Ask the Experts: Amy Wolaver on the economics of healthcare
October 21, 2011
Please note: You are viewing an archived Bucknell University news story. It is possible that information found on this page has become outdated or inaccurate, and links and images contained within are not guaranteed to function correctly.
LEWISBURG, Pa. — Amy Wolaver, associate professor of economics, talks about how changes in Medicaid coverage influence family planning and the relationship between binge drinking and college grade point averages.
Q: You study changes in Medicaid coverage for family planning services and how they affect women's sexual activity. What are some of the issues you examine?
A: Medicaid is the public health coverage for predominantly low-income, low-asset people. The federal government sets minimum standards of eligibility and decides what benefits are covered. The amount of these benefits paid for by the federal government as opposed to states is called the federal matching rate, and that matching rate differs by the kinds of services offered. States can be more generous in their Medicaid coverage than the federal minimum if they pay for the extra coverage, and they also can apply for what's called a Section 1115 waiver to receive additional matching funds in certain areas. Under these waivers, if states receive additional funds in one area, they have to make up for it with decreased reimbursement in other areas. So it has to be budget neutral from the federal government's point of view. Starting in the early 1990s, some states began to apply for and receive section 1115 waivers for family planning services as a way of expanding the number of women who can receive services such as contraceptives, a yearly checkup and sexually transmitted disease testing. The states that applied for this funding argued that increasing coverage for these services is budget-neutral because it saves in the long run by preventing unwanted pregnancies.
In my research, I look at data over time on young women in states that have these waivers and states that don't. There is a lot of variation about when the states put the waiver in and how they structured it. What I looked at was whether women with access to additional coverage changed their rate of sexual activity. One of the arguments against increasing Medicaid coverage for family planning services is that if you give birth control to women, particularly to teens, they are going to increase their rate of sexual activity. I wanted to look at that and see if that was true. I also wanted to see if these women increased their use of contraception, and if they did, what did it do to the rate of pregnancies? The data I am using looks at the same people over time, so I know if they reported a pregnancy in one year, if there's a baby in the following time period or if something happened in between.
Q: What has your research shown?
A: What I found is that there was no increase in sexual activity with the increased Medicaid coverage, but there was an increase in consistent use of birth control and big drops in the probability of pregnancy. Most of the drops are pregnancies that I predict would not have ended up with a live birth. Since I don't have a direct measure of abortions, this outcome is one that encompasses abortions and miscarriages — and the policy seems to reduce pregnancies that would end with one of these outcomes. My study is different than most studies on this policy because I am looking at the same women over time. There are other studies of these policies that look at either the state average birth rate before and after the policy, or they look at Medicaid recipients.
Q: You also have examined college binge drinking and how it affects GPA and study hours. What group were you looking at and what were your conclusions from that research?
A: In my research on binge drinking and GPA, I looked at data collected by Harvard University, which used to conduct a college alcohol survey. The survey included a nationally representative sample of students from about 140 colleges and asked them about their drinking behavior, drug use and other risk-taking activities. The survey also included students' self-reported GPAs and their self-reported hours of study. I looked at the survey data from 1993, 1997 and 2001, and determined that binge drinking, with all else being equal, lowers the GPA of the average student by about a third of a point on a four-point scale. The impact on study hours was odd. It wasn't always consistent and in some cases even seemed to increase study hours. But, study hours are not a very good measure of student effort. On the one hand, really able students can perform well with less time. But as I observe students today, if they are studying for five hours but their computer is on and they are checking e-mail every three seconds, it's not a great measure of the quality of the study effort.
The measure that is used in this data for binge drinking is five drinks in one sitting for males and four drinks in one sitting for females. Some people quibble about whether that is the right number to use, but if you look at college-aged students, that level of drinking is associated with an increased risk of violence, crime, injury, vandalism, having sex without planning to and having sex without protection.
Q: As an economist, you are measuring human behavior. How does that differ from more traditional medical research in this realm?
A: We are definitely studying human behavior, but what I am doing is based on observational data. I am looking at how human beings are responding to marginal costs and marginal benefits. If you change the cost of birth control by increasing access or decreasing it, how much of a behavioral change will that make? These questions are interesting because you are asking if people will abstain from sex if they know there is a risk of pregnancy and/or STDs, or — if they have contraception, which lowers the probability of those bad things happening — are they more likely to have sex.
If you are doing medical clinical research, you can randomize people into treatment and control groups. Usually economists are working with observational data. So we see what happens to people given the choices they have already made, and we have to understand they made those choices in a context. If I observe, for instance, in the study of binge drinking that the heaviest student-drinkers have lower GPAs, that doesn't mean the drinking caused the GPA to be lower. There could be some other factor that is simultaneously leading that student to be a heavy drinker and have low grades. Or, it could be that a student was depressed because of a low grade and drank a lot. I am trying to figure out if there is something else hidden that's affecting both outcomes.
For the Medicaid research, I can make a pretty good argument that these state policies are a random treatment for these women in terms of those eligible face lower contraceptive costs than those otherwise similar women who were not eligible. The method I used to try to figure out the true causal impacts of changing the price of contraception is known as a natural experiment or a quasi-experimental design. States are doing this on their own, but I can make an argument that it's random to the individual women in the sample. That allows me to act as if I have done a clinical trial where the treatment is cheaper contraception. The trick in a lot of the work I do is to try to figure out the cause from correlation.
Interviewed by Julia Ferrante
New editions of "Ask the Experts" will appear on the Bucknell website on most Thursdays during the fall and spring semesters. If you have ideas for future topics or are a faculty or staff member who would like to participate, contact Molly O'Brien-Foelsch.
To learn more about faculty and staff experts who can speak on a variety of news topics, visit Bucknell's searchable Experts Guide.
The following links are virtual breadcrumbs marking the 27 most recent pages you have visited in Bucknell.edu. If you want to remember a specific page forever click the pin in the top right corner and we will be sure not to replace it. Close this message.