Anticipated Emotions about Unintended Pregnancy in Relationship Context: Are Latinas Really Happier?

Abstract

This study examined differences in women's anticipated emotional orientations towards unintended pregnancy by relationship status and race/ethnicity. Data from a prospective survey of 437 women aged 18-44 who intended no more children for at least two years were analyzed along with 27 in-depth interviews among a diverse sub-sample. Cohabiting women and women in a romantic relationship not living together were less likely to profess happiness (OR=0.42, p<.05, OR=0.25, p<.01, respectively), even when partners’ intentions/feelings were controlled. The most prominent factor underlying negative feelings was partners’ anticipated lack of engagement with the emotional, physical, and financial toll of unintended childbearing. Contrary to conventional wisdom regarding the “Hispanic paradox”, foreign-born and US-born Latinas were no more likely to profess happiness than non-Hispanic whites or blacks. Moreover, foreign-born Latinas whose survey responses indicated happiness often revealed highly negative feelings at in-depth interview, citing pressure to conform to sociocultural norms surrounding motherhood and abortion.

BACKGROUND

Unintended pregnancy is a persistent public health and policy issue in the United States; data from the National Survey of Family Growth (NSFG) indicate that just under half of pregnancies occurring each year are reported as unintended—either occurring too soon (mistimed) or not intended at any time in the future (unwanted) (Finer & Zolna, 2016). Large disparities between exist between demographic groups and have persisted over the last decade (Finer & Zolna, 2014; Finer and Zolna 2016). Rates of unintended pregnancy are higher among women who are in cohabiting compared to marital unions and among Latina and African-American women compared to non-Hispanic whites (hereafter, whites) (Finer & Zolna, 2016). Similar demographic patterns also apply to unintended births, where disparities by race/ethnicity and union status persist at every level of socioeconomic disadvantage (Sweeney & Raley, 2014).

Although the majority of research to date has emphasized variation by relationship context and race/ethnicity in the intention status of births, differences have also been demonstrated with respect to women's feelings about those births that are unintended. In particular, Latina women are commonly viewed as being happier about unintended births—the so-called “Hispanic paradox”. Previous research using the 2002 and 2006-2010 NSFG has indicated that Latina women reported greater mean happiness (on a 1-10 scale) about unintended births compared to African American or white women (Chandra et al., 2005; Hartnett, 2012), and that foreign-born Latinas are particularly likely to profess happiness (Hartnett, 2012). Many researchers have suggested that family-oriented cultural norms and religious influence in Latin America precipitate positive attitudes towards children. An ethnographic study of Mexican women in the 1980s found that women generally reported wanting to have a greater number of children than they felt they could afford (LeVine, 1993), and other studies have suggested that better maternal behaviors and neonatal outcomes among Mexican mothers and infants may be attributable to positive cultural orientations towards childbearing (Scribner & Dwyer, 1989; Zambrana et al., 1997). Differences in feelings about unintended births by union status have not previously been explored, but prior work has suggested that desire to have a baby with a male partner is highly dependent how well established the relationship is and the degree to which the male partner can be relied upon to successfully adapt to fatherhood (Wilson & Koo, 2006).

These prior insights highlight three important gaps in current understanding of differentials in unintended fertility by union status, race/ethnicity, and nativity. First, most studies are limited to analyzing the intention status of births, either because only women who have had a live birth are included in the sampling frame or because of under-reporting of pregnancies resulting in abortion. Focusing on births rather than pregnancies introduces an element of selection bias, because those pregnancies that are most unintended and unwelcome are less likely to result in a live birth and thus to be included in the sample. Second, most studies rely upon retrospectively reported pregnancy intentions and feelings, which may be very different to those formed prior to discovery of pregnancy or before the decision to continue the pregnancy to birth was made. Findings from retrospective data are thus likely to be subject to recall bias, whereby intentions and feelings about a pregnancy asked about once birth has occurred tend to be more positive than those elicited during or before pregnancy (Joyce et al., 2002; Koenig et al., 2006; Miller & Jones, 2009; Rosenzweig & Wolpin, 1993), and they become increasingly positive over time, as the child increases in age (Bankole & Westoff, 1998). Demographic patterns in prospectively measured intentions and feelings about pregnancy are yet to be fully explored, particularly with respect to the Hispanic paradox. It is possible that the greater happiness about unintended births reported among Latina women is partly accounted for by the differential influence of retrospective bias in the context of cultural norms surrounding childbearing identified by prior work on the Hispanic paradox (Scribner & Dwyer, 1989; Zambrana et al., 1997).

Finally, most previous studies focusing on relationship status and race/ethnic differences in unintended fertility have employed only single-item survey categorizations (intended versus unintended or happy versus unhappy), which likely represent an oversimplification of women's complex conceptualizations of pregnancy (Bachrach & Newcomer, 1999; Klerman, 2000; Miller, 2011; Santelli et al., 2009). In reality, not all unintended pregnancies are unwelcome or undesired; some may be unexpected but welcome, whereas others may be unexpected and potentially detrimental (Aiken et al., 2015; Borrero et al., 2015). Taking into account women's feelings about pregnancy no doubt helps us to distinguish happy accidents from unhappy ones, but answers to single-item survey questions cannot provide insight into factors underlying why a pregnancy would be greeted with positive or negative emotions. Life circumstances, opportunity costs, and partner influence, among other factors, may all shape women's emotional responses to unintended pregnancy (Aiken et al., 2015; Borrero et al., 2015; Edin & Kefalas, 2005; Sable & Libbus, 2000), but comparisons of these underlying influences by relationship status and race/ethnicity are yet to be undertaken.

