The Long-Term Impacts of Teachers: Teacher Value-Added and Student Outcomes in Adulthood
TL;DR: This paper found no evidence of bias in VA estimates using previously unobserved parent characteristics and a quasi-experimental research design based on changes in teaching staff, concluding that good teachers create substantial economic value and that test score impacts are helpful in identifying such teachers.
Abstract: Are teachers' impacts on students' test scores ("value-added") a good measure of their quality? This question has sparked debate largely because of disagreement about (1) whether value-added (VA) provides unbiased estimates of teachers' impacts on student achievement and (2) whether high-VA teachers improve students' long-term outcomes We address these two issues by analyzing school district data from grades 3-8 for 25 million children linked to tax records on parent characteristics and adult outcomes We find no evidence of bias in VA estimates using previously unobserved parent characteristics and a quasi-experimental research design based on changes in teaching staff Students assigned to high-VA teachers are more likely to attend college, attend higher- ranked colleges, earn higher salaries, live in higher SES neighborhoods, and save more for retirement They are also less likely to have children as teenagers Teachers have large impacts in all grades from 4 to 8 On average, a one standard deviation improvement in teacher VA in a single grade raises earnings by about 1% at age 28 Replacing a teacher whose VA is in the bottom 5% with an average teacher would increase the present value of students' lifetime income by more than $250,000 for the average class- room in our sample We conclude that good teachers create substantial economic value and that test score impacts are helpful in identifying such teachers
Citations
More filters
TL;DR: In this article, the authors test for bias in value-added measures using previously unobserved parent characteristics and a quasi-experimental design based on changes in teaching sta¤.
Abstract: Are teachers’impacts on students’test scores (“value-added”) a good measure of their quality? One reason this question has sparked debate is disagreement about whether value-added (VA) measures provide unbiased estimates of teachers’causal impacts on student achievement. We test for bias in VA using previously unobserved parent characteristics and a quasi-experimental design based on changes in teaching sta¤. Using school district and tax records for more than one million children, we …nd that VA models which control for a student’s prior test scores exhibit little bias in forecasting teachers’impacts on student achievement. Although teachers have substantial impacts,
996 citations
Posted Content•
TL;DR: It is demonstrated that kindergarten test scores are highly correlated with outcomes such as earnings at age 27, college attendance, home ownership, and retirement savings, and it is documented that students in small classes are significantly more likely to attend college and exhibit improvements on other outcomes.
Abstract: In Project STAR, 11,571 students in Tennessee and their teachers were randomly assigned to classrooms within their schools from kindergarten to third grade. This article evaluates the long-term impacts of STAR by linking the experimental data to administrative records. We first demonstrate that kindergarten test scores are highly correlated with outcomes such as earnings at age 27, college attendance, home ownership, and retirement savings. We then document four sets of experimental impacts. First, students in small classes are significantly more likely to attend college and exhibit improvements on other outcomes. Class size does not have a significant effect on earnings at age 27, but this effect is imprecisely estimated. Second, students who had a more experienced teacher in kindergarten have higher earnings. Third, an analysis of variance reveals significant classroom effects on earnings. Students who were randomly assigned to higher quality classrooms in grades K–3—as measured by classmates' end-of-class test scores—have higher earnings, college attendance rates, and other outcomes. Finally, the effects of class quality fade out on test scores in later grades, but gains in noncognitive measures persist.
995 citations
TL;DR: It is argued that skill-building interventions should target “trifecta” skills—ones that are malleable, fundamental, and would not have developed eventually in the absence of the intervention.
Abstract: Many interventions targeting cognitive skills or socioemotional skills and behaviors demonstrate initially promising but then quickly disappearing impacts. Our paper seeks to identify the key features of interventions, as well as the characteristics and environments of the children and adolescents who participate in them, that can be expected to sustain persistently beneficial program impacts. We describe three such processes: skill-building, foot-in-the-door and sustaining environments. We argue that skill-building interventions should target "trifecta" skills - ones that are malleable, fundamental, and would not have developed eventually in the absence of the intervention. Successful foot-in-the-door interventions equip a child with the right skills or capacities at the right time to avoid imminent risks (e.g., grade failure or teen drinking) or seize emerging opportunities (e.g., entry into honors classes). The sustaining environments perspective views high quality of environments subsequent to the completion of the intervention as crucial for sustaining early skill gains. These three perspectives generate both complementary and competing hypotheses regarding the nature, timing and targeting of interventions that generate enduring impacts.
