Alana Barton
Media Technical Advisor
PRB staff offer summaries of new research and insightful sessions from the annual demography meeting.
April 27, 2023
Media Technical Advisor
Senior Program Director
Associate Vice President, U.S. Programs
Research Analyst
Senior Writer
Research on the impact of the COVID-19 pandemic, census data issues, and reproductive health were among the many topics discussed at the Population Association of America meeting in New Orleans in mid-April. The annual conference offered researchers and practitioners a platform to discuss new work in 265 sessions and more than 800 poster presentations.
Below, we highlight some sessions attended by PRB staff. Most of the research studies described are ongoing and not yet published; the full conference program can be accessed here.
In 2020, a COVID-19 infection in pregnancy was associated with a higher risk of preterm or very preterm childbirth, or birth occurring between 28 and 37 weeks of gestation, found an analysis of California data by Florencia Torche of Stanford University and Jenna Nobles of the University of Wisconsin-Madison. The impact of a COVID-19 infection on premature births declined in 2021 and disappeared in 2022, they found. Torche and Nobles suggest that the waning impact of infection may reflect new variants of the virus and access to vaccines and antiviral therapies, underscoring the importance of vaccination to reducing the burden of infection on vulnerable groups.
The COVID-19 pandemic threatened to derail gains in family planning programs in many low- and lower-middle incomes countries. A global team led by Avenir Health examined the impact in 15 countries in Asia and sub-Saharan Africa using service statistics to calculate the annual estimated number of modern contraceptive users in 2019 and 2020. They found that only three countries (Madagascar, Pakistan, and Zimbabwe) saw declines in the total number of estimated modern users, mainly driven by declines in the use of short-term methods. Contraceptive implant programs grew in 12 countries and the number of estimated injectable users increased in nine countries. While COVID-19 caused temporary disruptions to family planning services in most of the countries, most were able to continue to provide commodities, leading to a lower negative impact than initially feared, they found.
For 2020 Census data, the U.S. Census Bureau is applying a relatively new disclosure avoidance framework called differential privacy that protects respondent confidentiality by inserting random noise into the data. In a lively session, census data experts representing state government agencies and academia expressed concerns about the effects of the new privacy-protection procedures on 2020 Census data products, including:
Sallie Ann Keller, Chief Scientist at the Census Bureau, described their challenges in balancing data privacy, accuracy, and availability; she also said the Census Bureau is committed to involving stakeholders in the planning process. PRB is also working in partnership with the Census Bureau to produce a series of user-friendly briefs that describe how disclosure avoidance procedures are being applied to 2020 Census data products and the implications of for data users. For more information, see Disclosure Avoidance in the 2020 Census: What Should Data Users Know About Respondent Privacy and Data Accuracy?.
Many observers point to the shift to more childbearing at older ages in the United States to explain the rise in severe pregnancy complications (known as severe maternal morbidity). But Alison Gemmill and colleagues at Johns Hopkins University find that the upward trend is driven by increases in severe complications at all ages. The role of increased maternal age was minimal for all subgroups except Black women, where births at older ages represented 17% of the rise in severe complications, they reported.
Among women in low- and middle-income countries who stopped using contraception for method-related reasons, the monthly probability of resuming contraception was higher than the probability of pregnancy in the year after discontinuation, a team of researchers found. Alison Gemmill and Dana Sarnak of Johns Hopkins University, Sarah Bradley of Abt Associates, and Eve Brecker and Kaitlyn Patierno of PRB used pooled data from Demographic and Health Surveys in 49 countries after 2009. On average, close to half of women who discontinued for method-related reasons resumed contraception within three months, they found, but there was wide variation across settings. Country-specific analyses showed that the probability of resuming contraception within three months ranged from 15% in Liberia, Mali, and Sierra Leone to 85% in Bangladesh. Many women do not resume contraception and continue to be at potential risk of unintended pregnancy post discontinuation, they point out. Policymakers and program planners should continue focusing on interventions that meet the contraceptive needs of women who want to avoid pregnancy, including initiatives that strengthen contraceptive counseling as well as those that ensure different contraceptive options are available.
Migration is an important driver of county-level growth in the United States, and tax dollars going to storm relief—including relocating victims—have risen steadily during the past several decades. Despite this, little research has been done to collect data around how tropical storms affect excess migration, or migration that wouldn’t have happened without climate-related factors. And the relative risk of experiencing a flood migrants faced before and after migration has not been included in prior studies, reports Eugenio Paglino at the University of Pennsylvania, who worked to measure both these phenomena. While the number of people at risk from tropical storms has risen, Paglino’s study showed small rates of excess migration after storms. He also found that migrants’ risk levels did not decrease, since they tended to migrate to other high-risk areas. One significant limitation of the study was lack of recent data: Paglino measured migration between 1990-2010 due to a lack of available data from his primary source, the Internal Revenue Service (IRS).
In a health survey question about physical activity, the examples of exercise listed were golf and bowling. Are those examples broadly representative, or just representative of the people who were in the room when the questionnaire was written? Might people who play soccer or basketball, or those who ride a bicycle to work, be missed? At a session hosted by the Census Quality Reinforcement Taskforce, Census Bureau Director Robert Santos and U.S. Deputy Chief Data Scientist Dominique Duval-Diop spoke about the importance of having a diversity of voices behind the scenes in the collection, production, and publication of official statistics. Both gave concrete examples of why representation matters—from survey question phrasing to choosing which data are used in disaster recovery fund distribution. Both explained how they are working to bring more voices to the table and why diversity matters for good decision-making.
How should we define motherhood? If you’re a researcher studying the wage penalty for women with children, the definition you use will likely affect your results, according to Alexandra Killewald at Harvard University. The motherhood wage penalty describes how women with children earn less than childless women with similar characteristics. And, according to Killewald and colleagues, the wage penalty is larger if you use a more inclusive definition of motherhood that includes not only mothers living with their children, but also moms with older children who have left the home.
The California Complete Count Committee, an advisory panel established by California’s Census 2020 Office, commissioned the RAND Corporation to conduct an independent survey of 173 census blocks to evaluate the data accuracy of the latest decennial Census. The survey found that while the total population of blocks was comparable to official census numbers, subpopulation totals were highly variable. Researchers found evidence suggesting Asian people and people with broadband internet access were more likely to be over counted.
At the beginning of the pandemic in 2020, Americans on average got more sleep and got a bit more exercise than in 2019, Sarah Flood, Janet Wang, and Phyllis Moen of the University of Minnesota show based on American Time Use Survey data. U.S. adults also watched more television, a sedentary and passive activity, they found. Women—but not men—spent much more time preparing food, and men without jobs or college degrees were less physically active than in 2019, which may reflect the mental health impact of unemployment.