In this paper, we begin to address these gaps by measuring prospective emotional orientations towards pregnancy among a diverse sample of women who are using contraception and who are not intending to have more children for at least two years. Our aims are to: 1) examine whether anticipated emotional orientations towards unintended pregnancy vary by relationship status and by race/ethnicity; and 2) to provide insight into the factors that influence and shape these orientations. We accomplish these aims using a mixed-methods approach that includes both survey and in-depth interview components. This approach allows empirical associations to be identified as well as offering detailed explanations for these associations from women's perspectives.

METHODS

Survey

The quantitative component of our study draws upon an analytic survey sample of 437 women aged 18-44 years, living in Austin and El Paso, Texas who, at the time of interview, were not intending to have more children for at least two years and were not pregnant or sterilized. Our analytic sample is part of a cohort of 803 women recruited immediately following delivery from three hospitals in Austin and El Paso between July and November 2012 and followed over 9 months at 3-monthly intervals (full details in (Potter et al., 2014)). Postpartum recruitment offers the dual advantages of: 1) ensuring that all women in the sample are parous (since pregnancy intentions and feelings among nulliparous women are likely to be different); and 2) allowing prospective measurement of pregnancy intentions over time, because women are not pregnant at the time of recruitment. Moreover, women are at high risk of experiencing unintended pregnancy in the postpartum period—a third of unintended pregnancies take place within 18 months of a previous birth (Kuroki, 2008)—allowing us to shed light on a particularly vulnerable population. Future pregnancy intentions were assessed by asking women whether they intended to have any more children in the future. Those who answered “yes” were then asked when they intended to have another child (responses were given in number of years). Proportions of women in each racial/ethnic and relationship status group who intended to limit versus space future childbearing were not significantly different. Previous findings using nationally representative data from the NSFG similarly indicate that Latina women (which represent the majority of our sample) are no more likely than white women to experience a mistimed birth (Sweeney & Raley, 2014). Thus, in our analyses, we followed the standard NSFG definition of unintended pregnancies as those that would be either mistimed or unwanted.

Our sample is drawn from the 9-month time-point because in this survey women were asked about their own and their perceptions of their partners’ pregnancy intentions and feelings about pregnancy. At the time of the survey, all women were using some form of contraception. The 9-month follow-up survey took place by phone and women were compensated $15 for their time. Participant retention at 9 months was 84%. We excluded from our sample 116 women who were sterilized following delivery or during follow-up or whose current partner had a vasectomy, because these women were not asked about their or their partners’ feeling about pregnancy; 28 women who had become pregnant by the 9 month follow-up; 9 women who self-identified as Asian, Asian-American, or Native American, because they are too few to be meaningfully analyzed as separate racial/ethnic groups; 48 women who intended to have another child within one year (because for women who intended more children within one year, a pregnancy in the next three months would not be unintended); 23 women who were not sure about their feelings about pregnancy in the next three months and 11 women who were not sure about their partners’ intentions or feelings, because their responses are likely to represent something different about the anticipated impacts of a pregnancy that are beyond the scope of this study, and because they are too few in number to be represented as a separate group. While Latina and African-American women are more likely to hold ambivalent desires to avoid pregnancy (Schwarz et al., 2007), we did not measure desire to achieve or avoid pregnancy in this study.

Variables

Happiness about a future pregnancy was measured by asking, “How would you feel if you became pregnant in the next three months?” with responses recorded on a four-item ordinal scale: “Very happy”, “Somewhat happy”, “Somewhat upset” and “Very upset”, as well as a “Don't know” option (as stated above, women who chose this option are excluded from our analyses). These question constructs follow those employed in the Border Contraceptive Access Study (BCAS), which involved a study population with similar demographic characteristics. We combined “Very happy” with “Somewhat happy” and “Very upset” with “Somewhat upset” to create a dichotomous variable (happy vs. unhappy) (Aiken & Potter, 2013).

Previous research has highlighted the importance of women's perceptions of their partners’ childbearing intentions in shaping their own fertility decision-making (Fischer et al., 1999; Stanford et al., 2000; Zabin et al., 2000). In our survey, perceptions of partners’ pregnancy intentions were measured by asking: “Does your husband or partner intend to have more children in the future?” and responses options were “Yes”, “No”, or “Don't know”. Perceptions of partners’ feelings about pregnancy were measured by asking: “How would your husband or partner feel if you became pregnant in the next three months?” with answers coded using the same four-item ordinal scale as for women themselves. These questions were asked only of women who were currently in a relationship (i.e. those who were married, cohabiting, or in a romantic relationship but not living together). Following previous work examining nuance in men's pregnancy intentions (Higgins et al., 2012), we combined women's answers to these questions into a variable representing women's perceptions of their partners’ attitudes—both his intentions and his feelings—coded according to the following categories: “Intends more and happy”, “Intends more and unhappy”, “Intends no more and happy”, “Intends no more and unhappy.” Those who intend more children in the future can be thought of as partners who want to space future children, while those who intend no more can be thought of as partners who want to limit future childbearing.