369 citations
01 Oct 2014
TL;DR: This paper reviewed over 200 pieces of research to identify the elements of teaching with the strongest evidence of improving attainment and found some common practices can be harmful to learning and have no grounding in research.
Abstract: This report reviews over 200 pieces of research to identify the elements of teaching with the strongest evidence of improving attainment. It finds some common practices can be harmful to learning and have no grounding in research. Specific practices which are supported by good evidence of their effectiveness are also examined and six key factors that contribute to great teaching are identified. The report also analyses different methods of evaluating teaching including: using ‘value-added’ results from student test scores; observing classroom teaching; and getting students to rate the quality of their teaching.
321 citations
Cites background from "The Long-Term Impacts of Teachers: ..."
...Chetty et al (2011) tested teachers’ valueadded estimates to see whether they were affected by key variables that had not been included in the models and found that there was no evidence of bias....
[...]
TL;DR: In this article, the authors discuss how new data may impact economic policy and economic research and outline some of the challenges in accessing and making use of these data, and consider whether the big data predictive modeling tools that have emerged in statistics and computer science may prove useful in economics.
Abstract: Executive SummaryMany believe that “big data” will transform business, government, and other aspects of the economy. In this article we discuss how new data may impact economic policy and economic research. Large-scale administrative data sets and proprietary private sector data can greatly improve the way we measure, track, and describe economic activity. They can also enable novel research designs that allow researchers to trace the consequences of different events or policies. We outline some of the challenges in accessing and making use of these data. We also consider whether the big data predictive modeling tools that have emerged in statistics and computer science may prove useful in economics.
305 citations
References
More filters
TL;DR: In this article, a principal-agent model that can explain why employment is sometimes superior to independent contracting even when there are no productive advantages to specific physical or human capital and no financial market imperfections to limit the agent's borrowings is presented.
Abstract: Introduction In the standard economic treatment of the principal–agent problem, compensation systems serve the dual function of allocating risks and rewarding productive work. A tension between these two functions arises when the agent is risk averse, for providing the agent with effective work incentives often forces him to bear unwanted risk. Existing formal models that have analyzed this tension, however, have produced only limited results. It remains a puzzle for this theory that employment contracts so often specify fixed wages and more generally that incentives within firms appear to be so muted, especially compared to those of the market. Also, the models have remained too intractable to effectively address broader organizational issues such as asset ownership, job design, and allocation of authority. In this article, we will analyze a principal–agent model that (i) can account for paying fixed wages even when good, objective output measures are available and agents are highly responsive to incentive pay; (ii) can make recommendations and predictions about ownership patterns even when contracts can take full account of all observable variables and court enforcement is perfect; (iii) can explain why employment is sometimes superior to independent contracting even when there are no productive advantages to specific physical or human capital and no financial market imperfections to limit the agent's borrowings; (iv) can explain bureaucratic constraints; and (v) can shed light on how tasks get allocated to different jobs.
5,678 citations
"The Long-Term Impacts of Teachers: ..." refers background in this paper
...Correspondingly, these 5 needed to determine how VA should be used for education policy in a high stakes environment with multitasking and imperfect monitoring (Holmstrom and Milgrom 1991, Barlevy and Neal 2012)....
[...]
Posted Content•
TL;DR: The authors disentangles the separate factors influencing achievement with special attention given to the role of teacher differences and other aspects of schools, and estimates educational production functions based on models of achievement growth with individual fixed effects.
Abstract: Considerable controversy surrounds the impact of schools and teachers on the achievement of students. This paper disentangles the separate factors influencing achievement with special attention given to the role of teacher differences and other aspects of schools. Unique matched panel data from the Harvard/UTD Texas Schools Project permit distinguishing between total effects and the impact of specific, measured components of teachers and schools. While schools are seen to have powerful effects on achievement differences, these effects appear to derive most importantly from variations in teacher quality. A lower bound suggests that variations in teacher quality account for at least 7« percent of the total variation in student achievement, and there are reasons to believe that the true percentage is considerably larger. The subsequent analysis estimates educational production functions based on models of achievement growth with individual fixed effects. It identifies a few systematic factors a negative impact of initial years of teaching and a positive effect of smaller class sizes for low income children in earlier grades but these effects are very small relative to the effects of overall teacher quality differences.
