Research from biology and psychology has shown that the prenatal period is sensitive to the environment and critical for later development. While the effects of toxins such as alcohol and nicotine on the fetus are well documented, the effect of maternal stress is more difficult to assess. The main reason is unobserved selectivity. Women who experience or report high levels of stress may be different from those who don’t in ways that affect their pregnancies, making it impossible to disentangle the effect of stress from its common correlates. The question is important because stress is widespread, stratified along socioeconomic and racial lines, and may be a central mechanism for the noxious effect of poverty or discrimination on children. We examine the effect of maternal stress and address the unobserved selectivity problem in a recent ASR article.
In order to address the unobserved selectivity problem, my coauthor Andres Villarreal and I examined the effect of month-to-month change in local violence in Mexico on birth outcomes over the 2008-2010 period, when drug trafficking-related violence skyrocketed across Mexican communities. Because such short-term variation in homicides from one month to the next is likely uncorrelated with unobserved local characteristics shaping birth outcomes and likely has important repercussions for maternal stress and women’s behavioral and reproductive choices, sudden changes in homicides provide a good instrument for stress.
The analysis yielded a very surprising finding: Increase in local violence early in the pregnancy results in an increase in birthweight and a decline in the probability of low birthweight. This finding contradicts the very reasonable expectation that stress is noxious for the developing fetus. What can account for it? The answer, our research suggests, is the behavioral responses by pregnant women who face a stressful environment. One possibility is that, in the face of violence, some women leave the area or postpone their fertility. We don’t find evidence of either process. Rather, we find that pregnant women who experience an increase in violence tend to increase their use of prenatal care. Compared to those not affected by violence, they are more likely to use prenatal care, doing so earlier during the pregnancy, and use prenatal care more often. This finding suggests that women, driven by altruistic fear for their pregnancies, use resources at their disposal to protect themselves when they face a stressful environment.
This effect is not homogeneous across women. Rather, it is stronger among lower-class and urban women, the same group of women who see their birth outcomes improved as a result of growing local violence. For example, the probability that poor women living in urban areas start using prenatal care before the third trimester of gestation increases by 15% on average as a result of increasing violence, and the birthweight of the children they give birth to increases by 180 grams, on average (Figure 1). The explanation for this response, we believe, is that poorer women are more vulnerable to violence than their wealthier counterparts. However, in order to respond with an increase in prenatal care use, they need access to the healthcare system, which is much easier to do in urban areas.
Figure 1. Effect of local violence on birthweight and on initiation of prenatal care before 3rd trimester. All women, by social class, and by urban residence.
How can we account for these diverging behavioral responses to environmental violence? Under which circumstances is maternal stress likely to induce health-enhancing versus health-depressing responses? Psychological theory offers an answer by proposing an inverted U-shaped relationship between perceived risk and performance—the Yerkes-Dodson hypothesis (Yerkes and Dodson 1908). According to this hypothesis, low levels of perceived risk do not provide sufficient motivation to act effectively, and very high levels trigger defensive mechanisms targeted at reducing anxiety rather than addressing the threat. It is moderate levels of perceived risk -possibly experienced by Mexican women- that provide motivation to act without overwhelming coping mechanisms.
This analysis contributes to a nascent literature using natural experiments to capture the causal effect of prenatal exposure to environmental stressors. It shows that behavioral responses by pregnant women are crucial for understanding the effect of violence on wellbeing. Depending on the context, these behavioral responses may exacerbate the negative effect of violence (if, for example, mothers resort to alcohol or cigarette consumption as a result of stress) or, as in this case, may compensate for the negative effect of violence.
Torche, F. and A. Villarreal. 2014. “Prenatal Exposure to Violence and Birth Weight in Mexico: Selectivity, Exposure, and Behavioral Responses” American Sociological Review 79(5): 966-92.
Yerkes, R. and J. Dodson. 1908. “The Relation of Strength of Stimulus to Rapidity of Habit-Formation.” Journal of Comparative Neurology and Psychology 18(5):459–82.