Relationship status was measured as married, cohabiting, in a romantic relationship but not living together, and single (including staying friends with a former partner), with categories based upon those from the Fragile Families and Child Wellbeing Study (Reichman et al., 2001). Because we were able to measure relationship status longitudinally in our survey, we also constructed a variable representing relationship stability between baseline and the 9-month follow-up survey. We ordered relationship types according to a categorical scale from married to single, with cohabiting and romantic relationship not living together as two successive middle categories. We then coded change in type of relationship according to three categories: 1) relationship type remained stable (e.g. married at baseline and remained married at 9 months); 2) relationship type changed from higher to lower on the scale (e.g. from married to single); 3) relationship type changed from lower to higher on the scale (e.g. from romantic relationship not living together to cohabiting).

Race/ethnicity takes account of both participants’ self-reported racial/ethnic group and their country of birth, with participants falling into one of four categories: non-Hispanic white, African-American, US-born Latina, and foreign-born Latina. The vast majority of foreign-born Latinas in our sample were born in Mexico, but we include the small number born in other Latin American countries (n=8) in the same category. Data on demographic and socioeconomic variables shown in previous literature to affect fertility decision-making, including age, education, parity, and income (Finer & Zolna, 2011, 2014) were collected at the baseline interview. Type of health insurance, a marker of access of contraception and healthcare services, was ascertained at baseline and at each subsequent follow-up interview. We treat age as a categorical variable: 18-24 years, 25-29 years, 30-34 years, and 35 years and over. Education was measured as years of formal schooling, which we then divided into three categories: less than high school, completed high school, and beyond high school. Parity is categorized into 1, 2, and 3 or more children. All of the women in our sample were recruited immediately postpartum and have given birth to at least one live born child. Income was asked about as either a monthly or yearly household estimate and was self-reported using an iPad rather than asked by the interviewer to maximize response rates and accurate reporting. We converted all answers into yearly estimates and categorized them as less than $10,000, $10,000-24,999, $25,000-49,999, and $50,000 and over. Health insurance status at the time of the 9-month follow-up was categorized as public insurance (Medicaid or Women's Health Program), private insurance, or no insurance.

Statistical analyses

We performed chi-squared tests to examine differences in the distributions of each categorical variable between the group of women who would be happy about a future unintended pregnancy and the group who would be unhappy. We then modeled the association between positive emotional orientation towards an unintended pregnancy and both relationship status and race/ethnicity using binary logistic regression and controlling for other key demographic variables described above. We then limited our analysis to the sub-sample of women who were currently in a relationship to test further the associations between positive emotional orientation towards an unintended pregnancy and both relationship status and race/ethnicity when perceived partner intentions and feelings are taken into account. Again, we used binary logistic regression and controlled for the same panel of demographic covariates described above.

We also conducted several sensitivity analyses, including treating our dependent variable (happiness versus unhappiness) as a ranked categorical variable (very happy, somewhat happy, somewhat upset, and very upset). The results of the ordinal logistic model did not differ substantively or significantly from those of the binary logistic model and are thus not shown. We also ran all analyses both including and excluding the group of women who were unsure about their future childbearing intentions (n=52). Results did not differ substantively or significantly. Thus, in the analyses presented here, we opted to include those who were unsure of whether or not they intended to have more children in the future, because they also fall on the spectrum of unintended pregnancies. All analyses were conducted using Stata version 13.0 and statistical significance was set at an alpha level of .05.

In-Depth Interviews

The qualitative component of our study consists of semi-structured in-depth interviews with a sub-sample of 27 survey participants interviewed between October 2013 and January 2014 (shortly after completion of the survey). Participants were aged between 18-44 years, consistently intended no more children for at least two years over the course of the survey, and professed either happiness or unhappiness about a pregnancy in the next three months. Our aim was to document a range of experiences and thus our sample size was driven by gaining desired diversity and by thematic saturation, which was determined by continuous monitoring during the interview phase. Participants were selected to obtain both women who said they would be happy and unhappy at the prospect of an unintended pregnancy, a mix of white, U.S.-born Latina, and foreign-born Latina women (we limited our in-depth interviews to these groups of women because they are the main groups represented in our survey sample), and a mix of publicly and privately insured women. Our sample contained both married and cohabiting women, and was diverse in terms of income, age, and parity. Sample characteristics are shown in Table 1.

Table 1.

Characteristics of In-Depth Interview Participants

Characteristic	Mean or Frequency (N=27)
Age (years)	29.5 (range: 19-40)
Parity	3 (range: 1-7)
Relationship Status
Married	17
Cohabiting	10
Ethnicity/Nativity
Foreign-born Latina	11
US-born Latina	8
Non-Hispanic white	8
Level of Education
Less than High School	7
High School	10
Greater than High School	10

The research team consisted of the first author and two research assistants. We used insights from the survey to develop an in-depth interview guide, which focused on exploring anticipated emotional reactions to discovering a pregnancy, eliciting the factors underlying such emotions, discussing deviation from survey responses to questions regarding feelings about pregnancy, and exploring the nature of close relationships, including with partners and family members. While we endeavored to cover these topics during the course of the interview, we allowed the conversation to be guided by women's responses and often spontaneously pursued issues they raised. We interviewed women in-person in their homes, offering $30 in compensation for their time. Interviews were audio recorded with participants’ permission, transcribed verbatim, and translated from Spanish where necessary by a native Spanish-speaker who had experience working with the study population. Two members of the research team then coded each transcript using an iteratively developed coding guide. Through a series of team meetings, all three team members met to review and discuss each coded transcript and resolve any points of disagreement. We used Atlas.ti qualitative software to help organize the coded data, and created a matrix of coded categories for married versus cohabiting and foreign-born versus US-born and white women to help with the visual identification of key themes. All team members were involved in the analysis and interpretation of the coded transcripts and in the identification of key themes.

