GSJP Volume 16 final Graduate Student Journal of Psychology 2015 Vol. 16 Copyright 2015 by the Department of Counseling and Clinical Psychology Teachers College, Columbia University 47 Pregnancy Intentions and Happiness: Psychological Predictors of Breastfeeding in a National Sample Lindsey Wallace, Rebecca Chad, and Melissa Rodriguez Teachers College, Columbia University Unintended pregnancies (mistimed or unwanted) are linked to lower rates of breastfeeding, yet little research to date has investigated the role of feelings of happiness about the pregnancy on maternal health behaviors. Using data from the 2002 National Survey of Family Growth, we examined whether breastfeeding initiation/duration was explained by (1) socio-demographic characteristics of mothers and (2) pregnancy intentions/feelings of happiness RI �PRWKHUV��3UHJQDQF\�LQWHQWLRQV�KDSSLQHVV�IDFWRUV�VLJQLÀFDQWO\�H[SODLQHG�DQ�LQFUHDVH�LQ�WKH�PXOWLSOLFDWLYH�RGGV�RI � initiating breastfeeding better than a model with socio-demographic variables alone, G2 (5, 798) = 3622.5. Partici- pants who indicated their pregnancies were unwanted or too soon breastfed for a shorter duration in comparison to those who said it was the right time; these effects remained after controlling for socio-demographic characteristics. Feelings of happiness about the pregnancy did not alter the odds of initiation or affect the duration of breastfeeding. Keywords: breastfeeding initiation, breastfeeding duration, pregnancy intentions, happiness about the pregnancy While much research has investigated unintended pregnancies and their implications for maternal and infant health, less attention has been placed on the relationship between health behaviors and women’s thoughts and feelings about being pregnant (Blake et al., 2007). One particularly important health behavior that is worthy of further investigation due to its prov- HQ�SK\VLRORJLFDO�DQG�SV\FKRORJLFDO�KHDOWK�EHQHÀWV�WR� both mothers and infants is initiation and duration of breastfeeding (Davis, Savitz, & Graubard, 1988; Lawrence, 2000; Mortensen, Michaelsen, Sanders, & Reinsich, 2002; Newcomb et al., 1994; Visness, Kennedy, & Ramos, 1997). The physiological bene- ÀWV�RI �KXPDQ�EUHDVW�PLON�DUH�QXPHURXV��DQG�LQFOXGH� anti-infective, immunologic, and allergy protective properties; infants who are breastfed demonstrate a reduction in infections and lower rates of child on- set malignancies, such as acute leukemia (Davis et al., 1988; Lawrence, 2000). Research has also demonstrat- HG�WKH�KHDOWK�EHQHÀWV�RI �EUHDVWIHHGLQJ�IRU�PRWKHUV�� including an association between breast stimulation and the control of excessive postpartum hemorrhag- ing, as well as a decreased risk of premenopausal breast cancer (Newcomb et al., 1994; Visness et al., 1997). Breastfeeding has also been shown to have advantageous psychological associations, includ- ing long range developmental and intellectual out- comes for the breastfed infant, and a unique moth- er-infant attachment bond (Mortensen et al., 2002). Despite these optimal health outcomes for both infants and mothers, fewer than one third of US in- fants are breastfed exclusively for the recommended duration of more than 6 months, followed by contin- ued partial breastfeeding for at least a year (“Amer- ican Academy,” 2013; Chen, Johnson, & Rosenthal, 2012). Successful breastfeeding depends on multiple factors related to the mother, infant, and the envi- ronment, and there are a variety of barriers which inhibit initiation and continuation of breastfeeding, such as the mother returning to work, lack of access to breast pumps, sore nipples and pain, and access to free formula (Haughton, Gregorio, & Pérez-Escamil- la, 2010). We recognize that not all women are phys- ically able to breastfeed and some may experience GLIÀFXOW\� LQ� ERWK� LQLWLDWLQJ� EUHDVWIHHGLQJ� DQG� FRQ- tinuing when it proves challenging (Kelleher, 2012). Socio-Demographic Factors and Breastfeeding � 'HVSLWH� WKH� ZHOO�NQRZQ� EHQHÀWV� RI � EUHDVW- feeding in the existing literature, disparities exist in breastfeeding behavior among various population groups (Centers for Disease Control and Preven- tion [CDC], 2011). A variety of factors are associ- ated with a mother’s decision or ability to initiate and continue breastfeeding, including personal char- acteristics such as age, race/ethnicity, education, socioeconomic status, intrapersonal characteris- tics, sources of support, childcare status of the in- fant, and available breastfeeding interventions (Celi, Rich-Edwards, Richardson, Kleinman, & Gillman, 2005; Dennis, 2002; Li, Darling, Maurice, Barker, & Grummer-Strawn, 2005). Research shows that across race, white non-Hispanic mothers breastfeed at a higher rate than black mothers (Smith-Gagen, 48 Hollen, Walker, Cook, & Yang, 2014). Additionally, studies have shown that black women have a short- er duration of breastfeeding after 6 months (27.9%) compared with white (45.1%) and Hispanic (46%) mothers. By 12 months, only 12.9 % of black moth- ers still breastfeed in comparison with white (23.6%) and Hispanic (12%) mothers (CDC, 2011). Further- more, breastfeeding mothers tend to be older with higher levels of education (Jones, Kogan, Singh, Dee, & Grummer-Strawn, 2011; Taylor & Cabral, 2002). Socio-economic status tends to be higher for those who initiate breastfeeding (Piper & Parks, 1996; Taylor & Cabral, 2002). Among low-income pregnant and postpartum women who receive nutrition and health services through the Women, Infant, and Chil- dren (WIC) program, the largest supplemental food program in the United States, only half of participants report breastfeeding their youngest child (Haughton et al., 2010). Research demonstrates that participation in WIC may actually inhibit women from breastfeed- ing. According to Ryan and Zhou (2006), for mothers with infants at least 6 months of age, WIC status was actually the strongest negative determinant of breast- feeding; mothers who did not participate in WIC (but were eligible) were twice as likely to breastfeed for at least 6 months compared to those who were en- rolled in the program (Ryan & Zhou, 2006). Thuli- er and Mercer (2009) suggest that free samples and distribution of infant formula through WIC inhibits women from initiating and continuing breastfeeding. The role of WIC on low-income women’s breast- feeding behavior cannot be ignored when consider- ing the reasons why women of lower socio-econom- ic status tend to initiate breastfeeding less often and have trouble maintaining it for 6 months or more. Pregnancy Intention and Breastfeeding Since about half of the pregnancies in the United States are unintended, pregnancy intentions, includ- ing mistimed or unwanted pregnancies, have been studied in relation to health behaviors, including breastfeeding (Finer & Zolna, 2014). Several studies have linked unintended pregnancies to lower rates of breastfeeding (Cheng, Schwartz, Douglas, & Horon, 2009; Taylor et al., 2008). Unwanted pregnancy, in particular, is associated with less initiation of breast- feeding (Joyce, Kaestner, & Korenman, 2000a). Fur- thermore, research has demonstrated that when both parents intend to conceive, children are the most likely to be breastfed, while discrepancies between the parents regarding pregnancy intentions result in ORZHU�UDWHV�RI �EUHDVWIHHGLQJ��6SHFLÀFDOO\��FKLOGUHQ�DUH� breastfed