New IPR Research: June 2019

Below is this month's new research from IPR faculty.

Performance Measurement and Rewards

Overcoming Barriers to Safety Net Sign-Ups

Why do people fail to sign up for social safety net programs like the Supplemental Nutrition Assistance Program (SNAP), despite being eligible? In a working paper, IPR economist Matthew Notowidigdo and his MIT colleague Amy Finkelstein wanted to see if having additional information would increase SNAP enrollment. The researchers conducted a randomized experiment in 2016, contacting 30,000 elderly Pennsylvanians who were likely eligible for but not enrolled in SNAP. They divided potential enrollees into three groups, mailing two of them information about the program. One of these groups also received additional assistance—a toll-free number run by the nonprofit—with their application process. A third group, the control group, received neither the information nor assistance. Over the next nine months, only 6 percent of the control group enrolled in SNAP, while 11 percent of the information-only group and 18 percent of the information-plus-assistance group enrolled. The researchers interpreted their findings using an economic model and concluded that both interventions are a cost-effective way to redistribute income to low-income households, relative to other tax and transfer policies.

School Accountability and Teacher Turnover

How do teachers respond to the pressures of school accountability? To find out, IPR education economist David Figlio and his colleagues examine whether Florida’s changes to accountability measures affected teachers’ decisions to leave their schools. In 2002, the state introduced new standards for grading its schools. As a result, half of the schools received grades that differed, sometimes sharply, from those they had gotten previously. Some schools received a higher grade than they would have earned under the old system, and others received a lower grade. Teachers’ decisions to stay at their schools were unaffected in schools that would have received an “A,” “B,” or “C” grade under the original system. However, teacher turnover at schools that would have earned a “D” under the old system but received an “F” under the new accountability measures was 4–7 percentage points higher than the baseline rate. Teachers rated as above average were more likely to leave unless reforms such as smaller class sizes were made in the failing schools.

Child, Adolescent, and Family Studies

Using Text Messages to Help New Dads

According to recent estimates, about 10 percent of fathers experience depression after their child is born. Lurie Children’s Hospital Pediatrician and IPR associate Craig Garfield and his colleagues are exploring the use of a text-based intervention to reduce fathers’ depression and improve their children’s well-being. The SMS4dads program allows fathers to monitor their mood and provides them with online information. The researchers have recruited 800 fathers-to-be or new fathers from Australia, randomized into two groups. Fathers in the intervention group will receive 14 texts per month addressing their physical and mental health, their relationship with their child, and co-parenting with their partner. Fathers in the control group will receive generic health promotion texts twice each month. The study is the first trial to examine the efficacy of direct text support for men in their transition to fatherhood.

Poverty, Race, and Inequality 

Inheriting Gender Bias

Among countries with similar economic circumstances, India stands out for gender inequality and limited opportunities for women. In a recent study published in the Journal of Development Studies, IPR economist Seema Jayachandran and her colleagues investigated the intergenerational transmission of gender bias in India to determine the sources of children’s opinions about gender. Through survey data collected in 314 government schools in Haryana from nearly 5,500 adolescents and their parents, the researchers find that parents are the most significant influencers of their children’s attitudes about the appropriate roles and rights of women. According to the study, when a parent holds a discriminatory attitude toward women, their child is 11 percentage points more likely to hold the same value. Parents were shown to be more influential than peers, stay-at-home mothers more influential than working mothers, and mothers more influential than both working and non-working fathers. The study also finds that the impact of parental bias was strongest in Scheduled Caste—the “lowest” caste, also called dalit—communities, possibly due to greater social exclusion. The results of this study suggest the necessity of targeting both children and their parents in educational endeavors aimed at eliminating gender bias and providing more opportunities for women and girls in India.

Gender Inequality in the Spread of Information in Agriculture

Relying on people to spread information to each other may not help those who are less socially connected—particularly women, according to a study by IPR development economist Lori Beaman. Beaman and her Northwestern coauthor, Andrew Dillon, conducted an experiment that examined how to improve certain agricultural practices and how to spread that information to other farmers. They provided farmers in Mali with training on composting, and each of these farmers also received informational wall calendars to give to others. In Mali, men and women farm separate plots of land, but both genders are agricultural decision-makers and need to receive information. However, not all farmers ended up getting the calendars. People who were directly connected to the farmers who received the calendars were most likely to get the information, while friends of friends were significantly less likely. Overall, women were much less likely to get the information. According to Beaman, targeting the information to people who are very connected within a village also tends to leave out women, which could reinforce existing gender inequality. The results caution that while policymakers may be able to use social networks to spread information efficiently, the choice of whom to target within a network has implications for who will ultimately benefit from the information.

Building Sustainable Income for the Very Poor

Promising experimental programs to raise desperately poor people out of poverty have used a combination of providing assets such as goats, training and coaching, and access to a savings account. Is it necessary to use all of these methods together to help the world’s poorest families increase their household net worth, income, and consumption? To find out, development economists and IPR associates Dean Karlan, Christopher Udry, and their colleagues investigate the mechanisms of how each part of these multifaceted programs work. They then test whether only one of the parts, either the savings component or the grant of goats, will work as well as the combined pieces. By separating the parts in different arms of the experimental programs in Ghana, the researchers determine that the multifaceted program’s success is driven by enabling households to build profitable businesses that produce revenue. Neither component used alone led to the poor creating such businesses. The authors point out that in terms of policy, a simpler and cheaper program would be preferable but turns out not to be effective; in fact, they suggest additional components, such as mental health interventions, might increase the impact of these programs.

