Skip to content

The Frontline Research & Learning Institute Logo

The High Cost of Dropouts: The Value of Early Warning Indicators to Identify Students at Risk

Student Analytics Lab Hero Image


About half of a million high school students drop out of school each year. That’s almost 1,500 dropouts a day or one every 63 seconds.

However, high school graduation is an essential milestone that has large and lasting impacts on both personal and societal wellbeing. Therefore, maximizing the number of American students who earn a high school diploma is paramount to our public and democratic health.

Personal Effects of School Attainment

Evidence suggests that a high school diploma is the ticket to future success. Greater educational attainment is associated with positive outcomes including increased job opportunities, higher likelihood of employment, and higher pay.

About 70% of jobs in the United States require a high school diploma or equivalent, making employment less attainable for high school dropouts.

As a result, less than half of high school dropouts were participating in the workforce between 2021 and 2022, compared to over two-thirds of high school graduates who were not enrolled in college. And even when high school dropouts find employment, they earn much less than their high school and college-educated peers.

In 2021, the median annual earnings for 25 to 34-year-old full time employees with a Master’s degree was more than double the median annual earnings of those who had not graduated from high school.

But the financial impact is only the beginning – school attainment is also associated with health outcomes.

Those who complete more schooling tend to live happier, healthier, and longer lives. For instance, an additional four years of education is associated with a lower prevalence of heart disease and diabetes and on average, 25-year-old men with a bachelor’s degree outlive those without a high school diploma by nearly a decade (Hummer & Hernandez, 2015).

Further complicating the health-related implications, adults without a high school diploma are four times as likely to lack health insurance as those with a bachelor’s degree. Associations between school attainment and mental disorders (Breslau, Lane et al., 2008) and school attainment and incarceration (Duke, 2018) have also been detected.

Main Takeaway: A high school diploma is a ticket to future success.

Societal Effects of School Attainment

That ticket to future success is not an exclusive admit-one for the student who graduates. That’s because the impact of a single student graduating reverberates throughout the community (and even society) at large.

A higher graduation rate is associated with lower unemployment, higher wages, increased tax contributions, decreased demand for social services, decreased crime, and increased political participation (Belfield & Levin, 2007; Levin et al., 2007).

Moretti (2007) calculated that increasing the high school completion rate of all men ages 20-60 just 1% would save the US approximately $1.4 billion each year. When you account for inflation, those savings increase to about $2 billion dollars.

Improved education could also save taxpayers billions in public assistance costs like those made available through cash, food, and housing support programs (Magnuson et al., 2007).

Main Takeaway: Higher graduation rates are the ultimate win-win: better for the graduate, and better for society at large.

State Accountability for Increasing the Graduation Rate

Acknowledging the advantage that a high school diploma provides, the Every Student Succeeds Act (2015), requires states and local educational agencies to report adjusted cohort graduation rates (ACGR) and develop their own strategic plans for boosting them.

The ACGR is calculated by dividing the sum of the number of students who earn a regular diploma in four years and the number of students with cognitive disabilities who graduate with an alternate diploma, by the number of students in the entire cohort for the graduating class. The law also requires that the ACGR be disaggregated by the following 6 subgroups:

  1. People from every major racial and ethnic background
  2. Students who qualify as economically disadvantaged
  3. Children with disabilities
  4. Students learning English as an additional language
  5. Students who are experiencing homelessness
  6. Children who have experienced foster care

Understanding Student Risk

Several risk-factors, including membership to some of these sub-groups, have been identified as early warning indicators for school drop-out.

Status risk factors include demographic and historical aspects that are predictive of dropout but cannot be changed (Clemens et al., 2020), like qualifying for free or reduced-priced lunch, having a disability, or being an English learner (DePaoli et al., 2018).

Alterable risk factors, like a students’ attendance, behavior, and course performance, are those that can be reduced through intervention strategies, including policy and/or practice changes (Clemens et al., 2020).