at lower rates when the mother does not in- tend to conceive but the father does, as compared to when the mother does intend to conceive but the fa- ther does not (Korenman, Kaestner, & Joyce, 2002). Inconsistent with previous research, Cheng and col- leagues (2009) found that mothers with mistimed and unwanted births were as likely as mothers with intend- ed births to initiate breastfeeding; however, mothers with mistimed and unwanted pregnancies were less likely to continue breastfeeding for 8 weeks or more. Cheng et al. (2009) hypothesized that continuing to breastfeed when it becomes challenging may be more GLIÀFXOW�IRU�PRWKHUV�ZKR�EHFRPH�SUHJQDQW�XQLQWHQ- tionally compared to those with intended pregnancies. Psychological Factors and Breastfeeding Inten- tion vs Outcome Furthermore, researchers have investigated the effects of other psychological factors on breast- feeding intentions and outcomes. Noel-Weiss, Rupp, Cragg, Bassett, and Woodend (2006) found that the PRWKHU·V�FRQÀGHQFH�LQ�KHU�DELOLW\�WR�EUHDVWIHHG�ZDV� positively correlated with breastfeeding duration. Re- search by Fairlie, Gillman, and Rich-Edwards (2009) demonstrated that women with high pregnancy-re- lated anxiety and prenatal depressive symptoms are less likely to plan to breastfeed; however, these symptoms were not associated with failure to initiate breastfeeding in this study. Most research, though, indicates that prenatal breastfeeding intentions pre- dict postnatal breastfeeding behaviors (Forster, Mc- Lachlan, & Lumley, 2006; Lawson & Tulloch, 1995). Previous research has also demonstrated that women intending to breastfeed have higher scores on ma- ternal-fetal attachment than those not intending to breastfeed (Foster, Slade, & Wilson, 1996). Other research has found a modest relationship between prenatal attachment and exclusive breastfeeding in- tentions, in that high maternal-fetal attachment is correlated with the intention to breastfeed exclu- WALLACE, CHAD, RODRIGUEZ 49 sively; however, this relationship did not hold after controlling for other relevant covariates such as per- ceived support and employment status (Lear, 2013). Pregnancy-Related Happiness and Health Out- comes While previous research has investigated feel- ings of pregnancy-related anxiety, pregnancy inten- tion, and maternal-fetal attachment, some research has investigated mothers’ perceived well-being and happiness during pregnancy. Su (2012) demonstrated that mothers with unintended births experienced de- clines in overall happiness and well-being postpartum (rather than just pregnancy-related happiness) relative to childless women. Sable, Washington, Schwartz, & -RUJHQVRQ��������VSHFLÀFDOO\�H[SORUHG�ZRPHQ·V�IHHO- ings of happiness about having a baby. Pregnancy-re- lated happiness was positively correlated with social support and negatively correlated with loneliness and family relationship problems (Sable et al., 2007). Further research has investigated the link be- tween pregnancy intentions and happiness about being pregnant, and the associated risk for adverse effects for the mother and the infant (Blake et al., ������� $� VLJQLÀFDQWO\� KLJKHU� SURSRUWLRQ� RI � ZRP- en who were unhappy about being pregnant (in comparison to those who were moderately hap- py or happy) reported having smoked cigarettes in the past week, having been depressed in the past month, having experienced intimate partner vio- lence in the past year, and having drunk alcohol or used drugs during the pregnancy (Blake et al., 2007). Additionally, a woman’s happiness about be- ing pregnant was more strongly associated with the above-mentioned behavioral and psychosocial risk factors than pregnancy intention, indicating that pregnancy-related happiness may be more related to prenatal care initiation/utilization and pregnancy out- comes than pregnancy intentions (Blake et al., 2007). Investigation of Pregnancy-Related Happiness and Breastfeeding � %DVHG�RQ�WKH�ÀQGLQJV�GHPRQVWUDWLQJ�WKH�UHODWLRQ- ship between pregnancy-related happiness and pre- natal risk factors, further research into breastfeeding, which is one major component of maternal health behavior, may also reveal an association with pregnan- cy-related happiness. Even though Blake et al. (2007) found that pregnancy intentions and pregnancy-re- lated happiness are strongly correlated, Sable (1999) states that measuring pregnancy intentions may not EH�D�VXIÀFLHQW�SUR[\�IRU�SUHJQDQF\�DWWLWXGHV��&RQVLG- ering that pregnancy intentions and pregnancy-related happiness may differentially impact maternal health behaviors, research needs to expand on the role of maternal psychological factors. Although various so- FLR�GHPRJUDSKLF�IDFWRUV�KDYH�EHHQ�LGHQWLÀHG�DV�SUH- dictors of breastfeeding practices, the added impact of psychological factors (feelings of happiness and wantedness) remains limited. Given the public health costs of prenatal and postnatal programs which pro- YLGH� HGXFDWLRQ� DERXW� WKH� EHQHÀWV� RI � EUHDVWIHHGLQJ� behavior, it would be advantageous to better under- stand such psychological factors, rather than simply socio-demographic factors alone. In addition, a bet- ter understanding of these psychological factors may help explain low rates of breastfeeding by elucidating WKH�UHDVRQV�VRPH�ZRPHQ�KDYH�GLIÀFXOW\�ZLWK�LQLWLD- tion and prolonged duration, including experiences of physical discomfort and pain (Kelleher, 2012). Present Research � :H� VHHN� WR�ÀOO� WKLV� OLWHUDWXUH� JDS� E\� LQYHVWLJDW- ing the relationship between wantedness and feel- ings of happiness about the pregnancy in relation to breastfeeding initiation and duration. Our re- search question is: do wantedness and feelings of happiness about the pregnancy explain initiation and duration of breastfeeding when controlling for VRFLR�GHPRJUDSKLF� IDFWRUV"� &RQVLVWHQW� ZLWK� SUHYL- ous research, we hypothesize that mothers with un- intended pregnancies and feelings of unhappiness about the pregnancy are less likely to initiate breast- feeding and more likely to breastfeed for a shorter duration, compared with mothers whose pregnancy was intended and who are happy about their preg- nancy. In addition, we hypothesize that demograph- ic characteristics, such as education, race, age, and SRYHUW\� OHYHO�ZLOO� VLJQLÀFDQWO\� SUHGLFW� LQLWLDWLRQ� DQG� duration of breastfeeding, as is consistent with pre- vious research. Overall, however, we hypothesize that feelings of happiness and wantedness of preg- PREGNANCY INTENTIONS AND HAPPINESS 50 nancy will better explain variation in breastfeeding outcomes than socio-demographic factors alone. Method Data In order to investigate these questions, we used data from the National Survey of Family Growth (NSFG), which is a large periodic and popula- tion-based survey (NSFG, 2002). The NSFG has been conducted by the National Center for Health 6WDWLVWLFV� VLQFH� ������ ,W� ZDV� FRQGXFWHG� ÀYH� WLPHV� with a national sample of women periodically be- tween 1973 and 1995. A sixth periodic survey was conducted in 2002, which included both men and women. We used data from the 2002 wave that in- terviewed 12,571 men and women on factors affect- ing birth rates, pregnancy rates and women’s repro- ductive health. It contains information on 13,593 pregnancies and 9,148 births from 5,033 women. From the sample of 5,033 women, we created a sub-sample of women whose