Are Hispanic Women Happier About Unintended Births?

Abstract

Reducing unintended pregnancies – particularly among Hispanic and Black women, who have relatively high rates – is a key public health goal in the United States. However, descriptive literature has suggested that Hispanic women are happier about these pregnancies compared with White and Black women, which could mean that there is variation across groups in the consequences of the resulting births. The purpose of this study was to examine variations in happiness about unintended births by race-ethnicity and to assess possible explanations for these differences. Using data from the National Survey of Family Growth (n=1,462 births) I find that Hispanic women report being happier about unintended births compared with White and Black women. Higher happiness among Hispanics was particularly pronounced among a subgroup of women: those who were foreign-born and very religious. Overall, results confirm previous findings that intention status alone is incomplete for capturing pregnancy experiences. Happiness offers complementary information that is important when making comparisons by race-ethnicity and nativity.

Introduction

Unintended pregnancies and births are a major public health concern in the United States. According to 2006 figures, one-third of births are the result of unintended pregnancies (Finer and Zolna 2011). Unintended pregnancies and births are even more common among Hispanic and Black women in the United States and the reduction of these racial-ethnic disparities is one of the goals outlined in Healthy People 2020 (United States Department of Health and Human Services 2010). However, cross-tabulations from the 2002 National Survey of Family Growth (NSFG) suggest that there are racial-ethnic differences in women's happiness about unintended pregnancies (Chandra et al. 2005). If this is the case, it might indicate that pregnancy happiness is a useful complementary measure to traditional measures of pregnancy intentions, particularly when making racial-ethnic comparisons. The present study describes racial-ethnic differences in women's happiness about unintended pregnancies ending in births and seeks to explain these differences.

Background

Data on unintended pregnancy have been collected as part of the NSFG since 1973. The measure has evolved over time but the most recent version asks women to think back to the time before a pregnancy was conceived and answer whether she wanted a baby at some time in the future. If she did want a baby at some time, she is asked whether this particular pregnancy came on time, later than she wanted, or earlier than she wanted. Pregnancies that were not desired at the time or any time in the future are typically classified as “unwanted,” while those that occurred too early are classified as “mistimed.” “Unintended” pregnancies are defined as those that are either unwanted or mistimed (Chandra et al. 2005; National Center for Health Statistics 2010). Unintended pregnancies that are carried to term are commonly referred to as “unintended births” (see England, McClintock, and Shafer 2011; Guzman et al. 2010).

Although this measure of unintended pregnancy is widely used, a substantial body of literature indicates that unintended pregnancy is a heterogeneous concept and that this single measure is inadequate for understanding wantedness. One set of critiques argues that pregnancy intentions are multi-dimensional or exist on a continuum (Bachrach and Newcomer 1999; Lifflander, Gaydos, and Hogue 2007; Miller 1994; Stanford et al. 2000; Zabin 1999). Others question whether most women even try to plan their pregnancies. Indeed, focus group research has shown that “planning” a pregnancy is not a meaningful concept to many women (Moos et al. 1997; Trussell, Vaughan, and Stanford 1999). Even women who do hope to plan their pregnancies may not be sure of their lifetime childbearing preferences, which makes it difficult to categorize a particular pregnancy as intended, mistimed, or unwanted (Kaufmann, Morris, and Spitz 1997). These concerns are validated by the fact that a high proportion of reproductive-aged women – 23% in a national telephone survey – say they “don't care either way” if they get pregnant (McQuillan, Greil, and Shreffler 2010). There may, in fact, be good reasons for not deliberately planning pregnancies. Luker (1999) points out that deciding to have a baby is an onerous decision and that becoming pregnant “accidentally” allows women and men to avoid having to make that decision. Also, planning a pregnancy invites the possibility of disappointment and stress if a pregnancy does not occur (Moos et al. 1997). In addition, the consequences of a pregnancy might be shaped by a number of other people – partners, family members, and peers – who may be more or less supportive (Forte and Judd 1998; Kendall et al. 2005; Moos et al. 1997; Petersen and Moos 1997; Santelli et al. 2003; Santelli et al. 2009). The fact that all these people influence the likelihood that a woman gets pregnant, and the way she feels about the pregnancy once it occurs, complicates the task of evaluating whether she wanted it in the first place.

