1. Keeping Students on a Graduation Path in Philadelphia’s Middle-Grades Schools
Early Identification & Effective Interventions
Balfanz, Herzog, & Mac Iver (2007)
It is an honor and a pleasure to share what we have been learning about keeping students on a path that leads to high school graduation. We now know a lot about the early identification of students who are not likely to graduate from high school unless they experience effective, sustained interventions during the middle grades. We have also begun to understand better how to design prevention & intervention programs to help these students back onto a path that leads to graduation.
Balfanz, Herzog, & Mac Iver 2007

2. Balfanz, Herzog, & Mac Iver 2007
Research Questions
How many students exhibit early warning signs that they are beginning to disengage from schooling at the start of the middle grades in high poverty schools?
Can schools easily identify and effectively rescue students who have high odds of ultimately dropping out?
Everyone understands that most high school dropouts begin disengaging from school long before they actually dropout. But how early and easily can schools identify these students? Is it possible to help these students back onto a path that leads to graduation by providing the right kinds of supports?
Balfanz, Herzog, & Mac Iver 2007

3. Balfanz, Herzog, & Mac Iver 2007
Data and Methods
We followed all 13,000 of Philadelphia’s public school students who were enrolled in the 6th grade in October We followed them through October 2004 (1.25 years beyond their expected graduation date)
We also looked at 3 more recent cohorts of Philadelphia’s 6th-graders and at cohorts in two other cities to verify our findings
Today, I’m presenting findings from longitudinal analyses following all 13,000 of the students in Philadelphia who were sixth-graders in October 1996 until one and a quarter years past their expected graduation date.
The findings are robust, we have been able to replicate them in additional more recent Philly cohorts and in two other large cities.
Balfanz, Herzog, & Mac Iver 2007

4. Balfanz, Herzog, & Mac Iver 2007
Cont. – Data and Methods
We did a preliminary screen of about 20 variables (e.g., test scores, over age for grade, course marks, course failures, attendance, behavior marks, status and demographic indicators) to see which, if any, could identify as early as 6th grade students at high risk for slipping off a graduation pathway
We looked for variables with a high yield (i.e., about 75% or more of students with this characteristic do not make it to 12th grade on time)
Instead of designing a special diagnostic tool for measuring student disengagement, we wanted to see if any of the data that school districts already routinely collect would identify students who were beginning to disengage from school. We attempted to find a small set of simple predictive variables that would be useful to schools in identifying sixth graders who had low odds of ultimately graduating from the school district. We judged a variable to be useful if about 75% or more of the 6th graders flagged by that predictor did not make it to 12th grade on time.
Balfanz, Herzog, & Mac Iver 2007

5. Balfanz, Herzog, & Mac Iver 2007
Cont. – Data and Methods
Once we identified these flags, we then examined how well they predicted a student’s success in graduating from the Philadelphia school district on time or within one and a quarter extra years.
The graduation rates we report are lower than official calculations because they don’t adjust for transfers out of the district. 15% of those who left the district were transfers out.
We used this time window because Ruth Neild’s research shows that the vast majority of graduates from the school dist of Philadelphia get their diploma on time or within one extra year. Thus, If we had followed these students longer, it would have increased our graduation rate estimates by just a few percentage points.
We code every student as a graduate from the Philly sch dist or as a non-graduate. Both students who officially drop out and those who transfer out are counted as non-graduates. Some of these students may obtain a high school diploma from some other district or from a private school. However, Rumberger and other researchers have shown that adolescents who transfer after experiencing school difficulties ultimately drop out in high numbers. So the official practice of excluding transfers when computing graduation rates may to large overestimates of the actual graduation rates.
Balfanz, Herzog, & Mac Iver 2007