pregnancy resulted in a live birth, and who answered the questions based on WKHLU�ÀUVW�FKLOG��ZKHWKHU�LW�ZDV�DQ�RQO\�FKLOG�RU�QRW��,Q� order to create this sub-sample, we excluded individu- als who had lost a pregnancy, aborted their pregnancy, or were currently pregnant. The sub-sample includes 4,413 participants. It is worth noting then that the sam- ple only includes responses of those women referring WR�WKHLU�ÀUVW�FKLOG��:KLOH�WKLV�PD\�SRWHQWLDOO\�OLPLW�WKH� generalizability of the results, this exclusion criteri- on is based on previous research which suggests that mothers are likely to choose the same feeding meth- od for each of their children based on how they fed WKHLU�ÀUVW�FKLOG��UHJDUGOHVV�RI �WKH�QXPEHU�RI �FKLOGUHQ� they have (Taylor et al., 2008). Considerations of this assumption will be made in the limitations section. Measures Breastfeeding initiation/duration��7KH�ÀUVW�GH- pendent variable selected was initiation of breastfeed- ing. Participants were asked whether they breastfed their child at all, so this is a binary outcome measure for either yes or no. The second dependent variable selected was a continuous outcome measure for du- ration of breastfeeding measured in weeks. To check for non-normality we constructed a PP-plot and his- togram, which compared the cumulative probabili- ties of the normal distribution with the distribution of residuals. Our PP-plot showed that our data was VOLJKWO\�VNHZHG�WR�WKH�ULJKW��7R�À[�WKLV�QRQ�QRUPDO- ity problem, we transformed our Y dependent vari- able (duration of breastfeeding weeks) by taking the natural log and running our regression model again. Non-normality was no longer a problem after trans- forming the dependent variable. All further analy- ses with this continuous dependent variable were conducted with logged duration of breastfeeding in weeks. Those who breastfed their child for less than one week were coded as 0.5 to indicate an average between 0 weeks (no breastfeeding) and 1 week. Participants who reported that they were still breast- feeding their child were excluded from analyses be- cause duration of time was impossible to ascertain. Socio-demographic characteristics. Demo- graphic variables included age, poverty, education, and race. Both age and poverty level were mea- sured as continuous variables. Poverty level was the percentage of the poverty level for the year 2001; participants who exceed 500 percent of the pov- erty level were grouped together. Education was also a continuous variable, which measured years of completed schooling. Race was a categorical variable and included Hispanic, non-Hispanic black, non-Hispanic white, and non-Hispanic other. Hispanic was used as the reference category in all analyses. Intentions and feelings about pregnancy. In- tention variables included wantedness of pregnancy and feelings of happiness about the pregnancy. Want- edness of the pregnancy was scored as a categorical variable, and included the response options of too soon, right time, unwanted, indifferent, and don’t know. Participants whose answer was don’t know were coded as missing. Right time was used as the reference category in all analyses. Feelings of happi- ness were measured on a continuous scale where 0 PHDQV�WKH�SDUWLFLSDQW�ZDV�YHU\�XQKDSS\�WR�ÀQG�RXW� that she was pregnant, and 10 means the participant ZDV� YHU\� KDSS\� WR� ÀQG� RXW� WKDW� VKH�ZDV� SUHJQDQW� Statistical Analysis Initiation of breastfeeding. Logistic regres- sion was conducted to explore the relationship be- WALLACE, CHAD, RODRIGUEZ 51 tween the dependent variable (initiation of breast- feeding) and the independent variables. Omnibus tests were used to determine whether socio-demo- graphic factors (age, poverty, education, and race) predicted breastfeeding initiation, and whether so- cio-demographic factors with intentions/happiness variables predicted breastfeeding initiation. A log likelihood ratio test was used to investigate wheth- er a model with socio-demographic factors (age, poverty, education, and race) and wantedness/hap- piness about the pregnancy explains an increase in the multiplicative odds of breastfeeding better than a model with only socio-demographic factors. Indi- YLGXDO� FRHIÀFLHQWV� UHVXOWLQJ� IURP� WKH� VHFRQG� FRP- parison were examined to determine which factors DUH� VLJQLÀFDQW� LQ� SUHGLFWLQJ� EUHDVWIHHGLQJ� LQLWLDWLRQ� Duration of breastfeeding. Ordinary least squares (OLS) regression was used to examine the relationship between the dependent variables and the independent variable (the logged duration of breastfeeding in weeks). Since little research exists establishing an association between breastfeeding behaviors and happiness about the pregnancy, we decided to conduct two incremental F-tests. The ÀUVW� LQFUHPHQWDO� )�WHVW� FRPSDUHG�ZKHWKHU� D�PRGHO� with socio-demographic factors (age, poverty, edu- cation, and race) and wantedness/happiness about WKH� SUHJQDQF\� H[SODLQHG� D� VLJQLÀFDQW� SURSRUWLRQ� of the variation in logged duration of breastfeed- ing better than a model with socio-demographic variables alone. The second incremental F-test was structured the same, except that the second model included socio-demographic factors and only want- HGQHVV� YDULDEOHV�� ,Q� DGGLWLRQ�� LQGLYLGXDO� FRHIÀFLHQWV� resulting from the second comparison were ex- amined to determine which variables were signif- icant in predicting logged breastfeeding duration. Results Sample Characteristics Table 1 describes the characteristics of our sam- ple. Women who initiated breastfeeding, regardless of the duration, comprised 60.6% of the sample, while 39.4% did not breastfeed at all. The average duration of breastfeeding in our sample was 16.37 weeks (SD = 25.60) with a minimum duration of 0 and a maxi- mum of 208. This indicates that the average length of breastfeeding was about 4 months. Participants were 48.7% white, 22.5% black, 24.5% Hispanic, and 4.3% non-Hispanic other. The average percentage of the poverty level was 228.58, which for a family of 4 was about $40,000 per year in 2001 (“Annual Update,” 2001). Average number of years of schooling was 12.78, and the mean age was about 23 years old with a standard deviation of about 0.5 years. The average for feelings of happiness about the pregnancy was 7.97 out of 10. Almost half of the participants re- ported that their pregnancy occurred at the right time (47.5%), 7.8% reported they wanted it later, 31.5% in- dicated it was too soon or mistimed, 1.4% didn’t care, and 11.8% reported that the pregnancy was unwanted. Predictors of Breastfeeding Initiation Table 2 shows results of the logistic regression analyses. Results revealed that the socio-demograph- ic factors (age, race, education, and SES) explain an increase in the multiplicative odds of breastfeeding initiation,�ƶ2 (5, 3643) = 442.75, p < .001. The model predicts 67.6% of breastfeeding initiation correctly. Nagelkerke R2 = .151, meaning that 15.1% of the variation in breastfeeding initiation is explained by socio-demographic variables. The full model with socio-demographic and intention/happiness factors DOVR�VLJQLÀFDQWO\�H[SODLQV�DQ�LQFUHDVH�LQ�WKH�PXOWLSOL- cative odds of breastfeeding initiation, ƶ2 (11, 798) = 80.48, p < .001. The model predicts 72.5% of breast- feeding initiation correctly. Nagelkerke R2 = .127, indicating that 12.7% of the variation in breastfeed- ing