These are comments I delivered at a Congressional briefing on prevention hosted by Rep. Bobby Scott (D, VA), on September 10, 2013.


Dropping out affects many people and has significant costs. The problem is not so large as to be out of reach of action, though.

How numerous are dropouts? The Census Bureau reports that 14 percent of 25-29 year olds do not have a high school credential. That is about 3 million people. A problem affecting one in seven young Americans is indeed a problem. Equity is an issue too. Dropout rates for black and Hispanic students (born in the US) are twice the rate for white students. Dropout rates for Native Americans are three times higher.

The costs of dropping out are large. Reasonable estimates are that high-school graduates will earn $350,000 more over the course of their working lives. If we multiply lost earnings by the number of people dropping out, we arrive at a trillion dollars. A trillion dollars, and that’s just for 25 to 29 year olds. And this is an underestimate. It does not count benefits from reduced crime and incarceration, stronger families, less transmission of poverty, healthier people, and so on.

Should we focus on preventing dropping out, or treat it instead? Wait until young people have dropped out and work with them, or prevent them from dropping out at all?

It’s common to hear that prevention is better than treatment. But it depends on the relative costs of prevention and treatment. For example, from the medical system’s perspective, preventing diabetes is relatively inexpensive compared to treating it. Preventing a cold is relatively expensive compared to treating it. Rhinoviruses are everywhere. Preventing a cold means not touching just about everything and not interacting socially. Using this framework, we should treat colds and we should prevent diabetes. Of course, ultimately we do some of both. It’s a matter of emphasis.

Dropping out is more like diabetes than a cold. It affects nearly every aspect of a person’s life, and it does not go away. A crucial difference stands it apart from diabetes. Treatment is ineffective. A dropout can return to school, maybe get a GED, and maybe get job training. None of these has been shown to improve earnings. If dropping out were a medical condition, doctors might view it as untreatable and incurable. We have to prevent it.

The incidence of dropping out is highest for ninth and tenth graders, which means they aren’t in high school long. If we want to prevent dropping out, we need to focus on middle schools.

Do middle schools already invest in preventing dropping out? We don’t know. The National Center for Education Statistics recently released its first ever study of dropout prevention programs, but the survey questions ask whether districts operate programs and not how many students participate in them. There’s not much other information about what’s happening in middle schools to prevent dropping out.  Let’s assume more needs to be done in middle schools to prevent dropping out, a point bolstered by the large numbers of dropouts and its nearly-flat trend.

We can estimate how much we might need to spend on prevention and the effectiveness we would need to make it worthwhile. Suppose we know which of the 12 million middle school students is likeliest to drop out, and we focus spending on them. Based on current data, we think 14 percent of them will drop out. That’s 1.7 million students. For the sake of argument, let’s suppose we over-identify potential dropouts and focus on 2 million students. We put identified students in groups of 20, and hire adults to support a group’s progress in school. An adult supporting progress is the basis of the ‘Check and Connect’ program. We need 100,000 adults. Suppose they are paid like teachers with 10 years of experience. That’s about $70,000 a year including benefits. We need to spend $7 billion.

How effective do these adults need to be for the investment to be worthwhile? If reducing dropping out benefits an individual by $350,000, from the standpoint of earnings, we need 1 percent of the 2 million at-risk middle schoolers not to drop out, 20,000 students. For every five adults we hire, we need one student of the hundred they serve not to drop out. Suppose 20 percent of additional earnings are returned to the government as additional taxes. The entire program breaks even for the government if, for each adult we hire, one student does not drop out.  100,000 students.

I made a leap in the argument by assuming we can identify the right students, the ones who will drop out. It’s not that straightforward. Some students with risk factors don’t drop out and some students without risk factors do. Here is where early warning systems are promising. Building on research by Robert Balfanz and his colleagues, early warning systems with reasonable predictive accuracy now can be built around four pieces of information: whether a middle school student passes math, passes reading, does not have behavior issues, and attends at least 80 percent of school days. Failing any of these conditions increases the likelihood that a student drops out of high school.  Failing all of them increases the likelihood to a near-certainty. These systems are very new, however. Most districts and schools do not use them.

I also assumed there is one prevention program, an adult to support students. In fact many approaches could be used. Which one or which combination depends on the issues facing students and frustrating their progress. Some students need reading help, others struggle in math, others struggle with family dysfunction and personal issues.

We want to know what works for different problems but we run into a significant knowledge shortfall. The Federal government spends a lot of money to support educating disadvantaged students, and a lot of money to support educating students with disabilities too. It invests comparatively little in learning about what works to help students. The budget of the Institute of Education Sciences is about $650 million (not counting stimulus spending). The budget of the National Institutes of Health is $31 billion.  In the economy as a whole, the health and education sectors are about the same size. It’s puzzling that health research receives 48 times more Federal money than education research.

Federal responsibility for research in our highly decentralized education system is clear. Individual states and districts face a free-rider problem if they do research on their own. They will pay and others will benefit, so they do fewer studies than is socially valuable. In this situation, there is a legitimate role for the Federal government to pay and everyone to benefit.

We have a significant dropout problem, ineffective treatments, and no sense that prevention is effective or even occurring. But we have a sense that the problem is at a scale that can be tackled. The effectiveness of tackling the problem does not need to be high to break even or do better, because the benefits are substantial. The above calculations of costs and benefits are rough and can be refined. The orders of magnitude will not change though the answers will be more precise.

Franklin Roosevelt said ‘It is common sense to take a method and try it. If it fails, admit it frankly and try another. But above all, try something.’ He also said ‘The school is the last expenditure upon which America should be willing to economize.’  The situation seems ripe for policy. Let’s try something.



Leave a Reply

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

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Set your Twitter account name in your settings to use the TwitterBar Section.