All participants gave signed informed consent for participation in both the survey and in-depth interview study components; human subjects approval for this study was obtained from the institutional review boards at the University of Texas at Austin and Princeton University.

RESULTS

Survey

Table 2 shows demographic characteristics for our sample, which was diverse with respect to age, parity, race/ethnicity, relationship status, education, and income. Forty-six percent of the sample was married, 28% cohabiting, 12% in a relationship but not living together, and 14% single. Three quarters were in the same type of relationship as they were at baseline, whereas the others had transitioned either up or down the relationship scale in approximately equal proportions. Thirty percent were U.S-born Latina, 48% foreign-born Latina, 6% African-American, and 16% white. The sample was divided into approximately equal thirds with respect to parity one, two, and three or more (35%, 33%, and 33%, respectively). Thirty percent had less than a high school education, 29% had completed high school, and 41% had education beyond high school. Fifty-seven percent had no health insurance, 27% private insurance, and 16% public insurance. Just over a third (36%) had a household income of less than $10,000 per year, 31% between $10,000 and $24,999, 16% between $25,000 and $49,999 and 17% $50,000 or over. Among women who were in a relationship, the majority (58%) believed that their partner wanted more children in the future and would be very happy about a pregnancy in the next three months. Comparing those that would be happy about an unintended pregnancy (n=268) to those that would be unhappy (n=169), we observed significant differences in the distributions of relationship status and perceptions of partners’ intentions and feelings about pregnancy.

Table 2.

Sample Characteristics Among Women who Intend no More Children for at Least Two Years

Characteristic	Frequency (%) Entire Sample (N=437)	Frequency (%) Women with Positive Feelings About Pregnancy (n=268)	Frequency (%) Women with Negative Feelings About Pregnancy (n=169)
Age (p=.242)
18-24	37.5	35.4	40.8
25-29	31.6	33.2	29.0
30-34	18.5	20.5	15.4
35+	12.4	10.8	14.8
Race/ethnicity (p=.414)
US-born Latina	29.7	30.6	28.4
Foreign-born Latina	47.8	48.5	46.7
African-American	6.4	4.9	8.9
Non-Hispanic white	16.0	16.0	16.0
Relationship Status (p=.003)
Married	46.0	51.9	36.7
Cohabiting	27.7	27.6	27.8
Relationship not Living Together	11.9	9.7	15.4
Single	14.4	10.8	20.1
Relationship Stability (p=.272)
Remained in same type of relationship	75.1	76.9	72.2
Moved up relationship scale	13.3	11.2	16.6
Moved down relationship scale	11.7	11.9	11.2
Parity (p=.279)
1	35.0	37.7	30.8
2	32.5	32.1	33.1
3+	32.5	30.2	36.1
Education (p=.656)
Less than High School	29.7	31.3	27.2
Completed High School	29.1	28.4	30.2
Beyond High School	41.2	40.3	42.6
Insurance Status (p=.924)
Public	15.8	15.3	16.6
Private	27.0	26.9	27.2
None	57.2	57.8	56.2
Yearly Household Income (p=.968)
Less than $10,000	35.5	34.7	36.7
$10,000-24,999	31.4	32.1	30.2
$25,000-49,999	16.2	16.4	16.0
$50,000 or more	16.9	16.8	17.2
Perceptions of Partner's Pregnancy Intentions and Feelingsa (p=.000)
Intends no more and upset	11.8	3.4	26.7
Intends no more and happy	13.4	19.0	3.7
Intends more and upset	16.9	5.0	37.8
Intends more and happy	57.8	72.6	31.9

Note: All figures in percentages.

^aIncludes only those women who are in a relationship (n=372)

p-values indicate the results of chi-squared tests testing differences in the distribution of each variable between the group of women who would be happy about an unintended pregnancy and the group of women who would be unhappy.

Table 3 shows the results of our multivariable analysis of factors associated with positive emotional orientations towards a pregnancy that would be unintended for all women in our sample. We found that cohabiting women, women in a relationship but not living together, and single women were all significantly less likely to feel positively about an unintended pregnancy compared to married women. The magnitudes of the odd ratios indicate a progressive decline in happiness by relationship status from cohabiting women (OR=0.50, p<.05), to women in a relationship but not living together (OR=0.27, p<.01), to single women (OR=0.20, p<.001). Contrast testing by varying the reference category (not shown in table) indicated that while single women were less likely to profess happiness when compared to cohabiting women (OR=0.41, p=.014), no significant difference existed between cohabiting women compared to those in a relationship not living together (OR=0.64, p=.10). Age and parity were also associated with positive orientations towards unintended pregnancy. Women with two children or with three or more children were less likely to profess happiness about a potential unintended pregnancy compared to women with only one child (OR=0.57, p<.05, OR=0.45, p<.01, respectively). Compared to women aged 18-24, women aged 30-34 were more likely to profess happiness (OR=2.17, p<.05). We found no significant associations between happiness and race/ethnicity.

Table 3.

