You could say anthropologist Mary Shenk traveled to Bangladesh to complete a scholarly version of comparison shopping when she and a group of researchers left for the Matlab region of Bangladesh just over three years ago.
The MU assistant professor and her colleagues made the trip to see how that nation’s experience of what social scientists call “demographic transition” — a phenomenon typically defined as the process by which countries move from high birth and death rates to low birth and death rates during industrial development — stacked up against various theories purporting to explain it.
Her findings, published earlier this year in the Proceedings of the National Academy of Sciences, will help scholars, public officials and development workers better understand one of the world’s most pressing problems: too many babies in areas with too few resources.
Shenk describes herself as an evolutionary social scientist. “A lot of the theory I’m interested in comes from evolutionary theory, specifically, behavioral ecology.”
Behavioral ecologists, she says, study behavior from an evolutionary perspective, focusing on why behavior varies in different environments, including different cultures. Scholars have been considering for decades how demographic transitions happen from a multitude of theoretical perspectives, Shenk says. Fertility is measured by the number of offspring women have, and why it’s on the decline has been the subject of hundreds of scholarly articles.
“Everyone has a sense of some of the factors that are involved, but there’s been, comparatively speaking, very little research comparing across the different models that exist,” Shenk says.
Shenk was primarily interested in studying how these transitions take place in the lives of real people, particularly the women who would bear, or chose not to bear, the next generation of Bangladeshis. Fortunately for her, the International Centre for Diarrhoeal Disease Research, Bangladesh, ICDDR,B for short, has been collecting demographic data in the Matlab area since the late 1960s.
Using the Dhaka, Bangladesh-based international health research organization’s census data, Shenk’s team was able to draw a random sample of women between the ages of 20 and 64. From there, research assistants were deployed to the women’s homes.
“We received really good cooperation, probably because people are used to researchers asking questions in this area and because we are affiliated with ICDDR,B which has a long-standing relationship with people in the local villages,” Shenk says. “There really is not a culture of suspicion of researchers, and most people are happy to give an interview and very interested to speak with me, as I am a foreigner, and many people have not seen or talked to a foreigner before.” Even with all of the research and evidence available, the assistant professor found a dearth of comparative models that might help unify all of the theoretical models floating around. The idea to actually compare the models that help explain fertility decline came to her in 2006. At the time, Shenk was busy beginning a post-doctorate term at the University of North Carolina at Chapel Hill.
“I had just finished my dissertation at the time, and was publishing articles on that data,” Shenk says. “I also knew this was the kind of work I wouldn’t be able to do alone for a variety of reasons.”
Eventually, Shenk recruited a team of researchers and academics to work with her: a methods expert, Mary Towner of Oklahoma State University, Howard Kress, who had done work similar to her own in Ecuador, and Nurul Alam, a demographer in Bangladesh. Then she found a suitable field site in Matlab, Bangladesh, an area that was attracting scholarly interest thanks to a recent demographic transition. Despite being studied extensively, Shenk says, few of her colleagues had sought to apply different theoretical models to the situation in Matlab and compare them. Shenk and her colleagues contacted the National Science Foundation with a proposal to make it happen. The NSF gave a green light to the project in 2009.
Work began in early 2010. Shenk and Kress packed their bags and headed to Bangladesh to collect new data that the group could combine with census data already collected. The study was designed to be a mix of the quantitative — statistical analysis — and the qualitative, personal interviews meant to help the researchers understand cultural contexts that informed child-bearing decisions.
Not surprisingly, some of these questions involved asking participants to comment on their lives’ most intimate details. Nevertheless, Shenk says, most were unfazed.
“The types of questions we were asking might seem personal in a Western context, but they are not nearly as personal in the Bangladeshi context,” Shenk says. “There, it is polite to ask about people’s children, relatives and health. Also, these kinds of questions were asked by ICDDR,B, so people are used to them.”
In fact, she says, very few women refused to participate in the interview process. “A couple of women did not give interviews because they were ill, and one woman didn’t do the interview because she was getting married the next day and was very busy. But mostly people said yes. … Male relatives of our participants generally didn’t have any problems with the women answering questions because our interviewers were all women and we were asking mostly standard demographic research questions.”
Shenk knew going in that she wanted to focus on the three demographic transition models dominating the field: risk and mortality, cultural transmission, and economics and investment. Shenk says each model has its own advantages when it comes to explaining what drives fertility decline.