One relevant dimension of pregnancy wantedness is the woman's reported happiness (or unhappiness) about the pregnancy. Prior research has found that there is variation in post-conception feelings of happiness and that happiness predicts key outcomes. In several studies, factor analysis and other methods have found that happiness about a pregnancy is a distinct and important dimension of pregnancy desirability (Fischer et al. 1999; Miller 1994; Santelli et al. 2009; Stanford and Dewitt 1998; Stanford et al. 2000). This is echoed in qualitative research which finds that it is common for a woman to adapt easily to an unintended pregnancy and come to see it as a positive occurrence (Moos et al. 1997; Stanford et al. 2000). For some women, there are even more advantages than disadvantages associated with such a pregnancy, including a push to “grow up” and “find direction” (Edin and Kefalas 2005; Moos et al. 1997). Reported happiness about a pregnancy is predictive of pregnancy outcome (i.e., birth or abortion) and provides predictive power above and beyond the traditional intendedness measure (Fischer et al. 1999; Stanford et al. 2000; Santelli et al. 2009). Happiness about a pregnancy is positively correlated with prenatal care and birth weight, and in some cases it is more predictive of these outcomes than was pregnancy wantedness (Pulley et al. 2002; Sable et al. 1997; Sable and Wilkinson 1998).

Pregnancy intention might be a weaker or less meaningful measure for Hispanic women or women with low socioeconomic status generally, so it may be particularly important to explore other dimensions of pregnancy wantedness among this group of women. Research has shown that disadvantaged women might be less inclined to plan pregnancies. Qualitative research finds that poor, young, unmarried women may want a baby but believe it would be irresponsible to deliberately try to have a baby under their current circumstances, so they simply let it happen (Edin and Kefalas 2005). In addition, prior research suggests that compared with other women, Hispanics might be more welcoming of pregnancies even under challenging circumstances. For example, compared with Blacks and Whites, Hispanic pre-teen and teenage girls report the youngest desired age for first birth (East 1998). Moreover, in a clinic-based sample, Hispanic girls were more likely than other teenagers to report being ambivalent about a potential future pregnancy (Crowley and Farley 2001). Cross-tabulations from the 2002 NSFG suggest that Hispanic women are less unhappy about unintended pregnancies than other groups of women (Chandra et al. 2005). Other research has shown differences by nativity. Among women of Mexican origin, those born in Mexico report being happier about pregnancies than those born in the United States (Wilson 2008).

In order to better understand variation within the category of unintended births, this study examines whether feelings of happiness about unintended births differ across racial-ethnic groups and evaluates several hypotheses for why such differences might exist. If racial-ethnic differences in happiness cannot be easily explained, we might conclude that pregnancy happiness is an important complement to the traditional measure of pregnancy intentions when making comparisons across racial-ethnic groups.

Perhaps the most obvious explanation for why Hispanic women are happier about unintended births is that these women might be happier in general or happier about all births. Table 1 presents mean happiness scores for intended pregnancies by race-ethnicity and nativity using the 2006-10 National Survey of Family Growth. These results confirm that there is no statistically significant difference in happiness about intended births between Whites and Hispanics. The mean happiness score for intended pregnancies ending in births is 9.4 for non-Hispanic US-born White women, compared with 9.6 for foreign-born Hispanic women and 9.2 for US-born Hispanic women.

Table 1.

Happiness about intended (wanted) pregnancies ending in births (weighted)

	White US-born	Hispanic foreign-born	Hispanic US-born	Black US-born
Happiness about pregnancy, 1=very unhappy, 10=very happy (mean)	9.4	9.6	9.2	8.7 *
N = 1,822 births	996	320	203	303

^*Difference with Whites statistically significant at p<0.05

A second question is how abortion winnows the pool of unintended pregnancies that become births and how abortion rates might differ by race-ethnicity. Prior research has shown that the percentage of unintended pregnancies that end in abortion is nearly identical for White and Hispanic women: just under 40% in 2006 (Finer and Zolna 2011). This suggests that higher happiness about unintended births among Hispanic women compared with Whites is probably not due to variations in abortion.

The present study considers four possible explanations for racial-ethnic differences in happiness about unintended births. First the study assesses whether higher Hispanic happiness about unintended births is due to pre-conception feelings or post-conception feelings. In other words, are Hispanic women happier because these pregnancies were less unwanted before they occurred, or are they happier because they are more likely to warm to the pregnancy post-conception? If it is the case that Hispanic women were more open to these pregnancies before conception occurred, then the happiness gap will be explained by the fact that Hispanic women report having tried less hard to avoid these pregnancies and report having “wanted” the pregnancies more before conception.

After examining the extent to which happiness is accounted for by pre-conception feelings, I test three hypotheses about why Hispanic women might be happier about unintended births:

Hypothesis 1

Hispanic women may be happier about unintended births than White and Black women because they have lower socioeconomic status (Kochhar 2007) and therefore have lower opportunity costs associated with having a birth. According to this perspective, women with lower earning potential have a weaker incentive to limit childbearing because they lose less in earnings when they shift time from market work to childcare work (Becker 1991; Pollak and Watkins 1993). This theory has been extended to the differential incentives to prevent nonmarital and unintended pregnancies specifically (Hotz, Klerman, and Willis 1997; Miller 1986; Montgomery 1996; Willis 1999). Blake and colleagues (2007) find that lower levels of education are associated with higher levels of happiness about pregnancies generally. Based on their interviews, Edin and Kefalas (2005) observe that poor young women generally see few costs associated with having a baby, and having a first baby early will not do much to alter their educational or financial futures. Other research (also in the United States) confirms that in low-income communities, teenage mothers generally do not suffer large economic or educational penalties due to having early births (Furstenberg, Brooks-Gunn, and Morgan 1987; Kearney and Levine 2012).

