actually deterred districts from sharing [early warning] risk levels with parents and students just as a blatant 'here's your risk level' conversation," says Curtin. "Instead, we've encouraged districts to be a little bit more thoughtful in communicating information. If they're seeing that they have a high-risk student and that student has really poor attendance and course grades, and those indicators went into determining that they were high risk, then they should be having a conversation with the parents and the student about why those are important things, so that they stay on track."

Curtin adds that districts need to be thoughtful about digging into the root causes of warning signs like poor attendance.

"The early warning indicator system is not a diagnostic tool; it's a tool to flag students in a very systemic way," she says, "based on research and based on our own longitudinal analyses. But it's not going to tell educators the root causes of what's happening with students. We'll never be able to have that information at the state level. That's always going to have to be done at the local level."

Indeed, when the data is used for that kind of an examination, it can yield significant results. According to Mindee O'Cummings, principal researcher and the lead for AIR's dropout prevention work (who did not personally participate in the Massachusetts research), two different districts in Ohio that looked deeper at the root causes found surprising information that helped them make significant improvements in their approach to their communities.

In one, indicators showed that a large number of students were failing one or more courses in ninth grade. Educators suspected the challenging courses that come at that stage — ninth-grade algebra or English, for example — were to blame. But when they examined the data, they found the culprit was a ninth-grade biology class that required students to do more writing about biology than they were ready for. As a result, educators worked with teachers to ease the amount of writing and also moved the biology course to a later year, replacing it with physical science.

In another Ohio district, prompted by the indicators, educators decided to take a second look at an historical truancy problem. What they found was much more complex than the hooky-playing they'd expected. Most of the students were "newer truants." When officials called the families to ask what was going on, they found that many parents were working double shifts. As a result, their older children had taken on the responsibility of getting younger siblings to school and were unable to get to school themselves. Consequently, educators in that district are currently considering creating other ways for these older children to go to school — possibly a night school for credit recovery — now that the issue is clear.

O'Cummings says the value of early warning indicator systems' close scrutiny of the data — like naming names of mysteriously missing first-graders, as Superintendent Starr is hoping for in Montgomery County — is moving educators beyond assumptions: behavioral or demographic.

"It's equal opportunist," O'Cummings says. "And we need to be equal opportunist with our indicators, and it's hugely important. I mean, I'm a living example of it. I'm white, I came from a middle-class family, I went to a public school. I have an orthopedic impairment — I use a wheelchair and if you actually look at my graduation likelihood based off my demographics, it's up there in the 98th or 99th percentile. But at 16, I dropped out of my high school. If anyone was using an early warning system way back when in the dark ages, it would have identified me."

— Jonathan Sapers, a graduate of Harvard College, is a freelance writer in New York who writes about education for Scholastic Administrator and Teachers College Alumni Magazine.