initiation is explained by socio-demographic and intention/happiness factors. Results from the likeli- hood ratio test indicate that the full model, including socio-demographic and intentions/happiness, ex- plains an increase in the multiplicative odds of initiat- ing breastfeeding better than a model with socio-de- mographic variables alone, G2 (5, 798) = 3622.5. � 8SRQ�DQDO\]LQJ�WKH�FRHIÀFLHQWV�LQ�WKH�IXOO�PRGHO�� the odds of initiating breastfeeding decrease by .1% for an increase in one year of age, holding constant all other factors, p = .024, (exp(b) = 0.999). The odds of initiating breastfeeding decrease by 15% for one more year of schooling, holding constant all other PREGNANCY INTENTIONS AND HAPPINESS 52 � � Table 1��Sample Characteristics of Women from their First, Live Birth � ____________________________________________________________________________� Variable Valid Missing Mean Std. Dev Percent� � Breastfeeding Initiation 3738 675 - - 60.6� Breastfeeding Duration 3403 1,010 16.37 25.60� Education 4413 0 12.78 2.55 -� Poverty 4413 0 228.58 151.16 -� Age at Pregnancy Outcome 4413 0 2308.4 529.14 -� Hispanic 1080 - - - 24.5� White 2150 - - - 48.7� Black 995 - - - 22.5� Other 188 - - - 4.3� Happiness 827 3586 7.97 2.94 - � Wantedness: Right Time 2094 - - - 47.5 Later 343 - - - 7.8� Too Soon, Mistimed 1390 - - - 31.50� Didn’t Care/Indifferent 61 - - - 1.4� Unwanted 519 - - - 11.8� Don’t Know 6 - - - .1� � WALLACE, CHAD, RODRIGUEZ 53 � � Table 2��Logistic Regression Analyses Predicting Breastfeeding Initiation from Socio- Demographic Characteristics and Wantedness of Pregnancy Variables� � Variable� � Model 1� � � � Model 2� � � Coefficient� SE� Exp(b)� Coefficient� SE� Exp(b)� Constant � 2.34**� .22� 10.37� 1.297*� .65� 3.66� Age at Pregnancy � -.001**� .0001� .99� -.0005*� .0002� .99� Poverty Level Income� -.001**� .0003� .99� -.001� .0006� .99� Education� -.15**� .02� .86� -.15**� .04� .86� White� .71**� .10� 2.03� .67**� .22� 1.94� Black� 1.32**� .11� 3.74� .97**� .26� 2.63� Other� .24� .20� 1.27� .47� .43� 1.60� Later� -� -� -� .10� .34� 1.11� Too Soon, Mistimed� -� -� -� .15� .22� 1.17� Indifferent� -� -� -� .77� .84� 2.16� Unwanted� -� -� -� .53� .33� 1.70� Happiness� -� -� -� .06� .04� .99� Model Chi-Square [df]� 442.751 [6]� -� -� 80.484 [11]� -� -� Block Chi-Square [df]� -� -� -� 3622.5 [5]� -� -� % Correct Predictions� 67.6� -� -� 72.5� -� -� Nagelkerke R2� .151� -� -� .134� -� -� * p < .05 ** p <.01� � � � � � � � � PREGNANCY INTENTIONS AND HAPPINESS 54 factors, p < .05, (exp(b) = .85). The odds of initiating breastfeeding increase by 94.4% for white mothers compared to Hispanic mothers, holding constant all other variables, p = .003, (exp(b) = 1.94). The odds of initiating breastfeeding increase by 163.4% for black mothers in comparison to Hispanic mothers, holding constant all other variables, p < .001, (exp(b) = 2.63). Socio-economic status, non-Hispanic other, feelings of happiness, and all of the intention vari- ables (i.e. later, too soon, indifferent, unwanted) were not VLJQLÀFDQW� IDFWRUV� LQ� WKH� LQLWLDWLRQ�RI �EUHDVWIHHGLQJ� Predictors of Breastfeeding Duration Table 3 describes the results of the OLS regres- sion on breastfeeding duration. In order to determine WKH�LQÁXHQFH�RI �VRFLR�GHPRJUDSKLFV�DQG�SUHJQDQF\� intentions/happiness about the pregnancy on breast- feeding behavior, an omnibus analysis of variance test was conducted. It showed that socio-demographics DQG�LQWHQWLRQV�KDSSLQHVV�ZHUH�VLJQLÀFDQW�LQ�H[SODLQ- ing the logged duration of breastfeeding, F(11, 650) = 3.53, p < .001. An incremental F-test, however, did not demonstrate that a model with socio-demograph- ic factors and wantedness/happiness variables better explains breastfeeding duration than a model with socio-demographics alone, )¨ (5, 435) = 1.54, p > .05. Therefore, we sought to evaluate a model that did not include happiness to be pregnant, only socio-de- mographic and intention variables. This decision ZDV� ODUJHO\� LQÁXHQFHG� E\� SUHYLRXV� UHVHDUFK� ZKLFK� has demonstrated that pregnancy intentions predict breastfeeding, and the lack of research establishing a link between feelings of happiness and breastfeeding outcomes (Dye, Wojtowycz, Aubry, Quade, & Kil- burn, 1996; Li et al., 2007; Taylor & Cabral, 2002). � ,Q� RUGHU� WR� GHWHUPLQH� WKH� LQÁXHQFH� RI � VR- cio-demographic and pregnancy intention variables on duration of breastfeeding, an omnibus analy- sis of variance test was conducted. It revealed that socio-demographics and wantedness of pregnan- F\� ZHUH� VLJQLÀFDQW� LQ� H[SODLQLQJ� WKH� ORJJHG� GX- ration of breastfeeding, F(10, 2163) = 4.70, p < .001 (Table 3). An incremental F-test then demon- strated that a model with socio-demographic and intention factors better explains breastfeeding du- ration than socio-demographic factors alone, )¨ (4, 2163) = 4.44, p = .001. It is important to note that while a model including socio-demographic factors and wantedness of pregnancy variables bet- ter explains breastfeeding duration, socio-demo- graphic factors only explain 1.1% of the variation in breastfeeding, and wantedness explains .6% more variation in outcomes, adjusted R2 = .017, p = .001. � $QDO\VLV� RI � WKH� FRHIÀFLHQWV� UHYHDOV� WKDW� DV� DJH� increases by one year, duration of breastfeeding in- creases, holding constant education, poverty, race, and wantedness of pregnancy; however this does QRW� UHDFK� VWDWLVWLFDO� VLJQLÀFDQFH�� ơ� = .00004, p > .05 (Table 3). As income level increases, duration of breastfeeding increases, holding constant all other variables; yet this increase is not statistically signif- icant, ơ = .00003, p > .05. As years of completed schooling increases by one year, duration of breast- feeding increases by .02 weeks, holding constant all other variables; however, this increase is not statis- WLFDOO\� VLJQLÀFDQW�� p > .05. The average difference in breastfeeding between Hispanics and whites is .03, holding constant all other variables, p > .05. This means that white women breastfed a longer duration than Hispanic women controlling for other factors, EXW�WKLV�GLIIHUHQFH�LV�QRW�VWDWLVWLFDOO\�VLJQLÀFDQW��7KH� average difference in breastfeeding between Hispan- ics and blacks is -.08, holding constant all other vari- ables, which means that blacks breastfed a shorter duration in comparison to Hispanics, holding constant all the other variables, but this difference is not sta- WLVWLFDOO\�VLJQLÀFDQW��p > .05. The average difference between Hispanics and individuals from other ethnic backgrounds, holding constant all other variables, is .18, indicating that individuals from other ethnic backgrounds breastfed a longer duration than Hispan- ics, controlling for all other variables, but this differ- HQFH�GRHV�QRW� UHDFK�VWDWLVWLFDO� VLJQLÀFDQFH��p > .05. In terms of wantedness of pregnancy, the av- erage difference between those whose pregnan- cy came later than desired compared to those who said it was the right time is -.12, meaning that those whose pregnancy came later than desired breastfed a shorter duration than those who said it was the right time�� WKLV�GLIIHUHQFH�ZDV�QRW� VLJQLÀFDQW� WKRXJK��p > .05. The average difference between those who said their pregnancy was too soon compared to those who WALLACE, CHAD, RODRIGUEZ 55 � � � � � � � � � � � Table 3��Multiple Regression Analyses Predicting Breastfeeding Duration