The risk and mortality model, Shenk explains, postulates that fertility — the number of children born — goes down as mortality rates decline. The converse is also true. “When you have really high infant mortality rate or child mortality rates, people have to have a lot of children because they don’t know how many of them are going to die,” Shenk says. “So even if they really only wanted three children, they have to have, say, six because they don’t know what’s going to happen.”
Thus necessity demands large families as a hedge against untimely deaths. “There are examples like this even in our data set among the older women. There’s a woman who had 12 children and eight of them died,” she says. “That’s not common, but it happens.”
Cultural transmission models, on the other hand, attempt to provide insight into “how the preferences, beliefs, and social norms that govern human behavior” are passed along from generation to generation, and how these practices yield both continuity and change.
It’s a set of models, Shenk says, that has been part of the published literature “for a long time, in different approaches and demography.” It tends to focus on who the parents talk to, what is being told to them and how much influence that has over time in changing their behavior.
Finally, Shenk looked at economic and investment models. The emphasis here, especially for a rapidly developing nation like Bangladesh, is on the fallout from economic change, typically the transition from predominantly agricultural to market economies. Here the main question is, “How much does it cost to invest in children?”
Using a model comparison approach, Shenk’s team used the field research findings from Matlab to see how each model predicted a similar set of fertility-related variables, says Oklahoma State’s Mary Towner, the woman in charge of crunching the numbers. “We’re particularly interested in looking at fertility and at how that connects to other variables, like access to resources, education, mortality risk, and occupations,” she says.
In social sciences, according to Towner, statistical approaches have been evolving for a while. The changes are due, she says, to both technological advances and a trend in academia toward comparing alternatives rather than rejecting individual hypotheses one at a time. Towner has been using technology to hone the latter approach for about a decade. She says she was excited to apply the comparison approach to a study such as the one Shenk proposed.
“We can take on a bigger theoretical framework and compare alternative models because we have the methods now to do it,” Towner says. “The number crunching is much more straightforward now than it would have been even 10 years ago because of bigger, faster computers. The advances in alternative statistical approaches, as well as having the computer ability to do it, are finally becoming more accessible.”
One of the biggest advantages, Towner says, is that it allows scholars and researchers to challenge the conventional theoretical thinking while more quickly developing alternative ideas or models.
For their part, Shenk, Towner, Kress, and Alam found that the economic and investment models did play the most important role in predicting Bangladeshi fertility trends. But they also discovered that other variables under consideration — some having nothing to do with money — also played an important part in family planning.
“We found a number of variables were important, especially in the economic model, but some other variables had a bigger effect than I imagined they would have had going into the study, which is neat,” Towner says. Some of the variables affecting fertility rates include women’s education levels, their family involvement in agriculture, local infant mortality rates, or whether there is good access to birth control and information on family planning.
According to Shenk, declining fertility is a global phenomenon that began in England and France in the 18th century and has been working its way around the globe ever since. In the United States, fertility began to drop during the rise of industrialism in the 19th century and continued falling through the Great Depression. Then American birth rates famously bounced back during the post-World War II period. “Again, [the decline] suggested a strong link to economics, and then you have this bounce back in this really strong age of economic expansion, post World War II in the U.S., the baby boom,” Shenk says.
All this might seem familiar to folks in Bangladesh. At a time when almost every nation on the planet is feeling the effects of global recession, Bangladesh is one of the few countries that has experienced growth over the last five years.
“Bangladesh in the ’80s had extremely high fertility, and programs were put in place to reduce infant mortality in ’80s and early ’90s,” says study coauthor Kress. Fertility declined in the 1990s and then “stalled out,” Kress says. Further reductions are occurring now that Bangladesh is entering into a period of economic growth — growth fueled chiefly by a surge in Chinese companies moving textile plants into the country to take advantage of its low wage labor force.
“When we look at the demographic transition work, we’re really looking at what it means to raise a family, what is the proper way to raise a family, and that is really sensitive to economic and risk factors in any given nation or local economy,” Kress says. “When we’re looking at developing public health programs or economic intervention programs, we should always be aware of the demographic consequences of introducing those programs.”
Adds Shenk: “Our study created a framework by which different explanations could be explicitly compared. Population data from any region could be analyzed using these methods to help researchers, government officials, health workers and others understand the key drivers of demographic change.