Binary Logistic Regression Showing Factors Associated with Happiness About Pregnancy in the Next 3 Months Among Women who Intend no More Children for at Least Two Years (N=437)

Variable	Odds Ratio	P-value
Age
18-24	ref
25-29	1.60	.083
30-34	2.17	.027
35+	1.18	.685
Race/ethnicity
Non-Hispanic white	ref
African-American	0.76	.607
US-born Latina	1.23	.566
Foreign-born Latina	0.82	.601
Relationship Status
Married	ref
Cohabiting	0.50	.014
Relationship not Living Together	0.27	.001
Single	0.20	.000
Change in Relationship Status
Stable	ref
Moved up Scale	1.55	.228
Moved down Scale	1.23	.557
Parity
1	ref
2	0.57	.037
3+	0.45	.007
Education
Less than High School	ref
Completed High School	0.71	.236
Beyond High School	0.60	.107
Insurance Status
Private	ref
Public	0.94	.887
None	0.95	.885
Yearly Household Income
Less than $10,000	ref
$10,000-24,999	0.82	.456
$25,000-49,999	0.60	.165
$50,000 or more	0.43	.103

Note: All coefficients are expressed as odds ratios.

ref = reference category for each variable

Table 4 shows the factors associated with positive emotional orientation towards an unintended pregnancy among the women in our sample who had a male partner at the time of the survey. Here, we see that women's perceptions of their partners’ pregnancy intentions and feelings about a potential pregnancy were very strongly associated with their own happiness. Women who perceived that their partners did not intend more children but would be happy about a pregnancy have the highest likelihood of feeling positively themselves (OR=52.4, p<.001). Women who perceived that their partners intended more children and would be happy about a pregnancy were next most likely to feel positively themselves (OR=20.2, p<.001). We also found persisting associations between feelings about pregnancy and relationship status. Women in cohabiting relationships or in a relationship but not living together were significantly less likely than women in a marital union to feel positively about the prospect of an unintended pregnancy (OR=0.42, p<.05, OR=0.25, p<.01, respectively). Again, contrast testing by varying the reference category (not shown in table) indicated no significant difference between cohabiting women compared to those in a relationship but not living together (OR=0.60, p=.27). The associations between happiness and age and parity found previously no longer persisted, and we found no associations between happiness and race/ethnicity. We also tested for interactions between women's perceptions of their partners’ intentions and feelings and relationship status, but found no significant associations (results not shown).

Table 4.

Binary Logistic Regression Showing Factors Associated with Happiness About Pregnancy in the Next 3 Months Among Women in a Relationship who Intend no More Children for at Least Two Years (N=372)

Variable	Odds Ratio	P-value
Perceptions of Partner's Pregnancy Intentions and Feelings
Intends no more and upset	ref
Intends no more and happy	52.38	.000
Intends more and upset	1.02	.975
Intends more and happy	20.19	.000
Age
18-24	ref
25-29	1.48	.296
30-34	1.75	.233
35+	0.99	.986
Race/ethnicity
Non-Hispanic white	ref
African-American	1.69	.486
US-born Latina	1.38	.533
Foreign-born Latina	0.74	.584
Relationship Status
Married	ref
Cohabiting	0.42	.015
Relationship not Living Together	0.25	.006
Change in Relationship Status
Stable	ref
Moved up Scale	1.77	.356
Moved down Scale	0.93	.857
Parity
1	ref
2	0.64	.236
3+	0.45	.055
Education
Less than High School	ref
Completed High School	0.47	.058
Beyond High School	0.68	.406
Insurance Status
Private	ref
Public	0.81	.724
None	1.11	.842
Yearly Household Income
Less than $10,000	ref
$10,000-24,999	0.88	.733
$25,000-49,999	0.50	.162
$50,000 or more	0.47	.279

Note: All coefficients are expressed as odds ratios. ref = reference category for each variable

In-Depth Interviews

To gain contextual insight into the empirical associations between happiness about unintended pregnancy and relationship status demonstrated in our models, we turn next to the findings from our in-depth interviews. We focus here firstly on the experiences of women in cohabiting relationships. Several key themes concerning the perceived psychosocial consequences and life impacts of an unintended pregnancy arose from our exploration of these women's emotional orientations. (None of the women in our in-depth interview sample were in a relationship but not living together or single, so we could not explore those associations in further detail). Secondly, although we did not observe empirical associations between race/ethnicity/nativity and happiness about pregnancy in our survey data, our in-depth interviews revealed stark differences between the answers many of the foreign-born Latina women we interviewed gave to the survey questions about their feelings and the emotional orientations they described in intimate conversations. The perceived life impacts underlying negative feelings among foreign-born Latinas were very different from those offered by white and US-born Latina women (whose answers about feelings in the survey and in-depth interviews were uniformly consistent).

Emotional Orientations among Cohabiting Women

Comparing women in married versus cohabiting unions, we found that married women tended to feel more positively towards unintended pregnancy—even in cases where they perceived that their partners would not be happy—because they held expectations of financial and/or emotional support from their partner if an unexpected pregnancy were to occur. This theme is well illustrated by Adriana, a 34 year-old US-born Latina woman who is married with five children. She told us of her husband:

“He would just go with the flow. If you have a baby you have to be able to feed it. You don't have a choice. That's his belief. If I said “how are we gonna do it?” he would say “just like we're doing everything else with all the others.” He's there financially. He's just like “Watch. We're going to be OK.” And when I'm feeling overwhelmed, he listens. When I need time to for me, like if I want to go work out, he says “Just go, I'll pick the girls up.” He has my back, he's what makes me stand straight.”