from Socio- Demographic Characteristics and Wantedness of Pregnancy Variables� � � Model 1� Model 2� Variable� Coefficient� SE� Coefficient� SE� 95% CI� � � � � � � Lower � Upper � Constant � 2.16**� 1.49� 2.43**� .17� 2.101� 2.758� Age at Pregnancy Outcome � .000*� .00� .000� .00� .000� .000� Poverty Level Income� .000� .00� .02� .00� .000� .000� Education� .02� .01� .03� .01� -.001� .05� White� .03� .07� -.08� .07� -.09� .16� Black� -.12� .08� -.08� .09� -.25� .08� Other� .21� .13� .18� .13� -.08� .44� Later� -� -� -1.2� .09� -.31� .06� Too Soon, Mistimed� -� -� -.18**� .07� -.31� -.05� Indifferent� -� -� -42� .24� -.89� .04� Unwanted� -� -� -.34**� .09� -.53� -.15� Adjusted R2� .01� -� 0.17� -� -� -� F� 4.85� -� 4.70� -� -� -� � R2 - - .01 - - - � F - - 4.44 - - - * p < .05 ** p <.01 � PREGNANCY INTENTIONS AND HAPPINESS 56 said it was the right time was -.18, holding constant all other variables; this indicates that those who felt their pregnancy was too soon breastfed a shorter du- ration than those whose pregnancy was at the right time��WKLV�GLIIHUHQFH�LV�VWDWLVWLFDOO\�VLJQLÀFDQW��p = .006, 95% CI [-.31, -.05]. The average difference between those who said they felt indifferent about their preg- nancy timing was -.42, holding constant all other variables. This indicates that those who felt indifferent about whether they wanted their pregnancy breast- fed a shorter time than those who felt it was the right time��EXW�WKLV�GLIIHUHQFH�ZDV�QRW�VLJQLÀFDQW��S�!� .05. The average difference between those who said their pregnancy was unwanted and those whose was at the right time, holding constant all other variables, was -.34; therefore, women whose pregnancy was un- wanted breastfed a shorter duration than those who indicated it was at the right time.� 7KLV� FRHIÀFLHQW� LV� VWDWLVWLFDOO\�VLJQLÀFDQW��p < .001, 95% CI [-.53, -.15]. � ,Q� VXP�� WKH� RQO\� VWDWLVWLFDOO\� VLJQLÀFDQW� FRHI- ÀFLHQWV� DUH� too soon and unwanted; women who en- dorsed their pregnancies as either too soon or unwant- ed breastfed for a shorter duration than those who said theirs was at the right time. As mentioned pre- viously, preliminary analyses revealed that the data was non-normal, so we transformed our depen- dent variable (duration of breastfeeding in weeks) by taking the natural log. Other assumption checks were performed to identify any problems relating to homoskedasticity, linearity, outliers, leverage, or LQÁXHQFH�SRLQWV�� DQDO\VHV� UHYHDOHG� WKDW� UXQQLQJ� WKH� incremental-F� WHVW� ZLWKRXW� LQÁXHQFH� SRLQWV� \LHOG- ed a negligible change in outcome variation and ]HUR� GLUHFWLRQDO� FKDQJH� LQ� WKH� YDULDEOH� FRHIÀFLHQWV� Discussion Overall, our results replicate previous research which indicates that breastfeeding initiation is explained by socio-demographic variables (age, education, SES, and race); however, socio-demographic and inten- tion/happiness factors better explain the initiation of breastfeeding than socio-demographics alone. The full model is able to predict breastfeeding initiation with 72.5% accuracy and explains 12.7% of the vari- ation in initiation of breastfeeding. Even though our VDPSOH�RQO\�LQFOXGHG�GDWD�IURP�PRWKHUV·�ÀUVW�ELUWKV�� these results might possibly generalize to breastfeed- ing outcomes at any birth because previous research has demonstrated that mothers are likely to choose the same feeding method for each of their children EDVHG�RQ�KRZ�WKH\�EUHDVWIHG�WKHLU�ÀUVW�FKLOG��7D\ORU� et al., 2008). However, in consideration of the fact that pregnancy intentions, pregnancy-related anxiety and happiness, social support, and a multitude of other factors can vary from birth to birth, it seems likely that mothers may have different initiation/du- ration of breastfeeding from birth to birth. There- fore, these results should be interpreted with cau- tion, and may be generalizable only to new mothers. In line with previous research, we found that older mothers are more likely to initiate breastfeed- ing than younger mothers. Inconsistent with pre- vious research, in our sample, individuals with less education are more likely to initiate breastfeeding than those with more years of schooling (Jones et al., 2011; Taylor & Cabral, 2002). Although most studies have demonstrated that more education is associated with an increased likelihood to initiate breastfeeding, the interaction between race/ethnici- ty and education as it relates to initiation of breast- feeding remains to be determined (Chin, Meyers, & Magnus, 2008). It is possible that such an inter- action went undetected in our sample, resulting LQ� ÀQGLQJV� WKDW� DUH� LQFRQVLVWHQW�ZLWK� WKH� OLWHUDWXUH�� Consistent with recent studies, we found that white mothers are more likely to initiate breastfeed- ing than Hispanic mothers (CDC, 2011). Our model suggests that black mothers are more likely to initiate breastfeeding in comparison to Hispanic mothers, though this is not the typical trend in the literature (CDC, 2011). There is literature to suggest that ac- culturation status affects breastfeeding behaviors for Hispanic women, such that more acculturated women tend to have lower initiation/duration of breastfeeding. Therefore, varying degrees of accul- turation among the Hispanic women in our sample (perhaps more acculturated women) may contrib- ute to why our results indicated that black women were more likely to initiate breastfeeding (Ahluwalia, D’Angelo, Morrow, & McDonald, 2012). Our hy- pothesis that intentions/happiness factors help to predict initiation of breastfeeding better than so- WALLACE, CHAD, RODRIGUEZ 57 FLR�GHPRJUDSKLF�IDFWRUV�DORQH�ZDV�FRQÀUPHG��KRZ- ever, these results should be viewed with caution because pregnancy intentions and happiness about WKH�SUHJQDQF\�GR�QRW�VLJQLÀFDQWO\�DOWHU�WKH�RGGV�RI � initiation of breastfeeding as individual covariates. Furthermore, our results show that breastfeed- ing duration is explained by socio-demographic fac- tors and wantedness of the pregnancy. Our original hypothesis was not supported because feelings of KDSSLQHVV�GLG�QRW�KHOS� WR�H[SODLQ�D� VLJQLÀFDQW�SUR- portion of the variation in duration of breastfeeding, but a model with just wantedness of the pregnancy and socio-demographic factors does indeed explain a higher proportion of the variance in breastfeeding duration than just socio-demographic factors alone. This result should be interpreted with caution, how- ever, because wantedness of the pregnancy only ex- plains .6% more of the variation in breastfeeding outcomes than socio-demographic factors alone. Consistent with previous research, as age, educa- tion, and income levels increase, duration of breast- feeding increases (Jones et al., 2011; Piper & Parks, 1996; Taylor & Cabral, 2002); these results should be viewed with care because they do not reach statistical VLJQLÀFDQFH��5HVXOWV�LQGLFDWHG�WKDW�ERWK�ZKLWH�ZRPHQ� and women from other racial/ethnic minorities (who are not white, black, or Hispanic) breastfed their child for a longer time than did Hispanic women, and black women breastfed a shorter duration than +LVSDQLF�ZRPHQ��$OWKRXJK�WKHVH�FRHIÀFLHQWV�DUH�QRW� VWDWLVWLFDOO\� VLJQLÀFDQW�� SUHYLRXV� UHVHDUFK� VXJJHVWV� that white women breastfeed for longer durations than minority women, though the results tend to be mixed (Chen et al., 2012; Jones et al., 2011; Thulier & Mercer, 2009). Other research suggests that