6. Findings – 4 Powerful 6th Grade Predictors of “Slipping Off Path”
Attending school 80% or less of the time
Receiving a poor final behavior mark or a suspension
Failing Math
Failing English
The bad news is that 48% of the sixth-graders in the district displayed at least one these risk factors. The good news is that, of those who displayed a risk factor, 61% had only one. Most of the others had just two of the factors. Very few students have three of the risk factors and virtually none of the students have all four as 6th graders. Let’s consider each of the risk factors in turn.
Balfanz, Herzog, & Mac Iver 2007

7. 6th-Grade Course Failure as a Predictor of Not Graduating
Course failure was a much better predictor of not graduating than were low test scores.
Students who failed either a math course or an English/Reading course in sixth grade rarely graduated from the district.
Balfanz, Herzog, & Mac Iver 2007

8. % of students who actually graduated 13 6 Cum. % who left the District
Did 6th-graders who FAILED MATH in 1997 (n=1801) Graduate On Time or 1 Yr. Late?
% who were in ….
In 2003 (on time)
In 2004 (1 yr. late)
9th grade 4 1
10th grade 9
11th grade 8 2
12th grade 19
% of students who actually graduated 13 6
Cum. % who left the District 61 75
14% of the sixth graders in our sample failed math. These students had less than a 1 in 5 chance of making it to 12th grade on time.
Only 13% graduated on time.
Another 6% graduated with one extra year.

9. % of students who actually graduated 12 6 Cum. % who left the District
Did 6th-graders who FAILED ENGLISH in 1997 (n=1409) Graduate On Time or 1 Yr. Late?
% who were in ….
In 2003 (on time)
In 2004 (1 yr. late)
9th grade 5 1
10th grade 9 2
11th grade 8
12th grade 16
% of students who actually graduated 12 6
Cum. % who left the District 62 74
11% of the sixth-graders in our sample failed English. Only 16% of these students made it to 12th grade on time.
Only 18% graduated on time or with one extra year.

10. 6th-Grade Attendance as a Predictor of Not Graduating
Attending school less than 90% of the time increases the odds that a student will not graduate.
When a sixth-grader’s attendance dips below 80% (missing 36 days or more in the year), the student has only a 1 in 6 chance of graduating from the district on time or one-year late.
Balfanz, Herzog, & Mac Iver 2007

11. % of students who actually graduated 13 Cum. % who left the District
Did LOW ATTENDING 6TH Graders in 1997 (n=1934) Graduate On Time or 1 Yr. Late?
% who were in ….
In 2003 (on time)
In 2004 (1 yr. late)
9th grade 3 1
10th grade 6
11th grade 4
12th grade 17 5
% of students who actually graduated 13
Cum. % who left the District 69 79
15% of the sixth-graders in our sample attended school less than 80% of the time. Only 17% of these low attenders made it to 12th grade on time.
Likewise, only 17% graduated on time or with one extra year.

12. Poor Behavior in 6th-Grade as a Predictor of Not Graduating
Students who were suspended slipped off the graduation path in large numbers.
845 (6%) of the sixth-graders received one or more out of school suspensions. Only 20% of these students graduated on time or one year late.
222 sixth-graders received in-school suspensions. Only 17% graduated on time or one year late.
Students who were suspended slipped off the graduation path in large numbers. I’ll pause so that you can read the details.
Balfanz, Herzog, & Mac Iver 2007

13. Balfanz, Herzog, & Mac Iver 2007
Poor Behavior (cont.)
Receiving a final unsatisfactory behavior grade in any subject in the sixth-grade significantly reduced the chances that sixth-graders would graduate from the school district.
A very large number (4,893) and percent (38%) received at least one final unsatisfactory behavior grade.
Balfanz, Herzog, & Mac Iver 2007

14. % of students who actually graduated 24 5 Cum. % who left the District
Did 6th-Graders With an Unsatisfactory Behavior Grade in 1997 (n=4893) Graduate On Time or 1 Yr. Late?
% who were in ….
In 2003 (on time)
In 2004 (1 yr. late)
9th grade 3 1
10th grade 7
11th grade 2
12th grade 31 8
% of students who actually graduated 24 5
Cum. % who left the District 52 64
Only 29% of the sixth-graders who earned an unsatisfactory behavior grade in at least one subject, graduated on time or with one extra year.