These factors, commonly referred to as “The ABCs,” are the most predictive of high school dropout (Allensworth, 2013; Mac Iver & Messel, 2013) and are used in early warning systems to systematically predict and improve student outcomes.

See Table 1 for examples of data used to predict a student’s risk level for each indicator.

Table 1 – Dropout Risk Factors Related to the ABCs


  • Frequent tardiness
  • High absenteeism rates
  • Truancy


  • Behavior difficulties in school and in community
  • Frequent discipline referrals
  • Frequent suspensions or expulsion

Course Performance

  • Poor academic achievement, based on grades and scores
  • Over-age for grade level
  • Low student engagement
  • High rates of retention
  • Low credits
  • Low homework completion

Note: See Freeman et al., 2015.  

Early warning systems can be implemented as early as kindergarten and use predictors to begin forecasting which students will be on-track to graduate high school. While that might seem much too early to know how students will develop, catching students early also gives them the best chance at positive future outcomes.

That’s the goal: intervene as early as possible so that even students with irreversible predictors like being an English learner don’t lead to a drop out. On that note, it’s important to remember that indicators don’t cause the outcome (ever heard correlation does not equal causation?).

Well-kept systems populated with robust data that is easy to navigate and interpret can help districts identify those who are not on track for the purpose of providing targeted interventions. It is important to note however that “Early warning indicators are used only for prediction—they do not cause students to drop out. Rather, they should be treated as symptoms of the dropout process that is in progress” (Scala, 2015, p. 8).

It is important to note however that “Early warning indicators are used only for prediction—they do not cause students to drop out. Rather, they should be treated as symptoms of the dropout process that is in progress” (Scala, 2015, p. 8).

Stressful life events related to housing, money, criminal or legal issues, accidents or health problems, suspensions, pregnancy, and personal relationships amplify potential for dropout (Dupéré et al., 2018).

Identifying a causal link between a single risk factor and dropping out is difficult. According to Freeman et al., 2015, “the likelihood that a student will drop out increases when multiple risk factors are present” (p. 291). Therefore, a composite of students’ risk factors should be created to inform intervention efforts.

Profiles of Students at Risk

Much research on dropout positions it as a binary construct, grouping all dropouts in a single category to compare against graduates.

More recently, researchers have acknowledged that nuances within the group exist, suggesting that students disengage with school for different reasons, their disengagement manifests in different constellations of risk-factors, and efforts to re-engage them should vary accordingly.

An analysis of nearly 2,000 American students who dropped out of high school resulted in the identification of three significantly different classes or types of dropouts (Bowers & Sprott, 2012).

The three student profiles below illustrate these classes.

Brian – Obviously At-Risk

Low Medium High
Attitude towards school Low Type Icon
Test scores Low Type Icon
Grades Low Type Icon
Credits Low Type Icon
Homework Low Type Icon
Extracurricular Low Type Icon
Reading outside of school Low Type Icon
Attendance High Type Icon
Behavior High Type Icon
Suspensions High Type Icon
  • The bell signaling the beginning of class rang at least five minutes ago, and Brian slowly and reluctantly walks to his desk, which is directly in front of his math teacher’s.
  • The teacher, who is already circulating the classroom checking for last night’s homework expects his tardiness at this point. Still, he greets Brian and suggests that he take out his homework before continuing with his rounds.
  • Brian doesn’t have his homework today. In fact, he hasn’t completed a single assignment since school started nine weeks ago.
  • When Jason, who sits next to Brian asks why he was late to class, Brian shoves him causing him to fall onto the tile floor. Brian knows that he’ll be written up again and will likely land in in school suspension, but that might be better than sitting through another math class.
Note: Adapted from Bowers & Sprott, 2012. The “Low,” “Medium,” “High” distinctions are in comparison to the other dropout types and not to school completers.