However, not all research finds that the relationship between socioeconomic status and fertility is driven by differences in opportunity costs. Musick and colleagues (2010) for example, find that measures of earnings do not explain the association between education and fertility. Moreover, it is possible that higher socioeconomic status might be associated with less unhappiness about unintended births, rather than more, given that resources act as a buffer against many of the potential stresses of childrearing.

Hypothesis 2

Hispanic women may be happier about unintended births than White and Black women because they have more social support. Specifically, I examine support from partners, including marriage or cohabitation, whether the partnership provides a good environment for raising a child, and the partner's enthusiasm for the pregnancy.

Women who are married may be less unhappy about unintended births compared with unmarried women, since marriage is the normative arrangement for childbearing and since stable relationships enable mothers to share childrearing responsibilities and costs (Willis 1999). Similarly, cohabitation is likely to be associated with higher pregnancy happiness compared with not being in a coresidential union. Blake and colleagues (2007) find that married women are happier about their pregnancies than are unmarried women. If Hispanic women are more likely to be in stable unions at the time of an unintended pregnancy, this could help explain differences in happiness. Prior research shows that Hispanics (particularly those of Mexican origin) marry earlier than Whites, but are less likely to be married at older ages (Oropesa and Landale 2004).

Whether the woman has support for the pregnancy from her partner is also likely to influence her happiness. Interviews with women who experienced an unintended pregnancy revealed that positive responses from partners were often enough to make the women happy about the pregnancies (Stanford et al. 2000; see also Kerns et al. 2003). Blake and colleagues (2007) find that having a partner who considers the pregnancy unintended is associated with being less happy about the pregnancy. Hispanic women (particularly those who are foreign-born) might be especially likely to have a partner who supports the pregnancy, for several reasons. Hispanic men who are foreign-born might be particularly enthusiastic about a partner's pregnancy because in Latin American culture masculinity was traditionally tied to having children, though this is weaker than it once was (Sable et al. 2006). Research from focus groups with foreign-born Hispanic women also suggests that in some couples, males have full control of contraceptive decisions; such an arrangement would probably increase the chance of disagreeing on pregnancy intentions, such that the woman does not consider the pregnancy intended but the man does (Sable et al. 2009). Moreover, in communities with low socioeconomic status, men may have a desire to father children even under less-than-ideal circumstances since they often enjoy some of the benefits of fatherhood while bearing relatively few costs (Montgomery 1996; Willis and Hagaa 1996; Willis 1999).

Hypothesis 3

Hispanic women might be happier about unintended births compared to other women due to religious differences. The association between religious involvement and pregnancy feelings could operate through several avenues. First, religious beliefs may help women cope with the unexpected event of a pregnancy. Prior research has shown that those who are disadvantaged often use these beliefs to cope with a lack of control over their lives (Chatters, Levin, and Taylor 1992; Krause 1992). This extends to pregnancy – focus group research reveals that women often use their religious beliefs to rationalize and accept unintended pregnancies (Moos et al. 1997). Through a religious lens, unintended pregnancies are often viewed as “meant to be.” According to interviews, many Mexican immigrant women view their fertility as being in God's hands (Hirsch 2003). Religiosity is also associated with more “family-oriented” values and sex role segregation (Goldscheider and Goldscheider 1988; Thornton 1985; Thornton and Camburn 1989). Therefore, religious beliefs may provide support and encouragement for women to adopt roles as mothers, even under challenging circumstances. Catholic women in particular might be less likely to view unintended pregnancies and births as negative events due to the anti-contraceptive stance of Catholic Church (Vatican 1968). Finally, involvement in a religious congregation might be a source of social support (Neff and Hoppe 1993). Prior research has established racial-ethnic differences in religious participation, importance of religion, and denomination (Neff and Hoppe 1993; Rew and Wong 2006; Smith et al. 2002).

There might also be an interaction whereby race-ethnicity moderates the association between religiosity and pregnancy happiness. The relationship between religiosity and premarital childbearing differs for young Hispanic women compared with young White women, with religiosity being slightly less protective for Hispanics compared with Whites (Abrahamse, Morrison, and Waite 1988). Other authors have also found the relationship between religious beliefs and coping differs for Blacks compared with Whites (Chatters et al. 2008; Rosen 1982).