15. Percent of Sixth-Graders Graduating on Time or 1 Year Late
Failure & Behavior Combinations
On-Time Grads
1-Yr-Late Grads
Fail English but Good Behavior (n=176)
14% 7%
Fail English & Poor Beh. (n=725) 6% 5%
Fail Math but Good Behavior (n=298)
16% 8%
Fail Math & Poor Beh. (n=1006)
In addition to being a significant early warning signal in an of itself, poor behavior magnified the impacts of the other predictors. For example, students who fail math or English and also have a poor behavior grade in any of their classes slip off the path to graduation at even greater rates than students who fail math or English but have good behavior marks.

16. Graduation rates for 6th Graders with Different Numbers of Risk Factors
Percent Who Graduate
None
6265
56%
Only 1
3498
36% 2
1329
21% 3
619
13% 4
326 7%
1 or more
5772
29%
In fact, the more of the big four risk factors that a sixth grader has, the lower their odds of staying on a path that leads to graduation. This table shows how the odds of graduating decrease as the number of risk factors increase. Using the big four predictive indicators – failing math, failing English, attending less than 80% of the time, and receiving a poor final behavior mark – we were able to identify 60% of the 6th graders who would not graduate from the district within one and one-quarter years of their expected graduation date. Collectively, students with at least one of these predictive indicators have only a 29% graduation rate from the school district.

17. Balfanz, Herzog, & Mac Iver 2007
Discussion
We were able to find four variables with a very high predictive yield that identify the majority of sixth-graders who fall off the graduation path
These variables are each commonly measured and collectively capture a significant portion of a district’s future dropouts
Balfanz, Herzog, & Mac Iver 2007

18. Balfanz, Herzog, & Mac Iver 2007
Implications
Students fall off the graduation path in different but identifiable ways.
In 6th grade, most future dropouts have just one of the big four risk factors – especially poor behavior or poor attendance
Some have two risk factors, especially poor behavior plus course failure (in English or mathematics)
Less than 8% of the sixth-graders had more than two of the big four indicators.
I regard these findings as hopeful because most of the students who begin slipping off the graduation path at this point in their schooling, are not demonstrating the multiplicity of difficulties that characterize struggling high school students. By providing targeted supports in 6th grade and beyond, schools may be able to stop this slide and prevent these sixth graders from developing additional risk factors as they progress through school.
Balfanz, Herzog, & Mac Iver 2007

19. Balfanz, Herzog, & Mac Iver 2007
Why do you think so few sixth-graders recover once they display one of the big four warning signs?
Balfanz, Herzog, & Mac Iver 2007

20. Balfanz, Herzog, & Mac Iver 2007
Implications
Academic and behavioral problems at the start of the middle grades do not self-correct (at least in Philadelphia and the two other cities where we have replicated this work)
The most common and very harmful response to students who struggle in 6th grade is to wait and “hope they grow out of it”. But, they do not typically recover, they drop out.
Early intervention is absolutely essential.
Balfanz, Herzog, & Mac Iver 2007