Nina – Quiet

Low Medium High
Attitude towards school High Type Icon
Test scores Low Type Icon
Grades Medium Type Icon
Credits Medium Type Icon
Homework High Type Icon
Extracurricular Medium Type Icon
Reading outside of school High Type Icon
Attendance Medium Type Icon
Behavior Low Type Icon
Suspensions Low Type Icon
  • As soon as she descends the bus steps, Nina rushes to her English teacher’s classroom. She looks forward to chatting with her teacher who is also the faculty advisor of her creative writing club which meets once a week after school.
  • Sometimes, she uses the time in between bus drop off and first bell to finish her homework. Though she emigrated from Haiti four years ago, and is now a proficient English speaker, homework that is reading-heavy continues to challenge her.
  • This is especially true in subjects that depend heavily on technical and domain-specific vocabulary, like science. She’s proud of the work she’s doing in English class but is in danger of failing science, math, and history.
Note: Adapted from Bowers & Sprott, 2012. The “Low,” “Medium,” “High” distinctions are in comparison to the other dropout types and not to school completers.

Kabree – Involved

Low Medium High
Attitude towards school High Type Icon
Test scores High Type Icon
Grades High Type Icon
Credits High Type Icon
Homework High Type Icon
Extracurricular High Type Icon
Reading outside of school High Type Icon
Attendance Low Icon
Behavior Medium Type Icon
Suspensions Low Type Icon
  • Kabree has been a cheerleader since kindergarten and this year she earned a spot as a flyer on the varsity cheer team. The team practices every day after school and cheers year-round for the football and basketball teams.
  • She refuses to miss a day of school, unless she’s too sick, because she can’t attend practice if she’s absent and loses next-day-game cheering eligibility. Her school requires student athletes to maintain a 2.0 grade point average to participate in athletics, but her coach requires a 3.0.
  • With the support of her coach and teammates, Kabree maintains mostly A’s and B’s.
  • A perfect storm of team interpersonal conflict, financial stressors at home, and unrelenting social media bullying causes a series of lunchroom outbursts which both led to disciplinary referrals and temporary probation from the cheer team.
Note: Adapted from Bowers & Sprott, 2012. The “Low,” “Medium,” “High” distinctions are in comparison to the other dropout types and not to school completers.

Create Your Own Early Warning System

Districts can take a proactive approach to identify students, like Brian, Nina, and even Kabree, who are at-risk of dropping out by analyzing the student-level data they already collect related to the early warning indicators.

However, collecting data points on multiple indicators for several hundred if not thousand students each year can quickly inundate a district making it difficult to glean meaningful insights and identify those in immediate need of intervention.

For every student, districts document grades, achievement test scores, diagnostic and progress monitoring assessment scores, attendance records, discipline referrals, detentions, suspensions, demographic information, Individualized Education Plan status, English learner status, and extracurricular activity involvement.

Early warning systems depend on accurate and efficient data compilation and analysis. However, student-level data is often input by different people and stored in different spreadsheets or databases, making it more challenging for districts to analyze the risk-factors in concert and create composite profiles of at-risk students.

Software may assist in organizing student-level data and automating risk-analyses for each student based on data inputs.

Figure 1 displays an example of one district’s early warning system that was created by a software analysis tool using the district’s existing data. The system provides a score ranging from 0-2 for each indicator based on district-established thresholds. A score of 0 indicates the lowest risk and a score of 2 indicates the highest risk for each indicator. In this district, students earn a 0 for coursework or grades if they have A’s, B’s, and C’s only. Students like Nina who have two or more D’s and/or F’s earn a 2 indicating a higher level of risk.

Figure 1 – District-Level Early Warning System Dashboard

District-Level Early Warning System Dashboard -Image Placeholder

With a student analytics software system, districts can easily review historic data to ensure that their established thresholds are effectively identifying dropouts and those who continuously struggle with disengagement.

For instance, if your district decided to use less than 90% attendance as your metric for highest risk but then discovers that several students with 94% attendance dropped out the previous year, you can adjust the threshold to catch future dropouts more accurately.

