Harry Holzer—Co-Principal Investigator of the CALDER postsecondary initiative, AIR Senior Research Fellow, and Professor of Public Policy at Georgetown University, answers our questions about CALDER postsecondary research.
1. What are the goals of CALDER’s new research initiative to examine postsecondary education and the labor market?
There are two broad sets of goals: one, to better understand the determinants of postsecondary outcomes and two, to better understand which of those outcomes and credentials are rewarded by the labor market. We will focus on these issues with an emphasis on how these things relate to disadvantaged students – such as students from low-income families, minority students, as well as those that may have been weak performers in grades K-12.
We know right now that, in the United States, a lot of people attend some kind of college; but completion rates are pretty low, especially among disadvantaged workers and those within non-elite, non-flagship schools and in community colleges. We want to better understand why that is. What enables some people, especially among the disadvantaged, to complete credentials while others do not? We want to understand which credentials people can complete when they complete them, at what types of institutions, and from which populations these completers come from. We not only want to understand personal and institutional characteristics that affect those outcomes but also the types of policies, programs, and practices that seem to work as well. Understanding these different components will help us understand the policy actions we can take to try to raise low postsecondary completion rates.
The second issue is that, out of all of these credentials what does the labor market really reward? If we’re going to target and invest our resources in some credentials over others for these students, we should know what the labor market really rewards and what’s really worth investing in. We have some data on that right now from national surveys that indicate some certificate programs in more technical areas and some associates degrees pay pretty well. However, there are still a couple of big questions. One, do those results hold up when you control for the characteristics of the people who obtain these credentials? This is what economists call the problem of “selection bias.” Is it the credentials themselves driving these results, or really just the characteristics of people obtaining those credentials? When we’re looking to create effective policies, it’s very important to distinguish between these two effects. The second issue is one of short-run rewards versus long-run rewards. If we train some people for very specific occupations with specific training programs and they get a certificate or associates degree, what happens over time when they change jobs? Do they move to a different sector of the economy? What happens when the demand in the economy shifts from the sector in which they were trained to other sectors of the economy? Do those rewards hold up? To what extent are the skills that people get through obtaining these credentials portable across different economic sectors?
2. How will this new research in the postsecondary space complement CALDER’s ongoing work focusing on K-12 education and teacher effectiveness?
It’s really the logical next step in this research process. The CALDER research on K-12 education has been really productive and very informative. Now we’re trying to get a sense of how the policies and practices that seem to be effective in the K-12 years hold up across the postsecondary realm and ultimately into the labor market. To what extent do the early outcomes of young people – in terms of achievement, test scores, course completion, etc. – help determine their success in obtaining postsecondary credentials and ultimately in the labor market?
Also, if the U.S. wants to decide if we should really be investing our resources heavily in traditional academic measures of success, whether it be test scores or completing traditional core subjects, versus investing in other areas such as high quality career education, this research is a way in which we could better understand the pathways and the links between K-12 outcomes and postsecondary and labor market results.
3. What data will be used for these analyses?
This project will bring together some very new and important administrative datasets in four states. We have academic teams focusing on Florida, North Carolina, Texas and Washington state. They’re working on connecting the K-12 data with the postsecondary and labor market data. This is really the first time we’re going to have administrative data following people from kindergarten – and sometimes before kindergarten – all the way into the labor market. The administrative data on higher education will come from all of the public institutions in the states we’re looking at and the labor market data will come from unemployment insurance earnings records. Many states already collect this data and make it available to researchers, but this will be the first time this labor market data will be combined with both sets of education data. This compilation will create enormously powerful datasets for our researchers to explore the issues we’ve been talking about.
4. What are some emerging trends in this research area?
Based on the work of researchers outside of our team, we’re starting to see three broad sets of findings that we plan to shed more light on. First, there’s this question of the extent to which early outcomes are predictive of later outcomes in college and the labor market. There is some very interesting evidence from Raj Chetty of Harvard who finds that early class size effects can persist over time and can affect outcomes when students reach college. We want to examine issues like this more broadly -- to what extent do those factors have effects on later outcomes in college and perhaps in the labor market?
The second trend we’re seeing deals with the issue of completion. What enables students to complete postsecondary education once it’s been started? We’re beginning to see research findings in that area. We’re beginning to understand the role of financial constraints, different kinds of financial aid, different kinds of remediation, and providing information, counseling, and other kinds of services.
The third set of findings in the research so far has to do with labor market data. A few of our researchers have been using the data we have so far from Florida and from national datasets and what we’re finding is really interesting. We know that, on average, associates degrees pay better than certificates, and bachelor’s degrees pay better than associate’s degrees, etc. But there’s really high variance around those averages. There are some credentials from technical certificate programs that seem to pay better than associates degrees and even some bachelor’s degrees. There are even some associates degrees in the technical and healthcare fields that pay more than many liberal arts bachelor’s degrees. It’s quite intriguing to some people that this variance exists and many economists are interested in understanding how and why these students don’t respond more to those market signals.
5. What kinds of policy implications will this research have?
The policy implications of this research could be quite important. These findings could have a huge impact on higher education policy. What do we have to do if we want to raise the college completion rates of disadvantaged students? What kinds of financial aid, counseling and services should be available to these students? What kinds of curricula at these institutions help students complete their credentials? What kinds of courses should we offer and invest in and expand capacity in beyond what exists now? I think there will not only be implications for federal and state policy but also the institutions themselves in terms of improving their practices.
And of course the second set of policy implications has to do with the labor market. Ultimately what we care about the most is raising the earnings potential of disadvantaged workers when they reach the labor market. Right now, there’s some evidence that sectoral policies – where you target very specific sectors with training or career pathways, a mixture of classes and work experience – seem to deliver strong returns for low-income students. But again, we want to better see if those outcomes hold up over the long run and after we control more effectively for student characteristics. The answer has big policy implications when it comes to education and workforce policies in the U.S. How much should we be focused on general skill building as opposed to targeting specific sectors that seem to be in high demand at this point time? I think that this work will give us a better sense of how workforce policy should unfold and how workforce policy should be better integrated with higher education policy at the federal and state levels.