Vanessa, a 29 year-old white woman who is married with two children, echoed Ariana's experience:

“My husband and I work well together and I think we can rise to any kind of challenge that life brings us. He's the only one working right now and so he'd probably be a little scared at first, but once he got used to the idea he'd be excited and wouldn't foresee any problems. He's the support person...I'm confident we'd be able to handle it together.”

By contrast, a strong theme articulated by cohabiting women was that their partner would not understand the huge investment of time, resources, and physical and emotional energy another child would entail, nor the opportunity costs in terms of her own personal goals. Karina, a 31 year-old US-born Latina woman who has five children and is in a cohabiting relationship explains:

“I tell him that I don't want any more children. I would rather leave him than have more children. He would be happy with another one, but I'm so tired...your life revolves around your children, and my goal is to raise healthy children that grow up in a stable home and don't end up passed from relative to relative. I want them to reach that goal of going to university and being somebody. And if I can't pay for college, I will be limiting their future. I thought before that if I didn't have sex when he wanted to, he would go find somebody else. So I ended up with a lot of kids. But he doesn't have to get up if the baby is crying. He doesn't know anything about that. He doesn't know if the baby has a fever, if she sleeps or doesn't sleep, if the children go to school or anything. I tell him sometimes “I wish you could have a taste of your own medicine so you would know what giving birth is like.” It's nice to say you're a father. It's a good feeling, but it doesn't mean you know how to raise your child.”

Marta, a 29 year-old Mexican-born Latina woman, who is in a cohabiting relationship and has four children, echoes Karina's worries about a lack of help and support from her partner:

“He doesn't help when they are babies. We both work, but he doesn't do all the stuff a mother does. I'm the one who enforces the discipline. So for him to have another baby, it's not so hard. He doesn't feel responsible. He would just say ‘one more child, let's welcome him. He doesn't think about the cost of raising a child. If I got pregnant I would have to stop working, and then I would get depressed because I love my job and it gives me support. He's the reason I did not get my tubes tied. I had signed the form and everything, but he kept telling me reasons why I shouldn't have the surgery and talked me out of it. If I had done it, I think we would have ended up separated.”

Marta was not alone in this situation where, despite her own unhappiness at the prospect of another pregnancy, her partner was opposed to the idea of permanent methods of contraception and wanted to create a large family despite the stress that might ensue. Lydia, a 23 year-old US-born Latina woman who is in a cohabiting relationship and has one child told us:

“If I got pregnant...it would be like ‘are we going to stay together or are we not going to stay together?’ I wanted to get my tubes tied but he wants a big family and was totally against it. The biggest reason I had kids in the first place is because he wanted them. But the stress of everything makes it not enjoyable. My nephew lives with my mom, my niece lives with the dad, and CPS (Child Protective Service) was involved. I felt like if I'm not in a stable situation, the same could happen to me. Our parents thought that when we had our baby we might get married and do everything the right way because we're kind of doing it backwards.”

Commonly, cohabiting women explained that such positive attitudes towards childbearing among their partners arose because another baby would be “his kid”. Misty, a 19 year-old white woman who has one child and who is in cohabiting relationship tells us:

“Me and his dad had a lot of problems. He has two kids from a previous relationship and has to pay child support. He was trying to get with other girls so I took both of them [his daughter and their son] and left. That's when he changed. If I got pregnant now he'd be happy. It'd be his baby.”

Marie, a 26 year-old US-born Latina cohabiting mother of one echoed Misty's thoughts:

“He's the one who wanted a baby. Right from the beginning. And now he's like “you know he needs a brother or sister, right?.... You know you're going to have another one, right?” He has a daughter from when he was living with this girl for like 5 months and he pays child support but she doesn't really let him see her. So, it's hard for him and I think he wants more to kind of make up for it somehow. To start over.”

Overall, the overarching theme among cohabiting women in our sample was that while they viewed their partners as highly enthusiastic about the idea of having more children, they also perceived a serious lack of engagement with the day-to-day realities of bringing up a child.

Emotional Orientations among Foreign-born Latina women

Our empirical results show that foreign-born Latina women are no more likely than white women in our sample to profess happiness about an unintended pregnancy. Moreover, individuals who professed happiness about a potential unintended pregnancy in response to the survey question often revealed deeply negative emotions at in-depth interview. When asked about why their answers were different, a strong theme was perceived pressure to conform to prevailing social and cultural norms surrounding motherhood and childbearing, reinforced by partners, friends, family members and the wider community. Blanca, a 31 year-old Mexican-born Latina woman with five children illustrated these pressures when she explained:

“For me, the feeling would be total frustration. But with the last one I would complain a lot about being pregnant because I didn't want more children. I would say that it wasn't what I had planned, and people still hold it against me... I hope that God forgives me and he is not going to send me another one.....people say you shouldn't be complaining when a baby is a gift from God.”