21. Effective Interventions
Research comparing outcomes in Talent Development Middle Schools with matched other schools in Philadelphia suggest that comprehensive school reform can significantly reduce the number of students who develop a poor attendance habit, who fail math, or who fail English and can produce significantly higher graduation rates.
Of course, the first line of defense against student disengagement is to reform the instructional programs and teacher support systems in a school to ensure that in every classroom every day students are receiving engaging, substantive, and coherent lessons and are experiencing close productive relationships with both the adults and other students at the school. The Talent Development Middle Grades Model is one of the Forum’s whole school reform models that assists schools to make the comprehensive changes they need in curriculum and instruction and in teacher and student supports. Three of the middle schools that served 6th-graders in the cohort that we have been following were Talent Development Middle Schools. (These were the first 3 TD middle schools in Philly.) At the start of TD in Philly, the district’s research office had selected 3 control schools that were very similar to the 3 TD schools, so that both we and the district would have comparison data against which we could judge the progress of students in the TD schools. The TD model combines research-based instructional programs with extensive teacher training and support to enable more active and engaging pedagogy. It also provides targeted extra academic help to struggling students. Evaluations of the instructional programs and extra-help labs have shown that they significantly improve achievement when implemented with reasonable fidelity.
After we had finished identifying the big four risk factors, it occurred to us that it would be interesting to go back and see if Philly’s TD schools were successful in reducing the number of students who developed a poor attendance habit, poor behavior, or who failed math or English. We found that 6th-graders in the 3 TD schools were significantly less likely to become poor attenders or to fail math or English than were the 6th-graders in the control schools. (They were equally likely, however, to earn an unsatisfactory behavior grade.)
Not surprisingly, students who had attended one of the TD schools during the middle grades also graduated from the district at a greater rate than did the control students. That is, students who attended a TDMG school for three years were 55% more likely to graduate on time than control students.. However, our research indicate that such reforms are not enough to rescue every student.
Balfanz, Herzog, & Mac Iver 2007

22. Comprehensive Reforms Must be Combined with Targeted Interventions
Additional interventions specifically focused on improving behavior and attendance must be added
Balfanz, Herzog, & Mac Iver 2007

23. Balfanz, Herzog, & Mac Iver 2007
What characterizes effective interventions for behavior and attendance?
Balfanz, Herzog, & Mac Iver 2007

24. Common Features of Effective Interventions for Behavior and Attendance
Positive behavior and good attendance is constantly recognized, modeled, and promoted
The first absence or incident of misbehavior brings a consistent, appropriate response
The intervention uses simple data collection and analysis tools that enable teachers and administrators to better understand the school’s absenteeism or misbehavior problems
Attendance and behavior teams regularly meet to analyze data and devise solutions
Here are some conclusions that we have drawn as we search for effective behavior and attendance programs. First, there must be schoolwide efforts to promote positive behavior and good attendance. Second, individually-targeted efforts must be taken to understand why certain students are continuing to misbehave or not attend despite the schoolwide efforts to promote positive behavior and good attendance. Reaching an unresponsive student typically requires assigning a specific adult, usually one of the student’s main teachers with the responsibility of shepherding the student (building a closer, more personal relationship with the student, checking in daily with the student and giving that student immediate feedback).
Balfanz, Herzog, & Mac Iver 2007

25. Shepherding of the Initially Unresponsive
If the student is a low attender, the shepherd might call the student each day the student is absent to communicate that the student is missed and to ask the reason for non-attendance.
If the student has behavior problems, shepherding might involve asking each of the student’s teachers to complete a simple behavioral record and then checking at the end of the day how the student did.
If these modest shepherding efforts do not succeed, then it is time to seek even more intensive, individualized, and clinical interventions often involving one-on-one services from helping professionals.
Balfanz, Herzog, & Mac Iver 2007

26. A 3-Stage Intervention Model
Schoolwide reforms aimed at alleviating 75% or so of the problem behaviors including poor attendance
Shepherding for the 15% to 20% of students who need additional supports beyond the schoolwide reforms
Intensive efforts involving specialists (counselors, social workers) for the 5% to 10% who need more clinical types of support
Balfanz, Herzog, & Mac Iver 2007

27. A Promising Path to Higher Graduation Rates
Identify those who need sustained intervention
Provide both comprehensive schoolwide reforms and more targeted and individually-focused interventions to prevent and alleviate student disengagement
A promising strategy for confronting the nation’s graduation rate crisis head on is to combine early identification of students who are slipping off the path with effective whole school reforms and individually-targeted outreach efforts to actively combat student disengagement
Balfanz, Herzog, & Mac Iver 2007