To the extent that Hispanics in the United States do have different attitudes towards unintended births compared with other groups, it is expected that these differences will be most pronounced among women born outside the United States. When immigrants move, they are introduced to a new set of cultural values and economic constraints, according to the classic assimilation perspective (Bean and Swicegood 1985). Immigrants who are socialized in contexts where different norms prevail carry these norms to the United States, but with each successive generation, norms and behaviors converge with those of the receiving community – either the mostly White “mainstream” (under “classic” assimilation theory) or relatively poor local communities (under “segmented” assimilation theory; Alba and Nee 2003; Bean, Swicegood, and Berg 2000; Portes and Zhou 1993; Portes 2007; Rumbaut 1994). In addition, the socioeconomic status of immigrants rises with duration of residence and across generations, helping to drive convergence in attitudes and behaviors. I hypothesize that US-born Hispanic women will be more similar to US-born White and Black women in their reported happiness about unintended pregnancies ending in births, compared with foreign-born Hispanic women.

Data and Methods

The data for this analysis were pooled across two cycles of the National Survey of Family Growth: 2002 and 2006-08. The NSFG is a large-scale, nationally representative study of women in the United States with a cross-sectional design. Information on pregnancy history, relationship history, and other topics is collected at each wave, and the survey is considered to be one of the best sources of information on childbearing behavior in the United States. It is also the most widely used data source on the topic of unintended pregnancy, although the scope of investigable research on this topic is constrained by the cross-sectional design and the fact that abortions are underreported (Jones and Kost 2007). The analyses in this paper were limited to those unintended pregnancies that ended in births in order to be confident that results are not biased by differential reporting of abortion by race-ethnicity.

Sample

Data on pregnancy happiness were only collected for pregnancies taking place within 3 years prior to the interview so the analysis is limited to recent pregnancies. Unintended pregnancies were identified using two questions – whether the respondent wanted a pregnancy at any time in the future, and if so, did the pregnancy come sooner, at the right time, or later than she wanted it. A pregnancy was considered unintended if it came when the woman wanted no future pregnancy, if the pregnancy came sooner than she wanted it, or if she did not know how she felt about the timing of the pregnancy. The dataset included 1,607 unintended pregnancies that took place within 3 years of the interview and ended in births. Births to women who were not Hispanic, US-born White, or US-born Black were excluded from the analysis (n=126), leaving 1,481 pregnancies. One case was dropped due to missing data on pregnancy happiness and 18 cases (1%) were dropped due to missing data on at least one of the independent variables. The final analytic sample included 1,462 unintended births. Stata 10 “svy” commands were used to account for the NSFG's complex sampling design. Since some women contributed more than one birth, the standard errors were adjusted for clustering within mothers.

The self-identified race, ethnicity, and nativity of respondents were summarized using four categories: non-Hispanic US-born Whites (hereafter “Whites”), US-born Hispanics of any race, foreign-born Hispanics of any race, and non-Hispanic US-born Blacks (hereafter “Blacks”). The unit of analysis is births. Of the 1,462 births, 573 were to White women, 253 were to US-born Hispanic women, 210 were to foreign-born Hispanic women, and 426 were to Black women. These births were drawn from 1,297 different women. A minority of Hispanic respondents (42%) answered the computerized portion of the survey in Spanish.

Variables

The outcome variable was happiness about the pregnancy, based on the following question, “Please look at the scale on Card 39 [1=very unhappy, 10=very happy]. On this scale, a one means that you were very unhappy to be pregnant and a ten means that you were very happy to be pregnant. Tell me which number on the card best describes how you felt when you found out you were pregnant.”¹ For the analysis, the 10-point happiness scale was dichotomized at the mid-point, such that scores of 6-10 were coded 1 and scores of 1-5 were the reference category. Half of the sample fell into each group.

The set of independent variables that were tested in the models included two scale variables capturing women's feelings before or at the time of conception. Respondents were told, “Look at the scale on Card 40, where a zero means trying hard not to get pregnant, and a ten means trying hard to get pregnant. If you had to rate how much you were trying to get pregnant or avoid pregnancy right before you got pregnant (this time/that time), how would you rate yourself?” They were also told, “Look at the scale on Card 41, where a zero means you wanted to avoid a pregnancy and a ten means you wanted to get pregnant. If you had to rate how much you wanted or didn't want a pregnancy right before you got pregnant (this time/that time), how would you rate yourself?” Each of these scales is treated as a continuous variable.

