Friday, March 29, 2019

Maria Fitzpatrick



placeholder 

For this post, I interviewed Professor Maria Fitzpatrick. She is an Associate Professor in the Department of Policy Analysis and Management at Cornell University. In addition, she is the Director of the Cornell Institute for Public Affairs. Her research focuses on policies that affect children and families, and in particular, policies related to education. She studies them through the lens of economics.  

I interviewed Professor Fitzpatrick about her paper, “Early Retirement Incentives and Student Achievement.” In this paper, Professor Fitzpatrick and her coauthor, Michael Lovenheim, look at how these early retirement incentives (ERI) programs for teachers affect students’ academic performance. The results might surprise you.  

Professor Fitzpatrick explained to me that schools, or any employers, offer ERI’s as a cost-saving measure, since more experienced workers are usually more expensive than less experienced workers.  Instead of laying off workers, employers give them bonuses to choose to stop working, and retire early. In some cases, employers are suffering from the effects of a recession, and this is a convenient way to reduce the workforce. But there are other cases where the technology of an industry has changed over time, and newer, more recently trained workers are better-suited for certain jobs.

In the case of teachers, their pension plans are often designed so that it makes the most financial sense for them to work until a particular cutoff date, and retire on that date. Because of the ERI programs, they can now afford to retire early. 

The question these authors are asking is: how do ERI programs affect student learning? Some previous work has found that teacher effectiveness increases with experience. So we might expect that, when these more experienced teachers are replaced with new teachers, student performance would decrease. 

To test this, these authors measure performance with scores on standardized tests in math and English of third, sixth, and eighth graders. Using data from Illinois in the 1990s, they compare scores before and after ERI programs. They actually find that test scores increase as a result of the program.  
So what’s happening here? This is a policy that gives teachers a choice to retire early. While on average teacher quality seems to improve with experience, it’s likely that the least productive teachers are the ones who are taking advantage of the ERI programs: teachers who no longer enjoy teaching and are just waiting to reach that cutoff date, or maybe someone who has recently been sick and hasn’t been able to consistently come to school. We probably all have examples from our own experience of teachers who would fall into this category. Once these teachers take advantage of an ERI program, the higher quality experienced teachers remain, and test scores increase. Hence, we have an example of positive unintended consequences!

The authors also poke around at how different students might be affected differently. One important finding is that the largest increase in test scores from teacher retirements happens in schools that are in low income neighborhoods. This makes sense, because previous work has found that teachers tend to move to wealthier districts over their careers, and moving is easier for higher quality teachers.  

Even though the unintended consequences of this policy were positive, I did ask Professor Fitzpatrick if the results make her think that there might be better ways to design this policy. She told me that this study sheds light on the design of the pension plans: keeping teachers around until that cutoff date, and/or giving them an incentive to retire early may not be the best way to go when thinking about retaining the highest quality teachers. Instead, policymakers should be thinking about creative ways to retain the best teachers. 

I asked Professor Fitzpatrick about some of her other recent work. In her paper, “Pension-spiking, Free-riding, and the Effects of PensionReform on Teachers' Earnings,” she finds some very interesting unintended consequences of pension reform. Pension spiking happens when teachers work a lot towards the end of their careers, because their pensions are based on their work in these years. The government of Illinois tried to reduce pension payouts by preventing teachers from working a lot in just their last couple of years. Teachers responded by working more in the preceding years; so even though pension spiking appeared smoother, the pension fund did not save any money. 

Let’s talk! I would love to know what you think about this example of unintended consequences. Please submit comments and questions.

Wednesday, March 13, 2019

Carly Robinson



 Carly Robinson


For this post, I interviewed Carly Robinson about her paper “The Demotivating Effect (and Unintended Message) of Awards.” Carly is a PhD candidate in Education at Harvard University. In her research, she focuses on experimentally testing specific measures that can be taken to improve students’ educational outcomes.

This paper is co-authored with Jana Gallus, Monica Lee, and Todd Rogers. The award that this paper looks at is one given to middle and high school students in California for excellent attendance. Several studies have shown that students with better attendance perform better in school. In many states, average daily attendance rates are also a determinant of funding for school districts. So, schools might be motivated to increase attendance for both of these reasons. 

The standard intuition that most of us have for awards more generally is that once someone receives one for an accomplishment, we would expect them to continue to strive to achieve in that realm.  In fact, in California, the government has encouraged schools to recognize students for good attendance. These authors wanted to see how effective different types of rewards were in encouraging good attendance after the announcement of the award. So, they conducted a field experiment. A field experiment is like an experiment in the lab, but it’s done out in the real world. These researchers conducted their experiment on more than 15,000 middle and high school students in California. There were three groups of students: one group didn’t get an award (the control group), one group was told that they had the chance to earn an award for having perfect attendance in February (a prospective award for their future behavior), and one earned a retrospective award for already having perfect attendance in a previous month. The retrospective award was an unannounced, surprise award.   

The researchers expected both award groups to have better attendance than the control group. But what they found was that those students who received the surprise, retrospective award missed 8% more days of school in February than those who did not receive an award. And the attendance of the students who had a chance to earn an award in February – the prospective group – was no different than that of the control group in that month. But in March their attendance decreased. So, both types of awards actually led to reduced attendance after the award period was over.

So, what’s happening here? The researchers think that, once students earn an award, they believe that  they have at least met, and likely surpassed what’s expected of them, and so they feel licensed to miss school going forward. And, interestingly, the researchers also find that academically low performing students have more absences after the award period than their high performing peers. Since these students are probably least likely to enjoy school, they may use having earned the attendance award to negotiate missing more school going forward. 

So, our standard intuition, and California’s recommendations, don’t appear to be working as intended! And, just in case you think that you should have known better: the researchers surveyed teachers and administrators; and only 2% of them predicted that an award would reduce attendance after the award period. 

I asked Carly what recommendations she has for improving school attendance, since awards don’t seem to work.  First, she pointed out that one reason these awards might not work is that good attendance is something that students should be doing all the time, and it is therefore not really award-worthy. A better way to reduce absences is to keep parents informed. Carly says that parents chronically underestimate their children’s absences. In this paper, you can see the results of another field experiment in which Carly and her coauthors gave parents real time information about the number of days their children had missed. As a result, attendance did improve.  

Let’s talk! I would love to know what you think about this example of unintended consequences. Please submit comments and questions.