3,882 citations
TL;DR: The authors proposed a variance estimator for the OLS estimator as well as for nonlinear estimators such as logit, probit, and GMM that enables cluster-robust inference when there is two-way or multiway clustering that is nonnested.
Abstract: In this article we propose a variance estimator for the OLS estimator as well as for nonlinear estimators such as logit, probit, and GMM. This variance estimator enables cluster-robust inference when there is two-way or multiway clustering that is nonnested. The variance estimator extends the standard cluster-robust variance estimator or sandwich estimator for one-way clustering (e.g., Liang and Zeger 1986; Arellano 1987) and relies on similar relatively weak distributional assumptions. Our method is easily implemented in statistical packages, such as Stata and SAS, that already offer cluster-robust standard errors when there is one-way clustering. The method is demonstrated by a Monte Carlo analysis for a two-way random effects model; a Monte Carlo analysis of a placebo law that extends the state–year effects example of Bertrand, Duflo, and Mullainathan (2004) to two dimensions; and by application to studies in the empirical literature where two-way clustering is present.
2,542 citations
TL;DR: This paper found large and statistically significant differences among teachers: a one standard deviation increase in teacher quality raises reading and math test scores by approximately.20 and.24 standard deviations, respectively, on a nationally standardized scale.
Abstract: Teacher quality is widely believed to be important for education, despite little evidence that teachers' credentials matter for student achievement. To accurately measure variation in achievement due to teachers' characteristics-both observable and unobservable-it is essential to identify teacher fixed effects. Unlike previous studies, I use panel data to estimate teacher fixed effects while controlling for fixed student characteristics and classroom specific variables. I find large and statistically significant differences among teachers: a one standard deviation increase in teacher quality raises reading and math test scores by approximately .20 and .24 standard deviations, respectively, on a nationally standardized scale. In addition, teaching experience has statistically significant positive effects on reading test scores, controlling for fixed teacher quality.
2,513 citations
Posted Content•
TL;DR: In this article, the authors developed estimation methods that use the amount of selection on the observables in a model as a guide to the amount that should be selected on the unobservables in order to identify the effect of the endogenous variable.
Abstract: We develop estimation methods that use the amount of selection on the observables in a model as a guide to the amount of selection on the unobservables. We show that if the observed variables are a random subset of a large number of factors that influence the endogenous variable and the outcome of interest, then the relationship between the index of observables that determines the endogenous variable and the index that determines the outcome will be the same as the relationship between the indices of unobservables that determine the two variables. In some circumstances this fact may be used to identify the effect of the endogenous variable. We also propose an informal way to assess selectivity bias based on measuring the ratio of selection on unobservables to selection on observables that would be required if one is to attribute the entire effect of the endogenous variable to selection bias. We use our methods to estimate the effect of attending a Catholic high school on a variety of outcomes. Our main conclusion is that Catholic high schools substantially increase the probability of graduating from high school and, more tentatively, college attendance. We do not find much evidence for an effect on test scores.
2,489 citations
"The Long-Term Impacts of Teachers: ..." refers background or methods in this paper
...In this subsection, we discuss four aspects of our methodology that are relevant for all the regression estimates reported below: (1) leave-out mean estimation of VA, (2) control vectors, (3) standard error calculations, and (4) the treatment of outliers....
[...]
...In every case, the magnitude of the test score change is significantly different from 0 with p 0.001 but is not significantly different from what one would 48When computing this change in mean VA, we weight teachers by the number of students they teach....
[...]
...To calculate the gains from deselecting the bottom 5% of teachers based on their estimated VA, note that (4) implies that σµ̂ = σµ √ r(nc) where r(nc) is the reliability of VA estimates based on nc classrooms of data....
[...]
...σ(2)ε using equations (2)-(4) in Kane and Staiger (2008)....
[...]
...We are forced to make this strong assumption because we have no way to estimate teacher impacts on earnings that are orthogonal to VA, as discussed in Section 2....
[...]