Like cohabiting women, foreign-born Latinas who revealed negative emotional reactions to unintended pregnancy perceived a lack of support from male partners. This time, there was a strong sense that such behavior was culturally sanctioned. Naomi, a 38-year old Mexican-born Latina woman with four children echoed the experiences of many others when she told us:

“I can say this about my family and even about Mexicans, because I am from Mexico: Men only have to take care of their work outside the home. But women? On top of having to work outside the home, we also have the responsibility of running the home, cooking, cleaning, and caring for the children. I have to do it all. In our culture, they think it's a woman's responsibility to do all those things. That's what we see in all of our families. But I can see from my co-workers that it's different: American men share more responsibilities at home. I have male co-workers that are black, and they cook. And some that are white, and they do it too. With Mexicans, the woman is the one with all the responsibilities.”

A third major theme among foreign-born Latinas, which was absent from the narratives of US-born Latina and white women was that they also perceived fewer options for dealing with pregnancy, including fear of being found out if they chose abortion. Aida, a 29 year-old Mexican-born Latina woman with four children explained:

“The biggest challenge for me would be to erase from my head the idea that abortion exists. It would be a challenge to deal with five children instead of four, but I don't want abortion to cross my mind. I can't think about abortion as a solution because people think that if you make love and you create this baby...it would be wrong. Everybody comes from big families and everybody loves babies. No, I live in fear of having another child, but I can never consider abortion, never even think that it exists.”

DISCUSSION

Our findings demonstrate variation in anticipated emotional orientations towards unintended pregnancies by both relationship and cultural context. Among women in our sample, cohabitors and those in a relationship but not living together were at higher risk of feeling negatively about an unintended pregnancy compared to married women, even when their socioeconomic status, educational level, and perceptions of their partners’ childbearing intentions and feelings about pregnancy were accounted for. Contrary to the conventional wisdom of the Hispanic paradox, which predicts greater happiness about unintended pregnancies among Latinas (particularly those who are foreign-born) on the basis of retrospective reports, we did not observe empirical differences in happiness about an unintended pregnancy by race/ethnicity or nativity in our prospective analysis. Moreover, we found that the negative emotions of the foreign-born Latina women who participated in in-depth interviews were not fully or accurately reflected in their answers to the survey questions. These women were unique in that they felt comfortable revealing their negative feelings about a future unintended pregnancy only during intimate conversations.

The negative feelings of both cohabiting women and foreign-born Latina women were often underpinned by anticipated adverse life impacts of unintended pregnancy, including increased psychosocial stress, high personal opportunity cost, such as giving up a job, and reduced physical and emotional wellbeing. Despite perceiving that their male partners would hold positive feelings about a pregnancy, these women tended to believe that their partners would be not only unaware of their own negative emotions, but also that they would play a substantial role in shaping them.

There are many possible explanations for how male partners might negatively influence cohabiting women's feelings about a future pregnancy, including worries about adverse effects on the relationship or anxiety that the male partner may not want the child. Yet the range of experiences related by cohabiting women in our sample revealed that their anticipated adverse life impacts of an unintended pregnancy appeared to operate most prominently through: a) their perceptions that male partners would be unlikely to play a major role in the day-to-day work of looking after a child; and b) their impressions of their male partners as having very little understanding of the personal opportunity cost or emotional toll another child would entail.

Women's perceptions that their male partners would offer little help caring for the child and would fail to appreciate the work involved may be a function of the inherent structural nature of some cohabiting unions, including higher levels of relationships stress, fewer pooled resources, and lower relationship quality (Brown & Booth, 1996), all of which may make weathering the demands of an unintended pregnancy more difficult. Other work pointing to the importance of relationship quality has also suggested that couples who believe they have a future together are more likely to agree that they would continue an unintended pregnancy (Sassler et al., 2009). Although we did not find an empirical association between prior relationship transitions and emotional orientations towards unintended pregnancy, some women did mention that future childbearing may affect future decisions about whether or not to stay together. Indeed, prior research has suggested that much of the inequality in maternal and child wellbeing associated with cohabiting relationships is a function of relationship instability (Cavanagh & Huston, 2008; Osborne & Mclanahan, 2007; Smock, 2000).

Another possible explanation for greater unhappiness among cohabiting women is of a selection effect whereby men who are more likely to form cohabiting unions may also be less willing to be involved in childcare (Miller & Sassler, 2012). Based upon the insights offered by women in our sample, a more salient explanation is that such men are more likely to remain in cohabiting unions rather than proceed to marriage. Indeed, previous ethnographic work has suggested that such a lack of support, commitment, and engagement in caring for children is part of the reason why women in cohabiting unions are reluctant to marry their male partners (Edin & Kefalas, 2005). On the other hand, a lack of economic resources in cohabiting unions may mean that some male partners feel more pressure to provide and less time to devote to being engaged fathers, despite desire to do so (Edin & Nelson, 2013). While cohabitors in general appear to hold less traditional gender roles than their married counterparts in terms of labor-force participation, division of household labor appears to be equally gendered in both types of union (Smock 2000). Cohabiting women may thus more frequently experience the pressure of holding both a breadwinning and a domestic role, which may exacerbate any lack of help with childcare by their male partners. We also note that although we did not include women who were in a romantic relationship but not living with their partner in our in-depth interviews, these women appeared to be no more likely to profess happiness about an unintended pregnancy compared to cohabiting women. This finding suggests that in our sample, the character of cohabiting unions might have ben more similar to non-residential romantic relationships than to marriages, lending further support to the idea that union quality plays a key role in emotional orientations towards future childbearing.