An analysis tool can also provide a composite risk-level based on each student’s scores for the three indicators combined.

For this district, the levels range from No Risk to High Risk. Students who earn 0 points across the three indicators are marked as “No Risk” whereas students who earn five or more points are marked as “High Risk.”

Using this software, teachers and administrators can quickly identify those most at risk and intervene quickly.

Figure 2 shows that the system classifies Kabree as low risk, but not no risk, because while her academics and attendance are on track, her recent disciplinary episodes earned her a point for behavior.

Figure 2 – Student Risk Analysis

Student Risk Analysis -Image Placeholder

A More Holistic View of Students

Using a comprehensive student-analytics software, teachers and administrators who want more information about a specific student, like Kabree, can easily access it by clicking on his or her name. The action will transport them to a student report page, like the one depicted in Figure 3, that presents all available data on the selected student.

Figure 3 – Student Report Page

Student Report Page -Image Placeholder

To contextualize a student’s performance on a particular indicator, a district can zoom out to district-level data and see how the student compares to his/her peers.

For instance, Figure 4 displays the student-level behavior data across an entire district.

Figure 4 – District-wide Student-level Behavior Data

District-wide Student-level Behavior Data -Image Placeholder

The benefits associated with earning a high school diploma are well-documented and schools are poised to help students achieve this essential milestone. Districts are already doing a lion’s share of the work by collecting data on early warning indicators for each student. More than that, there are districts and school leaders that know their students by name, strength and need.

They are the real heroes aggregating street-level data that goes beyond what an early warning system can provide. By taking it just one step further and adopting a software analysis tool, they can turn disparate data points into meaningful insights about those students who are on track to graduate and those who are at-risk of dropping out.

This is just one way district offices can support campuses in their work of knowing every young person by name, strength, and need. Prioritizing speed to insight around those who are struggling within the organization allows a building to focus its resources on fostering an environment more comprehensive in its’ development and nurturing of all young individuals.

Best Practices and Strategies to Address Drop Out Rates

As part of a comprehensive approach to dropout prevention, schools need to employ strategies that not only identify students at risk using early warning indicators, but also provide actionable steps to keep these students engaged and motivated.

Several evidence-based strategies have been found to be successful in reducing dropout rates. Let’s explore 5 of them.

  1. Targeted Interventions: Once students have been identified as at-risk using early warning indicators, they should be placed in targeted intervention programs. These interventions should be specifically designed to address the students’ individual needs in terms of academics, behavior, and/or attendance. The interventions can include tutoring, mentoring, counseling, and other support services. Importantly, interventions should be proactive, not reactive, seeking to support students before they reach a critical point of disengagement.
  2. Mentoring Programs: Establishing a school-based mentoring program can be an effective strategy to address students at risk of dropping out. Mentors, whether teachers, counselors, or trained peers, can provide at-risk students with additional academic support, guidance, and encouragement. Mentors can also serve as positive role models, fostering a sense of belonging and connection to the school.
  3. Family Engagement: Engaging families is an important aspect of reducing dropout rates. Schools should work to build strong relationships with families, involving them in their children’s education, and providing them with resources to support their child’s learning at home. When parents are actively involved in their children’s education, students are more likely to remain engaged and stay in school.
  4. Supportive School Climate: A supportive and inclusive school climate can contribute significantly to reducing dropout rates. This includes fostering positive relationships between staff and students, promoting a sense of safety and respect, and providing a welcoming and inclusive environment for all students.
  5. Career and Technical Education (CTE) Programs: CTE programs can make education more relevant and engaging for students by providing them with practical, hands-on learning opportunities. These programs can help students develop in-demand skills for high-growth industries, increasing their motivation to stay in school and their chances of post-secondary success.