there are no differences in breastfeeding duration between dif- ferent racial/ethnic groups, except that black women tend to breastfeed less often than women of other racial/ethnic categories (Thulier & Mercer, 2009). Women with unwanted pregnancies or pregnan- cies that occurred too soon breastfed for shorter durations than those who indicated theirs was at the right time; those who felt their pregnancies came later than desired or were indifferent about it also breastfed IRU�D�VKRUWHU�WLPH��EXW�WKHVH�FRHIÀFLHQWV�GLG�QRW�UHDFK� VWDWLVWLFDO� VLJQLÀFDQFH��2YHUDOO� WKRXJK�� WKHVH� UHVXOWV� replicate previous research which supports that unin- tended, unwanted, or mistimed pregnancies are asso- ciated with a shorter duration of breastfeeding (Cheng et al., 2009; Taylor et al., 2008). Results did not sup- port the hypothesis that feelings of happiness would explain breastfeeding outcomes, in that more positive feelings would yield a longer duration of breastfeed- ing, but a logistic regression revealed that feelings of happiness taken together with intention variables and socio-demographics did predict higher rates of breastfeeding initiation above just socio-demograph- ics. In sum, women’s intentions and desires about her pregnancy do help to explain breastfeeding initiation and duration better than socio-demographic factors DORQH�� EXW� WKHVH� UHVXOWV� QHHG� IXUWKHU� FODULÀFDWLRQ� Women with unintended pregnancies and nega- tive pregnancy-related feelings may have a challenging time initiating and sustaining breastfeeding for the recommended length of time. Unintended pregnan- cies and negative feelings surrounding the pregnancy PD\�PDNH�LW�PRUH�GLIÀFXOW�IRU�WKH�PRWKHU�WR�IDQWD- size about her infant and create a psychological bond with the infant that continues after birth. For these women, there may be a denial of the reality of the pregnancy and a lack of recognition of the infant, its needs, and the role of motherhood, possibly translat- ing to physiological and psychological disengagement in maternal health behaviors, including breastfeeding. Future research should explore these hypotheses. Limitations One limitation is that our sample only included ZRPHQ�ZKR�ZHUH�DVNHG�DERXW�WKHLU�ÀUVW�ELUWK��:KLOH� previous research suggests that a mother tends to replicate her breastfeeding behavior with each subse- quent birth (Taylor et al., 2008), feelings of happiness and wantedness of the pregnancy may change from pregnancy to pregnancy based on a variety of factors. Thus, it would be advantageous to examine a model that takes into account multiple births in order to bet- ter elucidate the relationship between breastfeeding behavior and happiness/wantedness of pregnancy. It ZRXOG�EH�EHQHÀFLDO�WR�FRPSDUH�D�ZRPDQ·V�SUHJQDQF\� intentions/feelings of pregnancy-related happiness and breastfeeding initiation/duration for each anal- ogous birth to determine if shifts in intentions/feel- PREGNANCY INTENTIONS AND HAPPINESS 58 ings correspond to shifts in breastfeeding behavior. Another potential confound in our sample is the LQÁXHQFH�RI �D�GLYHUVH�DUUD\�RI �VRFLHWDO�QRUPV�UHJDUG- ing breastfeeding behavior. Many women experience a cultural expectation to breastfeed, which has been equated to being a “good enough mother” (Stea- rns, 2009). Feminist scholars explain that the med- ical and cultural mandate for women to breastfeed places an undue burden on women, especially those who choose not to breastfeed or who cannot breast- feed (Stearns, 2009). Examining women’s internalized expectations of “good mothering” behavior during pregnancy, along with pregnancy-related anxiety and happiness, may help to better explain their prena- tal and postnatal maternal behaviors. Other groups may experience the opposite cultural standards re- garding perinatal behaviors, and are ensconced in communities that embrace formula as the preferred method of feeding (Chin, Meyers, & Magnus, 2008). These variations in societal expectations are critical to take into account when addressing breastfeeding behavior, yet were unaccounted for in our sample. Another limitation is that we did not investi- gate the father’s feelings of happiness or wantedness about the pregnancy. The father’s sentiments about WKH� SUHJQDQF\� PD\� LQÁXHQFH� WKH� PRWKHU� GXULQJ� gestation, as well as after birth (Vaaler et al., 2011). Previous research has demonstrated that disagree- ment between the parents’ intentions predicts the mother’s instability in pregnancy intention, which PD\�LQ�WXUQ�LQÁXHQFH�EUHDVWIHHGLQJ�EHKDYLRU��-R\FH�� Kaestner, & Korenman, 2000b). Further research by Chang, Valliant, and Bomba (2012) supports the strong evidence for a gender gap in breastfeeding knowledge and attitude (men tend to view breast- feeding as less favorable and formula feeding as more favorable in comparison to women). There- fore, the relationship between the mother’s feelings about the birth and breastfeeding may be far more complex, and interact with her partner’s feelings to LQÁXHQFH�WKH�LQLWLDWLRQ�DQG�GXUDWLRQ�RI �EUHDVWIHHGLQJ� One potential confound is that all the data from the NSFG was collected retrospectively. Joyce et al. (2000b) found that in their sample, 30% of the wom- en who reported their pregnancy to be unintended during pregnancy reported that their pregnancy was intended after delivery. Although other research has suggested that prospective assessments are not nec- essarily superior to retrospective reports of unintend- ed fertility (Joyce et al., 2000b), we cannot rule out the possibility that retrospective reporting may have confounded the results of our analyses. In regards to the data, another limitation is that our sample is drawn from the 2002 wave of the NSFG. Although this was the most recent wave of the study available, it was twelve years ago. These results should be in- terpreted with caution since the social climate sur- rounding breastfeeding has inevitably changed since then, making the results of this study slightly dated. Another limitation is the way that pregnancy-re- lated happiness was measured in the NSFG. The interview contained only one question about hap- piness, which asked participants to rate on a scale of one to ten how happy they felt when they found out that they were pregnant (one meaning that they were very unhappy to be pregnant and ten meaning they were very happy to be pregnant). The limited operationalization of the construct of pregnan- cy-related happiness, combined with a possible so- cial desirability bias, may have contributed to our LQVLJQLÀFDQW�ÀQGLQJV�ZKHQ�FRQGXFWLQJ�RXU�DQDO\VHV�� Directions for Future Research Future research should identify potential mod- erators in order to examine why those who indicate their pregnancies are too soon or unwanted tend to breastfeed for shorter durations. We suggest that women with unwanted or mistimed pregnancies may KDYH� D� GLIÀFXOW� WLPH� DWWDFKLQJ� WR� WKHLU� LQIDQWV� ERWK� before and after birth. Since research has demon- strated that women with higher levels of maternal fetal attachment have intentions to breastfeed (Fos- ter et al., 1996), one way in which challenges form- ing an attachment bond may manifest is through breastfeeding, because the mother may lack the mo- WLYDWLRQ�WR�IHHG�QDWXUDOO\�ZKHQ� LW�SURYHV�GLIÀFXOW�RU� when formula feeding is convenient. Another way that unwanted sentiments during pregnancy