A partner asked me to help improve high school graduation rates. My first request was a list of 50 students who had dropped out.<\/p>\n\n\n\n
About 80% of students were already graduating. There were likely many things working simultaneously to produce that result, making it difficult to know which were most important. I suspected we would find a clearer signal among the students who had left.<\/p>\n\n\n\n
Our team interviewed all 50, asking what had led them to stop coming to school. A few patterns emerged. We then surveyed a much larger sample of students who had dropped out to test those patterns and identify anything we had missed.<\/p>\n\n\n\n
The majority of students who dropped out described one or more of the following:<\/p>\n\n\n\n
We developed strategies to address these patterns. Within 12 months, retention rates, my partner\u2019s strongest short-term indicator of progress toward graduation, had improved.<\/p>\n\n\n\n
I almost left those three findings out. In education, we sometimes turn local findings into universal explanations, then build or borrow solutions around them. These findings described a particular group of students in a particular context. What I hope you take from this story is the approach:<\/p>\n\n\n\n
When most of your target population is already getting the desired result, the clearest actionable signal may be among the minority who are not.<\/em><\/p>\n\n\n\n Studying the 80% who graduate can produce a long list of factors that might contribute to success. Studying the 20% who leave can reveal a smaller number of specific breakdowns in your context. That shorter list can give you a much clearer place to start.<\/p>\n\n\n\n I have used versions of this approach elsewhere, including recently on chronic absenteeism. We asked students and families what was keeping them from school and looked for patterns in what they told us. Responding to those patterns improved attendance.<\/p>\n\n\n\n If you have a program or intervention that is working, say, 75 to 95 percent of the time, what might you learn by talking with the group for whom it isn\u2019t working and looking for patterns in what they tell you?<\/p>\n\n\n\n Reply if you like. I\u2019d love to hear what this brings up for you.<\/p>\n\n\n\n
\n\n\n