Identifying Excellent Teachers

Higher educational quality tends to reduce unequal outcomes in society (Jackson, Johnson, and Persico 2015), promote greater innovation (Aghion et. al. 2005), increase worker productivity (Moretti 2004), and promote good democratic citizenship (Dee 2004). As such, public education has played a central role in domestic policy for decades. Recent education policy in the United States has focused on improving teachers’ instructional quality (Jackson 2012). Even though teachers are tasked with imparting subject knowledge to students, at the broadest level a quality teacher is one that is able to provide students with the skills to be productive workers, good citizens, and happy adults (Douglass 1958, Jackson et. al. 2014). Using policy levers to attract, retain, and create high-quality teachers requires that we can identify them. To this aim, there has been much work toward identifying teachers that are able to raise cognitive skills, as measured by standardized test scores. However, recent evidence from psychology and economics shows that “soft” skills (as measured by both a broad set of behavioral outcomes and psychometric measures of traits, practices, and attitudes) are important for success in adulthood but are not well measured with standardized tests. Unfortunately, little work has been done toward identifying teachers that impact the broader set of skills likely needed to be productive members of society. My research seeks to provide evidence on how to better identify high-quality teachers that improve the broad set of skills students need to succeed into adulthood. I seek to answer five foundational questions. (1) How much do teachers improve a broad set of skills required for adult success? (2) How can we identify excellent teachers that improve both students’ test scores and also their soft skills? (3) What is the causal effect of having a middle-school teacher that improves soft skills on long-run adult outcomes such as college going, crime, employment and earnings? (4) How much better can we identify teachers who improve adult outcomes using both their effects on test scores and “soft” skills than using test scores alone? (5) How much can improvements in teacher quality reduce gaps in both childhood skills and adult outcomes across racial, ethnic, and income groups? This research requires uniquely detailed data on a large population of individual students (that includes both test scores and measures of soft skills) that are linked to other administrative datasets with key adult outcomes. Currently, no such dataset exists. In collaboration with the Chicago Consortium of Public School Research, I will compile such a uniquely comprehensive longitudinal data set. With these data, we will employ insights from economics and psychology to use these unusually rich data to create a holistic set of measures of student skills. We will then document how these measures of skills in childhood predict adult success. We measure adult success by educational attainment, criminal activity, employment, and earnings—outcomes that are indicative of being a productive member of a democratic society. We will employ value-added methods developed by statisticians and economists to estimate the causal effect of individual teachers on these skills, and document the extent to which teachers who raise test scores are also those who raise “soft” skills. Finally, we will document the extent to which teachers who improve these broad skills in childhood have a causal effect on these important adult outcomes, and document how the distribution of teachers’