A proactive approach that combines early identification of at-risk students with effective, evidence-based interventions can significantly reduce dropout rates. Utilizing a student analytics software system to organize available student data can greatly assist in the process of utilizing EWI data to identify students who may need support, but ultimately a successful drop out prevention program requires a concerted effort from educators, administrators, parents, and students themselves, but the benefits – both individual and societal – make this effort invaluable.

Recommended Resources

  • Mercedes ISD Image

    Case Study Data-informed Decisions with Campus-level Dashboards

    How Mercedes ISD empowers district and campus leadership with easily accessible student data.

  • Back to School Data Image

    Podcast Deep in Data, Short on Time: How Analytics Can Do Your Legwork

    Explore the single biggest question that district leaders can ask and answer with data to drive greater student outcomes and achievement

  •  Image

    Webinar Planning for Next Year: Student Performance Data Informing Summer Conversations

    See how districts are blending last year’s performance, state or national assessment and other student data sets to inform summer professional learning plans and identify students who need additional support.



  1. Allensworth, E. M. (2013). The use of ninth-grade early warning indicators to improve Chicago schools. Journal of Education for Students Placed at Risk (JESPAR), 18(1), 68–83.
  2. Belfield, C. R., & Levin, H. M. (Eds.). (2007). The price we pay: Economic and social consequences of inadequate education. Brookings Institution Press.
  3. Bowers, A. J., & Sprott, R. A. (2012). Why tenth graders fail to finish high school: A dropout typology latent class analysis. The Journal of Education for Students Placed at Risk (JESPAR), 17(3), 129-148.
  4. Breslau, J., Lane, M., Sampson, N., & Kessler, R. C. (2008). Mental disorders and subsequent educational attainment in a US national sample. J Psychiatr Res., 42(9), 708-716.
  5. Clemens, E. V., Lalonde, T., Klopfenstein, K., & Sheesley, A. (2020). Early warning indicators of dropping out of school for teens who experienced foster care. Child Welfare, 97(5), 65-88.
  6. DePaoli, J. L., Balfanz, R., Atwell, M. N., & Bridgeland, J. (2018). Building a Grad Nation: Progress and Challenge in Raising High School Graduation Rates. Annual Update 2018. Civic Enterprises.
  7. Duke, B. (2018). A meta-analysis comparing educational attainment prior to incarceration and recidivism rates in relation to correctional education. Journal of Correctional Education, 69(1), 44-59.
  8. Dupéré, V., Dion, E., Leventhal, T., Archambault, I., Crosnoe, R., & Janosz, M. (2018). High School Dropout in Proximal Context: The Triggering Role of Stressful Life Events. Child Development, 89(2), 107–122.
  9. Freeman, J., Simonsen, B., McCoach, B., Sugai, G., Lombardi, A., & Horner, R. (2015). An analysis of the relationship between implementation of school-wide positive behavior interventions and supports and high school dropout rates. The High School Journal, 98(4), 290-315.
  10. Hummer, R. A., & Hernandez, E. M. (2015). The effect of educational attainment on adult mortality in the United States. Popul Bull, 68(1), 1-16.
  11. Levin, H. M., Belfield, C. R., Muennig, P., & Rouse, C. (2007). The Costs and Benefits of an Excellent Education for All of America´ s Children.
  12. Mac Iver, M. A., & Messel, M. (2013). The ABCs of keeping on track to graduation: Research findings from Baltimore. Journal of Education for Students Placed at Risk (JESPAR), 18(1), 50-67.
  13. Magnuson, K. A., Ruhm, C., & Waldfogel, J. (2007). Does prekindergarten improve school preparation and performance? Economics of Education review, 26(1), 33-51.
  14. Moretti, E. (2007). Crime and the costs of criminal justice. In C. Belfield, & H. Levin (Eds.), The price we pay: Economic and social consequences of inadequate education (pp. 142–159). Washington, DC: Brookings Institution Press.
  15. Scala, J. (2015). Early warning systems. Rural Dropout Prevention Project, American Institutes for Research (AIR). Retrieved from