There were three sets of variables capturing socioeconomic status and potential opportunity costs. Mother's education was a four-category variable for the highest level of education completed: less than high school (reference category), high school graduate or GED, some college, or Bachelor's degree or higher. Respondents were also asked, regarding each birth or current pregnancy, “At any time while you were pregnant with (baby's name/this baby), were you employed at a job for pay?” (0= no, 1=yes). There was also a dummy variable for whether the respondent had health insurance at the time of the birth. This variable was coded 1 if the respondent listed (non-Medicaid) health insurance as a form of payment for the birth (either the only form of payment or in conjunction with other forms of payment) and the reference category consisted of those who did not use insurance to cover any part of the birth.²

There were three sets of variables summarizing social support. First, union status had three categories: married at conception, unmarried at conception but married or cohabiting at the time of the birth, and unmarried at conception and neither married nor cohabiting at the time of the birth (reference category). Marital status at conception and union status at birth were calculated by NCHS using birth and relationship histories.

The second support variable was a dichotomous variable for whether the respondent wanted to have a baby with her partner. In reference to unintended pregnancies, respondents were asked, “Right before (the/this/that) pregnancy, did you think you might ever want to have a(nother) baby with that partner? Would you say definitely yes, probably yes, probably no, or definitely no?” Pregnancies were set equal to 1 if the respondent answered “definitely yes” or “probably yes,” (the reference category was “definitely no,” “probably no,” “refused,” or “don't know”).

The third support variable was the woman's perception of whether her partner considered the pregnancy intended. As with respondent intentions, partner intentions were based on two questions – whether the partner wanted a baby at any time in the future, and if so, did the pregnancy come sooner, at the right time, or later than he wanted it. These responses were coded 1 for intended pregnancies (those that came at the right time or later than he wanted) and the reference category consisted of unintended, “didn't care,” “indifferent,” “don't know,” or “not sure.”

Religiosity was captured with a dummy variable set equal to 1if the respondent reported having a religious affiliation and said that religion was “very important” in her daily life. The reference category consisted of those who considered religion “somewhat important” or “not important,” or were not affiliated with a religion.

Analytic Approach

There were two parts to the analysis. First, cross-tabulations with t-tests were performed to show how women's feelings about unintended births and the context of unintended births differed by racial-ethnic group.

Second, regressions were estimated to test hypotheses about why happiness about unintended births differed across racial-ethnic groups. Logistic regressions were used to predict whether respondents reported a high happiness score (in the top half of the scale) about a given pregnancy.³ To examine whether racial-ethnic differences in happiness about unintended pregnancies were mediated by other factors, I estimated a series of models in steps (non-cumulative). The steps included: 1) race-ethnicity only, 2) race-ethnicity plus pre-conception feelings, 3) race-ethnicity plus socioeconomic status and opportunity costs, 4) race-ethnicity plus social support, and 5) race-ethnicity plus religiosity (including interactions). I compared the coefficients for foreign-born Hispanics and US-born Hispanics in each model to evaluate whether any set of factors acted as mediators for the relationship between race-ethnicity and happiness. In all of these models I tested controls for time since the pregnancy, age at conception, and survey wave, but these coefficients were not statistically significant, so they were not included in the models presented.

Results

Table 2 presents descriptive statistics for the sample of unintended pregnancies ending in births. These findings were similar to the racial-ethnic differences in pregnancy happiness suggested by previous research (Chandra et al. 2005). Of the four groups, Hispanic immigrants were the happiest about unintended births, with an average of 7.2 on the 10-point scale. US-born Hispanics were also significantly happier than Whites, with an average of 6.2 on the 10-point scale. There was no statistically significant difference in happiness between Whites and Blacks; the average happiness scores for these groups were 5.6 and 5.3, respectively. The same racial-ethnic differences appear when examining the dichotomous happiness variable. A total of 70% of foreign-born Hispanic women and 61% of US-born Hispanic women reported happiness scores between 6 and 10, compared with 45% and 48% of White and Black women, respectively. A similar racial-ethnic pattern emerged for pre-conception feelings: Foreign-born Hispanic women reported trying less hard to avoid pregnancy compared with Whites (3.8 versus 2.6 on a 10-point scale) and wanting to avoid it less (3.8 versus 2.4 on a 10-point scale).

Table 2.

Descriptive statistics for the sample of unintended pregnancies ending in births (weighted)