could manifest after birth is through a lack of warmth and responsiveness to the infant’s needs, which can cause an insecure attachment bond between mother and child (Ainsworth, Blehar, Waters, & Wall, 1978). WALLACE, CHAD, RODRIGUEZ 59 Since higher sensitivity to an infant’s needs is associ- ated with a longer duration of breastfeeding, an ab- sence of breastfeeding may be a sign that the mother is struggling with developing an attachment to her child (Britton, Britton, & Gronwaldt, 2006). Research should investigate whether feelings of unwantedness prenatally translate to a lack of warmth and respon- siveness after birth. Future studies should also explore different ways to measure feelings of happiness, want- edness, and satisfaction with the pregnancy through- out gestation in order to better understand how SV\FKRORJLFDO� IDFWRUV� LQÁXHQFH� SRVWQDWDO� EHKDYLRUV�� Similar to our research, other studies investigat- ing feelings about the pregnancy also use a single item question, rather than a valid scale, to measure preg- nancy-related happiness (Blake et al., 2007; Sable et al., 2007). It would be advantageous for researchers to develop and validate a scale measuring pregnancy-re- lated happiness in order to capture the nuances of the construct better than a simple one-item measure can provide. Another methodological consideration for future research is to use an acculturation scale in addi- tion to the demographic variable of “race”. Previous research indicates that among Hispanic women, those who are less acculturated initiate and continue breast- feeding longer than those who are more acculturated (Ahluwalia, D’Angelo, Morrow, & McDonald, 2012). The addition of acculturation as a socio-demograph- LF� IDFWRU�PD\� LQÁXHQFH� WKH� LQLWLDWLRQ�GXUDWLRQ� UDWHV� of breastfeeding and more accurately capture the experience of individuals who immigrate to another country. It would be interesting to learn how accul- turation interacts with pregnancy intentions/feelings of happiness about the pregnancy in regards to pre- natal and postnatal behaviors for immigrant mothers. Future studies should also make a distinction be- tween wantedness of the pregnancy and pregnancy intentions. In the present study and in current re- search, these two constructs are used interchange- ably; however, these are potentially two separate constructs. For example, a woman who may not have planned to get pregnant may still want the pregnan- cy once she learns that she is pregnant. Conversely, a woman who intends to get pregnant may realize that she does not want to be pregnant once it occurs. 7KH�ÀHOG�RI �UHVHDUFK�VXUURXQGLQJ�SUHQDWDO�DQG�SHUL- natal behaviors may make further advances by clari- fying the distinction between these two constructs. Furthermore, even though we suggest an associa- tion between pregnancy intentions and breastfeeding outcomes, there are several societal, environmental, and health factors that may impact or be associated with breastfeeding. These factors include prenatal care, working before and after birth, the amount of maternity leave taken, acculturation, and smoking before and after birth. Work barriers may limit the amount of time that the mother has with her infant DQG�PDNH� LW� GLIÀFXOW� IRU� KHU� WR� VXFFHVVIXOO\� EUHDVW- feed. In addition, socio-economic status may serve as a moderator between postnatal employment and breastfeeding. We acknowledge these factors and lim- itations, yet push researchers to ask what psychological IDFWRUV�PD\� LQÁXHQFH�RU� LQWHUDFW�ZLWK�EUHDVWIHHGLQJ� behavior as well, since experiences of happiness and prenatal attachment-seeking behaviors by the mother may predict the likelihood of breastfeeding outcomes. Future research should explore whether psycholog- ical factors may help to address why some women do not intend to breastfeed, do not initiate breast- feeding, experience pain and discomfort with breast- IHHGLQJ��DQG�ZK\�WKH\�VWRS�GHVSLWH�LWV�KHDOWK�EHQHÀWV� Conclusion Overall, our results show that when controlling for socio-demographics, pregnancy intentions may LQÁXHQFH� WKH� GXUDWLRQ� DQG� LQLWLDWLRQ� RI � EUHDVWIHHG- ing. Pregnancy intentions and happiness about the pregnancy did contribute to an increase in odds of breastfeeding initiation, but these variables were not VLJQLÀFDQW� SUHGLFWRUV� LQGHSHQGHQWO\�� %UHDVWIHHGLQJ� GXUDWLRQ� ZDV� LQÁXHQFHG� E\� SUHJQDQF\� LQWHQWLRQV� while controlling for socio-demographic factors. Breastfeeding behavior has been a topic of interest in a variety of sectors, from public health to psychology, GXH�WR�LWV�KHDOWK�EHQHÀWV�IRU�WKH�PRWKHU��VXFK�DV�SRVLWLYH� postpartum health outcomes (Newcomb et al., 1994; Visness et al., 1997) and advantageous psychological associations (Mortensen et al., 2002). Breastfeeding has also been associated with several physiological and psychological health bene- ÀWV� IRU� WKH�FKLOG�� VXFK�DV� LQFUHDVHG� LPPXQH�V\VWHP� PREGNANCY INTENTIONS AND HAPPINESS 60 functioning, long range developmental outcomes, and a secure mother-infant attachment bond (Davis et al., 1988; Lawrence, 2000; Mortensen et al., 2002). Research into breastfeeding behavior has fo- cused heavily on socio-demographic factors, partic- ularly on socio-economic status and how breastfeed- ing initiation and duration can be enhanced through health education for low-income parents. We argue that interventions designed to improve breastfeeding should include evaluation of the mother’s feelings about her pregnancy, and ways to address feelings of unwantedness/unhappiness. Poleschuck and Woods (2014) recommend that health psychologists should assist women who express surprise, ambivalence, or frustration about the pregnancy with exploring their feelings, reframing their expectations, garnering so- cial support and resources, and preparing for the im- pending changes of a newborn. The assistance from a psychotherapist may help to improve a variety of perinatal behaviors besides just breastfeeding, such as nutrition, smoking, and responsiveness to the infant’s needs. Research should explore how psychosocial ZHOO�EHLQJ�DQG�LQWHQWLRQV�DERXW�WKH�SUHJQDQF\�LQÁX- ence perinatal health behaviors and the psychological attachment formed between mother and child. While incentives and health education can be used to im- prove breastfeeding among women, attending to the SV\FKRORJLFDO� YDULDEOHV� ZKLFK� LQÁXHQFH� EUHDVWIHHG- ing outcomes could have lasting implications for the physical and mental health of both the mother and WKH�FKLOG�ZKLFK�WUDQVFHQG�IDU�SDVW�WKH�ÀUVW�\HDU�RI �OLIH� References Ahluwalia, I. B., D’Angelo, D., Morrow, B., & McDon- ald, J. A. (2012). Association between accultura- tion and breastfeeding among Hispanic women: Data from the pregnancy risk assessment and monitoring system. Journal of Human Lactation, 28(2), 167-173. doi:10.1177/0890334412438403 Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psycholog- ical study of the strange situation. Hillsdale, NJ: Erlbaum. American Academy of Pediatrics. (2013). Retrieved December 17, 2013, from http://www2.aap. org/breastfeeding/ Annual Update of the Health and Human Services Guidelines. (2001). Retrieved December 12, 2013, from http://aspe.hhs.gov/poverty/01fe- dreg.htm Blake, S. M., Kiely, M., Gard, C. C., El-Mohandes, A. E., El-Khorazaty, M., & NIH-DC Initiative (2007). Pregnancy intentions and happiness among pregnant