Our findings among foreign-born Latinas run counter to much of the previous literature documenting highly positive and accepting attitudes towards children in Latina, and particularly Mexican, culture (Giachello, 1994; LeVine, 1993; Scribner & Dwyer, 1989; Zambrana et al., 1997). While we did not find empirical differences in happiness towards unintended pregnancies by race/ethnicity in any of our model specifications, we might have expected to find a higher likelihood of happiness among Latinas, and particularly among those who are foreign-born (Hartnett, 2012; Chandra, 2005). The most likely explanation for our divergent findings is that rather than asking women about their intentions and feelings after the birth of a child (as has been the case in prior research using retrospective data), we asked women to consider their anticipated reaction to pregnancy. Given foreign-born Latina women's explanations about the social taboo of calling a child undesired elicited during in-depth interviews, survey participants may have felt able to give a more candid response when speaking hypothetically, and when referring to a pregnancy rather than a child.

Moreover, the disconnect we observed between survey and in-depth interview responses regarding emotional orientations towards pregnancy suggests that even when asked prospectively, survey questions may not elicit accurate responses among all foreign-born Latinas. Indeed, many women cited precisely the social norms and cultural expectations regarding motherhood and childbearing highlighted in prior research when explaining why they felt the need to conceal their unhappiness about a pregnancy. Norms and expectations may thus have constrained Latina women's responses to survey questions, with negative emotions being more likely to be revealed through private conversations with trusted parties outside their social circle.

The issues voiced by foreign-born Latinas in our sample surrounding lack of acceptability of abortion for dealing with an undesired pregnancy may also help to explain our findings. Foreclosure of abortion as an option, or having to consider and seek care covertly without the support of family or friends, is likely to make an already stressful situation much worse (Rocca et al., 2013). Moreover, accepting birth as the only possible outcome of an undesired pregnancy raises the stakes regarding anticipated life impacts, contributing to negative emotional orientations. Once an unintended pregnancy has occurred, however, many women may feel resigned to the idea that they have no choice but to deal with the situation and consider the pregnancy welcome, perhaps partly accounting for the higher happiness levels found in retrospective studies.

Although examination of the happiness about unintended pregnancy at the intersection of race/ethnicity and union status was beyond the scope of our study, our findings suggest that it is an important area for future research. Prior work demonstrates variation in unintended births by race/ethnicity within relationship contexts. African-American and Latina women are more likely than white women to have children outside of marital unions (Martin et al., 2013), as well as to classify births within marital unions as unintended (Guzman et al., 2010; Musick, 2002). By contrast, births within cohabiting unions are more likely to be classified as planned or intended among Latinas, particularly those of low educational status (Manning, 2001; Musick, 2002). Latinas are also more likely than whites to be willing to consider having a child outside of marriage (Oropesa, 1996), suggesting differences in the meaning, acceptability or character of cohabiting unions within different social and cultural contexts (Manning, 2001). Prior work has also suggested differences in the nature of cohabitation by nativity (Landale & Oropesa, 2007; Osborne et al., 2007; Phillips & Sweeney, 2005), with foreign-born Latinas in particular having lower odds of unintended births within cohabiting unions compared to whites (Guzman et al., 2010). Although prior work has demonstrated that differences in relationship context do not explain differences in emotional orientations towards unintended births among Latina women compared to white or African-American women (Hartnett 2012), differences in emotional orientations towards unintended pregnancies or births among Latina women compared to white or African-American women within the same relationship context have yet to be examined,

Despite key strengths of prospective measurement of pregnancy intentions and feelings and the ability to collect detailed information on the factors underlying women's emotional orientations towards pregnancy, our study has several important limitations. First, our empirical results were based upon a relatively small and localized sample of women, all of whom had at least one child and many of whom were of low-income. We thus cannot expect our findings to be readily generalizable across other settings. We also likely lacked some statistical power in testing interactions between variables of interest, which should be further investigated using a larger sample. Second, our sample is limited to women who have had at least one child, and thus we cannot generalize our results to nulliparous women for whom an unintended pregnancy would be their first with their partner, or speak to differences by race/ethnicity, nativity, or relationship status in the impacts of initiating childbearing (although others have investigated a similar question (Nomaguchi & Milkie, 2003; Woo & Raley, 2005). Third, although our in-depth interviews allowed us to document a range of women's experiences, we must acknowledge the heterogeneity of cohabiting relationships and are cautious about making broad generalizations. Race/ethnicity and nativity are not monolithic categories either, and we cannot draw conclusions about all foreign-born Latinas on the basis of our predominantly Mexican sample. Finally, our analyses included women's feelings about an unintended pregnancy at only one point in time. Tracking feelings about unintended pregnancy (as well as pregnancy intentions) over time could illuminate factors associated with changes in feelings, including the role of relationship and family structure transitions. Although we are able to account for relationship transitions during the survey, we did not have information on whether such transitions involve a change of partner.

Women's emotional orientations towards unintended pregnancies help us to distinguish between different types of unintended pregnancies and may ultimately be more predictive of adverse outcomes than pregnancy intentions (Blake et al., 2007). Overall, our findings suggested that taking a more nuanced view of unintended pregnancy and distinguishing between unexpected and welcome versus unexpected and unwanted pregnancies may be a crucial component of both understanding demographic patterns in fertility behavior and identifying women, men, and children most at risk of adverse health and social impacts. Future research aimed at identifying those at highest risk of undesired pregnancies will help inform public health and policy efforts to empower women and couples to realize their reproductive goals, and provide a starting point for efforts to reduce inequities in undesired fertility.