My Research: The Unconventional Story of Unintended Pregnancy

I am a public heath researcher who graduated last year with a PhD in Health Behavior and Health Education. I study social determinants of health disparities, specifically birth outcomes, reproductive health, and chronic disease. I want to share my work because I’m passionate about public health and social stratification. I want my ideas accessible to people who are interested in social science ideas. In the effort to practice writing and share my research, I’ve distilled one study from my dissertation into this post. I investigated the impact of educational advantages on unintended pregnancy in the United States and found that women who grow up with many educational advantages had an increased, not decreased, likelihood of unintended pregnancy relative to their more disadvantaged peers.

By definition, unintended pregnancy is divided into two sub-categories: “mistimed” (a pregnancy wanted at a later time) and “unwanted” (a pregnancy not wanted at all). These definitions, and the associated measurement data, are typically found in demographic and public health surveys. Examples include the National Survey of Family Growth and the Pregnancy Risk Assessment Monitoring System. In general, the majority of unintended pregnancies in the United States are mistimed — 31% of all pregnancies v. 19% unwanted (while 50% are intended).

Seemingly, many believe unintended pregnancy occurs among socially disadvantaged groups of women, like uneducated women; who are presumed likely to have inadequate access to quality health care. Because of this presumption, many so-called solutions to unintended pregnancy are centered on correct contraceptive skill, knowledge, attitude, or access. However, pregnancy intentions are complex and determined by contextual and background factors such as the social norms around sex, educational and work opportunity, and attitudes about becoming a mother. Counter intuitively, when we examine the impact of educational advantages on mistimed and unwanted pregnancies, we find that women with high educational advantage as youths were more likely (v. middle or low educational advantage) to report a mistimed pregnancy.

We conceptualize “educational advantage” as the environmental conditions that increase the likelihood of achieving academic success. This is a richer measure than education level because it captures things like college expectations, college-educated parents, and a college-preparatory curriculum. Growing up with educational advantage appears to create expectations in girls (e.g. college, career) that create conscious awareness around delaying pregnancy until later in life.

Table 1. presents descriptive information on the sample of women used for this analysis, stratified by level of educational advantage, National Longitudinal Survey of Youth 1979 (more info here)

Table 1 tells us that, compared to women who grew up with high educational advantage, low educational advantage is correlated to:

· Lower educational achievement

· Marital status (unmarried)

· Economic environment (impoverished childhood)

· Age at first pregnancy (younger)

· Race (Black/African-American)

· Percentage of wanted pregnancy (lower)

Yet, these indicators tell us very little about why unintended pregnancy is concentrated among the most disadvantaged women. Surprisingly, very few studies have actually tested hypotheses using quantitative methods. Many assume that it is due to inadequate contraceptive behavior on the part of women. However, studies based in interview data indicate that other factors, like social norms, relationship dynamics, and attitudes, may be at play and matter more than more proximate factors like contraceptive behaviors. But, perhaps, the same mechanisms that sustain adolescent achievement (which increase adulthood success), also encourage delayed pregnancy; so that girls with more advantages are more likely to find a “well-timed” pregnancy more salient.

Educationally advantaged girls and young women, who look forward to higher education and more valued occupations or identities to which they expect access, may not view motherhood as a primary social role when faced with roles and identities considered to be higher or immediate priority. Parents, educators, and other adults who provide many educational resources, opportunities, and support in home and school settings are likely also conferring social norms for success that also require delayed childbearing. Girls who experience more educational support learn to prioritize schooling and economic success, over motherhood, through vigilant use of contraception and limited sexual contact to prevent pregnancy.

In contrast, educationally disadvantaged students have fewer ways to control their academic trajectory, diminishing their expectations of adult opportunities, and simultaneously reducing the opportunity costs and the salience of the construct of having a “poorly-timed” pregnancy. Thus, mothers who grew up in less educationally advantaged circumstances are less likely to consider their births “mistimed,” compared to more educationally advantaged peers. These factors often precede more proximate contraceptive behaviors and can serve as a sort of motherboard for a variety of fertility and non-contraceptive behaviors.

So, once early educational opportunity and expectations are accounted for, we find that having more advantages actually increases the likelihood of unintended, specifically mistimed, pregnancy. You can this in Table 2 where I’ve limited the outcome to mistimed pregnancies for clarity:

Table 2. This table presents three sets of relative risk ratios of having a mistimed pregnancy, compared to a wanted one, from three separate models (all control for the previously mentioned indicators in Table 1). Relative risk ratios tell us the “risk,” or likelihood, of an event occurring, compared to a referent event. In this case, the event we are interested in is mistimed pregnancy compared to a wanted pregnancy. The asterisks tell indicate the level of statistical significance, for those of you familiar with p-values (*<.05, **<.01, ***<.001).

This set of relative risk ratios tell us a few things. The first model (M1) shows the risk of having a mistimed pregnancy by level of educational advantage. Women with median advantage were 45% more likely to have a mistimed birth compared to low advantage and women with high advantage were 71% more likely to have a mistimed birth.

The second model (M2) shows the risk of having a mistimed pregnancy by level of educational advantage and education by first birth, which is measured as years of schooling. The bottom line is the more years of schooling a woman acquired by the time she had her first birth, the more likely that birth was mistimed. You may notice that the statistical significance of educational advantage diminishes a bit when we include education by first birth in the model. That is because education by first birth is strong linked to both educational advantages and pregnancy intention, which makes some intuitive sense and is shown in Table 2.

The third model (M3) shows the risk of having a mistimed pregnancy by the level of educational advantage and years of schooling by age 25, which is a more conventional measure of education. In this model, years of education by age 25 is not statistically significant, but educational advantage is: women with median advantage are 42% more likely to have a mistimed pregnancy and women with a high advantage is 58% more likely.

Figure 1 is a graphical presentation of the predicted probabilities of a birth from a mistimed pregnancy over age and stratified by level of education advantage. Overall, the predicted probability of having a mistimed birth declines as age increases and begins to increase at later ages. Second, low advantaged women consistently have the lowest probability of a (birth from a) mistimed pregnancy out of the three groups.

Summary: While women with medium to high educational advantage tend to have more years of schooling, married, be older at first birth, be White, and come from a better economic environment; yet, once we account for these indicators, they are more likely to describe their first pregnancy as “mistimed” versus lower educationally advantaged women.

These findings complicate the pervasive narrative that unintended pregnancy is an objective characterization for public health and demographic research and suggest that it is a subjective construct that depends on the saliency of pregnancy timing. It also suggests that focusing on increasing contraception cannot be expected to reduce unintended pregnancy.

Thank you for reading my first post in this series. I invite you to comment below.

My next topic will describe how abortion fits into this discussion. I’ve done previous analysis that explore the likelihood and found that young with indicators of advantage were more likely to have abortion.