Black women at high risk for adverse infant health outcomes. Perspectives On Sexual And Reproductive Health, 39(4), 194-205. Britton, J. R., Britton, H. L., & Gronwaldt, V. (2006). Breastfeeding, sensitivity, and attachment. Pediat- rics, 118(5), 1436-1443. doi:10.1542/peds.2005- 2916. Celi, A. C., Rich-Edwards, J. W., Richardson, M. K., Kleinman, K. P., & Gillman, M. W. (2005). Immi- gration, race/ethnicity, and social and economic factors as predictors of breastfeeding initiation. Archives of Pediatrics & Adolescent Medicine, 159(3), 255. Centers for Disease Control and Prevention. (2011). National immunization survey. Provisional breastfeeding rates by socio-demographic fac- tors among children born in 2007. Retrieved from www.cdc.gov/breastfeeding/data/NIS_ data/2007/socio-demographic_any.htm Chang, Y., Valliant, M., & Bomba, A. K. (2012). Gen- der differences in knowledge and attitude regard- ing breastfeeding. International Journal of Con- sumer Studies, 36, 342-351. Chen, P. G., Johnson, L. W., & Rosenthal, M. S. (2012). Sources of education about breastfeeding and breast pump use: What effect do they have RQ�EUHDVWIHHGLQJ�GXUDWLRQ"�$Q�DQDO\VLV�RI �WKH�,Q- fant Feeding Practices Survey II. Maternal Child Health Journal, 16, 1421-1430. doi:10.1007/ s10995-011-0908-4 Cheng, D., Schwarz, E. B., Douglas, E., & Horon, I. (2009). Unintended pregnancy and associat- ed maternal preconception, prenatal and post- natal behaviors. Contraception, 79, 194-198. doi:10.1016/j.contraception.2008.09.009 Chin, A. C., Myers, L., & Magnus, J. H. (2008). Race, education, and breastfeeding initiation in Louisi- ana. Journal of Human Lactation, 24(2), 175-185. WALLACE, CHAD, RODRIGUEZ 61 Davis, M. K., Savitz, D. A., & Graubard, B. I. (1988). Infant feeding and childhood cancer. The Lan- cet, 332 (8607), 365-368. Dennis, C. L. (2002). Breastfeeding initiation and duration: A 1990/2000 literature review. Journal of Obstetric, Gynecologic, & Neonatal Nursing, 31(1), 12-32. Dye, T. D., Wojtowycz, M. A., Aubry, R. H., Quade, J., & Kilburn, H. (1996). Unintended pregnancy and breast-feeding behavior. American Journal of Public Health, 87(10), 1709-1711. Fairlie, T. G., Gillman, M. W., Rich-Edwards, J. (2009). High pregnancy-related anxiety and prenatal de- pressive symptoms as predictors of intention to breastfeed and breastfeeding initiation. Journal of Women’s Health, 18(7), 945-953. doi:10.1089/ jwh.2008.0998 Finer, L. B., & Zolna, M. R. (2014). Shifts in intended and unintended pregnancies in the United States, 2001–2008. American Journal of Public Health, 104(S1), S43-S48. Foster, S. F., Slade, P., & Wilson, K. (1996). Body im- age, maternal fetal attachment, and breast feed- ing. Journal of Psychosomatic Research, 41(2), 181-184. Haughton, J., Gregorio, D., & Pérez-Escamilla, R. (2010). Factors associated with breastfeeding du- ration among Connecticut special supplemental nutrition program for Women, Infants, and Chil- dren (WIC) participants. Journal of Human Lac- tation, 26(3), 266-273. Jones, J. R., Kogan, M. D., Singh, G. K., Dee, D. L., & Grummer-Strawn, L. M. (2011). Factors asso- ciated with exclusive breastfeeding in the United States. Pediatrics, 128(6), 1117-1125. Joyce, T. J., Kaestner, R., & Korenman, S. (2000a). The effect of pregnancy intention on child devel- opment. Demography, 37(1), 83-94. Joyce, T. J., Kaestner, R., & Korenman, S. (2000b). The stability of pregnancy intentions and preg- nancy-related maternal behaviors. Maternal and Child Health Journal, 4(3), 171-178. Kellehner, C. M. (2012). The physical challeng- es of early breastfeeding. Social Science and Medicine, 63(10), 2727-2738. doi:10.1016/j. socscimed.2006.06.027 Korenman, S., Kaestner, R., & Joyce, T. (2002). Con- sequences for infants of parental disagreement in pregnancy intention. Perspectives on Sexual and Reproductive Health, 34(4), 198-205. /DZUHQFH�� 5�� �������� %UHDVWIHHGLQJ�� %HQHÀWV�� ULVNV�� alternatives. Current Opinion in Obstetrics and Gynecology, 12, 519-524. Lawson, K., & Tulloch, M. I. (1995). Breastfeeding duration: prenatal intentions and postnatal prac- tices. Journal of Advanced Nursing, 22(5), 841- 849. Li, R., Darling, N., Maurice, E., Barker, L., & Grum- mer-Strawn, L. M. (2005). Breastfeeding rates in the United States by characteristics of the child, mother, or family: The 2002 National Immuniza- tion Survey. Pediatrics, 115(1), e31-e37. Mortensen, E. L., Michaelsen, K. F., Sanders, S. A., & Reinisch, J. M. (2002). The association between duration of breastfeeding and adult intelligence. The Journal of the American Medical Associa- tion, 287(18), 2365-2371. Newcomb, P.A., Storer, B.E., Longnecker, M.P., Mit- tendorf, R., Greenberg, R., Clapp, R.W.,… Mac- Mahon, B. (1994). Lactation and a reduced risk of premenopausal breast cancer. The New En- gland Journal of Medicine, 330(2), 81-87. Noel-Weiss, J., Rupp, A., Cragg, B., Bassett, V., & Woodend, D. (2006). Randomized controlled tri- al to determine effects of prenatal breastfeeding ZRUNVKRS�RQ�PDWHUQDO�EUHDVWIHHGLQJ�VHOI�HIÀFDF\� and breastfeeding duration. Journal of Obstetric, Gynecologic, & Neonatal Nursing, 35(5), 616- 624. Piper, S. & Parks, P. L. (1996). Predicting the duration of lactation: Evidence from a national survey. Birth, 23(1), 7-12. Poleshuck, E. L., & Woods, J. (2014). Psychologists partnering with obstetricians and gynecologists: Meeting the need for patient-centered models of women’s health care delivery. American Psychol- ogist, 69(4), 344. Ryan, A., & Zhou, W. (2006). Lower breastfeeding rates persist among the special supplemental nu- trition program for women, infants, and children participants,1978-2003. Pediatrics, 117(4),1136- 1146. PREGNANCY INTENTIONS AND HAPPINESS 62 Smith-Gagen, J., Hollen, R., Walker, M., Cook, D. M., & Yang, W. (2014). Breastfeeding laws and breastfeeding practices by race and ethnicity. Women’s Health Issues, 24(1), 11-19. Stearns, C. A. (2009). The work of breastfeeding. Women’s Studies Quarterly, 37(3&4), 63-80. Taylor, J. S. & Cabral, H. J. (2002). Are women with an XQLQWHQGHG� SUHJQDQF\� OHVV� OLNHO\� WR� EUHDVWIHHG"� The Journal of Family Practice, 51(5), 431-436. Taylor, J. S., Geller, L., Risica, P. A., Kirtania, U., & Cabral, H. J. (2008). Birth order and breastfeed- ing initiation: Results of a national survey. Breast- feeding Medicine, 3(1), 20-27. doi:10.1089/ bfm.2007.0006 Thulier, D. & Mercer, J. (2009). Variables associated with breastfeeding duration. Journal of Obstet- ric, Gynecological, & Neonatal Nursing, 38, 259- 268. doi:10.1111/j.1552-6909.2009.01021 United States Department of Health and Human Services. National Center for Health Statistics. National Survey of Family Growth, Cycle VI, 2002. ICPSR04157-v1. Ann Arbor, MI: Institute for Social Research [producer], 2004. Ann Ar- bor, MI: Inter-university Consortium for Politi- cal and Social Research [distributor], 2008-10-01. doi:10.3886/ICPSR04157.v1 Vaaler, M. L., Castrucci, B. C., Parks, S. E., Clark, J., Stagg, J., & Erickson, T. (2011). Men’s attitudes toward breastfeeding: Findings from the 2007 Texas Behavioral Risk Factor Surveillance Sys- tem. Maternal and Child Health Journal, 15(2), 148-157. Visness, C.M., Kennedy, K.I., Ramos, R. (1997). The duration and character of postpartum bleeding among breastfeeding women. Obstetrics & Gy- necology, 89(2). WALLACE, CHAD, RODRIGUEZ