	White US-born	Hispanic foreign-born	Hispanic US-born	Black US-born
Happiness about pregnancy 1=very unhappy, 10=very happy
Mean score	5.6	7.2 *	6.2 *	5.3
% reporting high happiness score (6-10)	47.8	70.1 *	61.2 *	44.6
How hard trying to avoid pregnancy (0) or to get pregnant (10) (mean score)	2.6	3.8 *	3.4	2.8
How much did you want to avoid (0) or to get pregnant (10) (mean score)	2.4	3.8 *	2.7	2.5
Mother's education
Less than high school	18.2	86.3 *	51.2 *	23.0
High school	38.3	3.9 *	23.9 *	40.2
Some college	28.5	6.6 *	20.5	22.4
Bachelor's plus	14.3	2.9 *	3.9 *	12.9
N/A (no mother figure)	0.8	0.4	0.5	1.5
Worked (at all) during pregnancy	60.9	40.5 *	42.2 *	62.3
Paid for birth with insurance	49.6	27.5 *	30.4 *	33.6 *
Union status at conception and birth
Married at conception	45.1	46.6	24.2 *	18.1 *
Unmarried at conception, married/cohabiting at birth	31.6	33.9	52.1 *	29.4
Unmarried at conception, not married/cohabiting at birth	23.3	19.5	23.7	52.5 *
Wanted to have a baby with this partner	68.8	66.6	67.8	56.7 *
Partner considered pregnancy intended	24.4	30.8	29.0	31.8
Religion very important in daily life	35.7	59.3 *	37.5	67.9 *
N = 1,462 births (Drawn from 1,297 women)	573 (508)	210 (196)	253 (221)	426 (372)

^*Difference with Whites statistically significant at p<0.05

In addition, the situations surrounding unintended births differed by mother's race-ethnicity. Compared with Whites, US-born Hispanics and Blacks were less likely to be married. Hispanic women (both foreign-born and US-born) were less likely to work during the pregnancy and less likely to pay for the birth with insurance. There were also differences in key demographic characteristics: Hispanic women (particularly foreign-born) had less educated mothers compared with Whites, and foreign-born Hispanic women (along with Black women) were more religious.

Table 3 presents coefficients and odds ratios from logistic regression models predicting happiness about unintended pregnancies ending in births. Consistent with the descriptive statistics, the baseline model in Table 3 reveals that Hispanic women were more likely to be happy about unintended pregnancies ending in births, compared with White women, and differences were particularly pronounced for Hispanic women who were foreign-born. The odds of happiness for foreign-born Hispanic women were 156% higher (2.56-1.00) than the odds for White women, while the odds for US-born Hispanic women were 72% higher than the odds for White women. Meanwhile, White and Black women had similar odds of being happy about these pregnancies.

Table 3.

Coefficients (and odds ratios) for logistic regressions predicting happiness about unintended pregnancies ending in births¹

	Baseline	Pre-conception feelings	Hypothesis 1: SES	Hypothesis 2: Support	Hypothesis 3: Religiosity
Race-ethnicity and nativity (Ref = White US-born)
Hispanic foreign-born	0.941 ** (2.563)	0.809 * (2.246)	0.882 ** (2.416)	0.981 ** (2.668)	0.462 (1.587)
Hispanic US-born	0.545 * (1.724)	0.566 * (1.762)	0.530 * (1.699)	0.530 * (1.699)	0.614 (1.848)
Black US-born	−0.129 (0.879)	−0.175 (0.839)	−0.118 (0.889)	0.022 (1.022)	−0.184 (0.832)
How hard trying to avoid pregnancy (0) or get pregnant (10)	--	0.005 (1.005)	--	--	--
How much did you want to avoid (0) or get pregnant (10)	--	0.294 ** (1.341)	--	--	--
Mother's education (Ref = Less than high school)
High school	--	--	−0.179 (0.836)	--	--
Some college	--	--	−0.080 (0.923)	--	--
Bachelor's plus	--	--	−0.204 (0.815)	--	--
N/A (no mother figure)	--	--	−0.334 (0.716)	--	--
Worked (at all) during pregnancy	--	--	0.108 (1.114)	--	--
Paid for birth with insurance	--	--	0.094 (1.099)	--	--
Union status at conception and birth (Ref = Unmarried at conception, not married/cohabiting at birth)
Unmarried at conception, married/cohabiting at birth	--	--	--	0.454 * (1.575)	--
Married at conception	--	--	--	0.276 (1.318)	--
Wanted to have a baby with this partner				0.857 ** (2.356)
Partner considered pregnancy intended	--	--	--	0.397 (1.488)	--
Religion very important in daily life	--	--	--	--	−0.029 (0.972)
Religion very important × Hispanic foreign-born	--	--	--	--	0.878 * (2.405)
Religion very important × Hispanic US-born	--	--	--	--	−0.181 (0.834)
Religion very important × Black US-born	--	--	--	--	0.094 (1.099)
N = 1,462 births

^**p<0.01

^*p<0.05.

¹The dependent variable equals1 for those who reported happiness scores of 6-10 (on a 1-10 scale). Odds ratios are in parentheses.

The subsequent models presented in Table 3 tested various hypotheses regarding racial-ethnic differences in happiness about unintended pregnancies ending in births. First, I examined whether racial-ethnic differences in happiness were due to differences in the degree of unintendedness – in other words, Hispanic women might have been happier about unintended pregnancies post-conception because they were more open to these pregnancies before conception occurred. Although the scale for wanting to get pregnant was positively and significantly associated with pregnancy happiness, controlling for these two variables (having wanted to get pregnant and having tried to get pregnant) did little to explain the difference in happiness between Hispanic and Whites. Thus, post-conception happiness was distinct from preconception feelings.