Safety Net Growth Since the 1990s Has Not Reached the Poorest Children

Growing up in poverty can harm a child’s health and long-term financial outcomes. The federal government invests over $200 billion in programs aimed at alleviating child poverty, which also benefits the broader economy by reducing public costs. But a working paper by IPR director and economist Diane Whitmore Schanzenbach finds that expansions of the social safety net since 1990 have gone almost exclusively to families with working parents and those with earnings above the poverty level. Additionally, spending on children as a share of total spending in the national economy has remained relatively flat. Schanzenbach and co-author Hilary W. Hoynes write that the availability of social safety net programs to low-income children has fallen, which could lead to worse outcomes in adulthood. Broadening the safety net will lead to many important benefits for individuals, including improved health, academic performance, and economic security. Schanzenbach suggests policymakers consider the long-term benefits of investments in children, which are not fully measured today but can be seen later in life.

How Status Differences in Marriage Affect Rates of Domestic Violence

Changes in societal context can tangibly affect the lives of individuals. How do we trace the effects of contextual changes in gender equality on the safety of individual women? In a recent study, IPR sociologist Julia Behrman investigates how declines in educational hypergamy, or marriages in which a woman has a lower educational attainment than her husband, influence rates of intimate partner violence. Using Demographic Health Survey data on 12,230 couples gathered between 2000 and 2010 in Kenya, Malawi, Uganda, and Zimbabwe, Behrman finds that women’s rising education levels—lessening the gap between them and their husbands—can influence their safety, but that the nature of this influence differs regionally. In Kenya, Uganda, and Zimbabwe, women’s rising status led to increased intimate partner violence, possibly due to men feeling that their social dominance was under attack. In Malawi, however, rising status led to fewer reports of domestic violence against women, perhaps due to matrilineal cultural norms and less public acceptance of violence against women. These two competing outcomes suggest that policy aiming to uplift women’s social status and standard of living must vary by region.

More Racial Discrimination in Job Offers than in Callbacks

Many previous experiments on racial discrimination in hiring focus on whether or not an applicant received an invitation to interview, often referred to as a callback, because it is much more difficult to conduct a field experiment that goes all the way to the job offer outcome. IPR sociologist Lincoln Quillian and his coauthors conduct a meta-analysis of 12 experiments encompassing more than 8,300 job applications to understand discrimination in terms of job offers, rather than just callbacks. The researchers find significant additional discrimination in hiring after the callback. White applicants received 52% more callbacks and 128% more job offers than comparable minority applicants. The results demonstrate that minority applicants face substantial additional discrimination even after they make it past the initial pile of resumes. Racial discrimination in the labor market is significantly more serious than is suggested by field experiments that focus on the callback as the outcome of interest to study.

The Intersection of Racial and Partisan Discrimination

In a new IPR working paper, IPR political scientist James Druckman and former IPR graduate research assistant Richard Shafranek explore the role race and political affiliation play in the desire of four-year colleges to communicate with prospective students. The study measured the likelihood of colleges and universities to respond to email requests for information from one of two racially distinct names, Jabari Washington or Dalton Wood, with one of four varying political mentions in the email. These mentions were no affiliation, active in a civics club, or politically involved with either a Young Democrats or Young Republicans club. The researchers find clear evidence of a racial threat bias. The average response rate to emails from the non-African American name that mentioned politics was about 75 percent; however, the response rate was roughly only 66 percent to an African-American name who referenced politics in any way (for example, expressing an interest in politics or stating a party affiliation). Additionally, there was no noticeable discrimination towards a non-political African-American prospective student. Therefore, the results suggest that, in some cases, the interaction of race and politics generates discriminatory behavior.

Urban Policy and Community Development 

More Students Carry Guns in Chicago than New York or Los Angeles

A study by community health scholar and IPR associate Joseph Feinglass finds between 2007 and 2013 more freshman and sophomore students reported carrying guns in Chicago than in New York or Los Angeles. The study, published in Injury Epidemiology, was based on self-reported data from the national Youth Risk Behavior Surveillance System, an anonymous, voluntary survey of public high school students. While self-reported gun carrying increased in Chicago between 2007–2013, it declined rapidly in Los Angeles and remained less than half the Chicago rate in New York. The 2013 prevalence of high school freshman and sophomore students who reported carrying a gun was 9 percent in Chicago, 4 percent in New York, and 6 percent in Los Angeles. When students who were exposed reported more violence risk factors, such as feeling unsafe in school, being exposed to fights, or doing illegal drugs, they were more likely to carry a gun, the study finds. Chicago’s students were exposed to more guns and these risk factors than their peers. The findings may provide insight into Chicago’s 2016 spike in gun violence, which occurred mostly among youth and young adults.

The Impact of Crack Cocaine Markets on Young Black Men

Crack cocaine markets were associated with a surge in violence in the United States during the 1980s and 1990s, and the arrival of these markets coincided with an uptick in murder rates for young black males. In a working paper, Healthcare economist and IPR associate Craig Garthwaite, William Evans of the University of Notre Dame, and Timothy Moore of Purdue University use cross-city variation in the emergence of these markets to show that the resulting violence has important long-term implications for understanding current murder rates by age, sex, and race. They estimate that the murder rate of young black males doubled soon after crack entered a city, and that these rates were still 70 percent higher 17 years after the drug arrived. The researchers document the role of increased gun possession as a mechanism for this increase. Following previous work, they show that the fraction of suicides by firearms is a good proxy for gun availability and that this variable among young black males follows a similar trajectory to murder rates. Access to guns by young black males explains their elevated murder rates today compared to older cohorts. The long-run effects of this increase in violence are substantial. Garthwaite and his co-authors attribute nearly 8 percent of the murders in 2000 to the effects of the emergence of crack markets. Elevated murder rates for younger black males continue through to today and can explain approximately one-tenth of the gap in life expectancy between black and white males.