Hypothesis 1

The third model examines the explanatory factors of socioeconomic background and opportunity costs. Surprisingly, there was no significant association between pregnancy happiness and these variables, which included mother's education, whether the woman worked during the pregnancy, and whether the birth was paid for by insurance (a marker of socioeconomic status). There was also no evidence that racial-ethnic differences in happiness were due to differences in the socioeconomic background or the opportunity costs associated with unintended pregnancy; controlling for these variables did almost nothing to attenuate the racial-ethnic coefficients.

Hypothesis 2

The fourth model added variables capturing support from partners – including marriage and cohabitation status, the woman's feelings about having a baby with her partner, and the woman's perception of her partner's pregnancy intentions. Women who married their partners between conception and birth were more likely to report being happy about the pregnancy compared with those who remained unmarried (odds ratio of 1.6) and women who reported wanting to have a baby with their particular partner were more likely to report being happy than those who did not (odds ratio of 2.4). However, the data did not provide evidence that support from partners explained differences in happiness between Hispanics and Whites. Controlling for support from partners did not attenuate the odds ratios for race-ethnicity.

Hypothesis 3

The final model added controls for religiosity and the interaction between religiosity and race-ethnicity. These results show that religiosity was significantly associated with pregnancy happiness among foreign-born Hispanic women, but there was no significant association between religiosity and happiness among other groups of women. Foreign-born Hispanic women who were very religious were the most likely to be happy about unintended births. Based on the coefficients in this model, the predicted probability of reporting a happiness score in the top half of the scale was 0.77 for foreign-born Hispanic women who were very religious and 0.59 for foreign-born Hispanic women who were not very religious. There was no difference by religiosity among Whites: the predicted probability was 0.47 for those who were very religious and 0.48 for those who were not. In sum, higher happiness scores among Hispanic women were largely explained by a subgroup of women who reported particularly high scores: foreign-born Hispanic women who were very religious. ⁴

Discussion

Hispanic women in the United States – particularly those who were foreign-born – reported being happier about unintended pregnancies ending in births, compared with White and Black women. Less than half of White and Black women (48% and 45%, respectively) reported happiness scores in the top half of the scale, compared with 61% of US-born Hispanic women and 70% of foreign-born Hispanic women. Relatively high happiness scores among foreign-born Hispanics were largely attributable to a subgroup of women – those who were very religious. Differences in happiness about unintended pregnancies between Whites and Hispanics were not explained by differences in opportunity costs or differences in support from partners.

Notably, the data demonstrated that stronger feelings of happiness about unintended pregnancies among Hispanic immigrants were explained by an interaction with religiosity; women who were both very religious and foreign-born reported being much happier about unintended pregnancy than others. The religious communities that Hispanic women were integrated in might have provided particularly good support in the face of unintended pregnancies. Alternatively, religious beliefs might have functioned especially well as a coping mechanism for this group of women (Koenig, George, and Siegler 1988; Pargament 2001). The findings presented here fit with prior research demonstrating that race-ethnicity moderates the association between religious beliefs and various outcomes (Abrahamse et al. 1988; Chatters et al. 2008; Rosen 1982). The findings presented here suggest that immigration status can also influence the association between religiosity and outcomes.

An alternative explanation for why religiosity and nativity explained the White-Hispanic difference in happiness about unintended births is that religiosity is a proxy for acculturation among foreign-born Hispanic women. Under this scenario, higher happiness among foreign-born Hispanic women may not be a function of higher (and different) religiosity, but due to broader cultural orientation. A number of studies have highlighted Hispanic convergence with Whites in fertility behavior and other outcomes (Fischer and Mattson 2009; Parrado and Morgan 2008). The evidence presented here suggests that cultural differences might still be relevant to Hispanics’ outcomes in some areas of fertility behavior, and the way women view unintended pregnancies appears to be one of those areas.

Not only was socioeconomic status not useful for explaining the difference between Whites and Hispanics, but it was not predictive of pregnancy happiness, which runs counter to much of the literature on socioeconomic status and fertility. The opportunity costs model, bolstered by qualitative research findings, predicts that women with low socioeconomic status perceive that they have little to lose from a poorly-timed birth. The findings presented here therefore do not support the reasoning that disadvantaged women allow unintended pregnancies to happen due to low opportunity costs. These findings are in line with those of Musick et al. (2010) who found that variation in opportunity costs is not related to variation in births, including unintended births (Musick et al. 2010; see also Heckman and Walker 1990). Perhaps since women with higher opportunity costs are also better able to buy goods and services that make childrearing easier, these two effects balance one another out.

I also investigated whether post-conception happiness about unintended births among Hispanic women could be attributed to pre-conception desire for pregnancy (in other words, the possibility that these pregnancies were less unwanted to begin with). Pre-conception attitudes (how much the woman wanted to get pregnant and how hard she tried to avoid the pregnancy) did not explain the difference in post-conception happiness between Hispanics and Whites. These results suggest that higher happiness among Hispanic women was mostly the result of adaptation to unintended pregnancies after conception has occurred. One implication of this is that higher Hispanic happiness is probably not a satisfying explanation for why unintended pregnancy is more common among Hispanics compared with Whites.

One challenge of this study is that several of the variables – including pregnancy happiness, pregnancy intentions, and how hard the woman was trying to avoid pregnancy – are inherently subjective and the way women responded to these questions might have varied by race-ethnicity and nativity. The fact that the meaning of unintended pregnancies in women's lives seemed to vary based on Hispanic ethnicity was, in fact, part of the motivation for the study. That there is little racial-ethnic difference in happiness about intended births increases confidence that the observed difference in happiness about unintended births is “real.”

An additional limitation was that the number of variables available in the NSFG to capture socioeconomic status and social support was limited. The finding that socioeconomic variables were relatively unimportant is bolstered by the fact that White and Black women reported nearly identical levels of happiness about unintended pregnancies, despite having different average socioeconomic status. As data become available, future research should examine the predictors of pregnancy happiness using a greater variety of socioeconomic and support measures, including the woman's income and the partner's income at the time of conception, as well as the availability of emotional and instrumental support from the woman's parents, siblings, and other relatives. Accounting for these various sources of support is important since prior research has demonstrated a link between social support and fertility behaviors (Newman 2009).

Future research could also try to disentangle the mechanisms by which religiosity is associated with pregnancy happiness. One puzzle is that Black women are more religious than Whites but the two groups have similar average happiness scores. This could be due to denominational differences or it could be that for Black women the religiosity “advantage” is offset by disadvantages in other areas, such as having less support from male partners. Or, it could be that religiosity is only important among foreign-born Hispanics because it is an indicator of acculturation. Further, it may be worth exploring possible interactions between marriage and religiosity. Because of religious sanctions on nonmarital sex and childbearing, it may be that the association between religiosity and pregnancy happiness differs for married versus unmarried women. In addition, the relationship between ethnicity and happiness might vary depending on whether the birth was slightly mistimed, greatly mistimed, or not wanted at any point in the future. Finally, future research might explore variation in Hispanic women's feelings about unintended births by subgroup. This analysis focused on the distinction between US-born and foreign-born Hispanic women, but there may also be variation by country-of-origin.

Implications for Research and Policy

Because happiness regarding pregnancy varies by race-ethnicity and nativity, future research on pregnancy intentions might be improved by accounting for women's happiness. Measures of pregnancy intentions are used for a range of purposes and happiness might be a particularly useful or important addition for some of these applications compared with others. First, pregnancy intentions are used as a predictor of prenatal care and child health outcomes. However, happiness may act as a mediator between intendedness and health outcomes. If Hispanic women are better able to cope with unintended births, this might result in better outcomes for children and mothers. This might help explain the “Hispanic Paradox” as it relates to birth outcomes. Future research should explore whether this is the case.

Second, pregnancy intentions are frequently used as an indicator of “unmet need” for contraception, and one of the goals of Healthy People 2020 was to reduce the incidence of unintended pregnancy (United States Department of Health and Human Services 2010). However, the findings presented here highlight the heterogeneity in women's feelings about unintended births and this heterogeneity may extend to unintended pregnancies as a whole. While policymakers should continue to pursue the goal of reducing unintended pregnancies, it should be acknowledged that it might be relatively difficult to prevent those pregnancies that women report being happiest about.

In addition, future research that seeks to document unintended pregnancy or births, or evaluate the consequences of unintended births, should consider using pregnancy happiness in addition to pregnancy intentions, particularly when making racial-ethnic comparisons. There are several possibilities for integrating this information. First, measures of pregnancy happiness and measures of pregnancy intentions might be used side-by-side, for example, in regressions where pregnancy intention is used as an independent variable. In analyses that compare the incidence of unintended pregnancy across populations or over time, it might be useful to report both the full proportion of unintended pregnancies and the proportion of untended pregnancies that women are unhappy about.

Conclusion

The findings presented here demonstrate that the experience of having an unintended birth is different for Hispanic women compared with White and Black women. For Hispanic women, unintended births are less unwanted both before and after conception. Hispanic women who are both foreign-born and religious report the greatest happiness compared to other groups. The results confirm previous findings that intention status alone is incomplete for capturing pregnancy wantedness, and demonstrate that pregnancy happiness offers a particularly important complement when making comparisons of unintended births by race-ethnicity and nativity.