Black-White Differences in Pregnancy Desire During the

Demography (2021) 58(2):603–630DOI 10.1215/00703370-8993840 © 2021 The AuthorsThis is an open ac cess ar ti cle dis trib uted un der the terms of a Creative Commons license (CC BY-NC-ND 4.0).

Published online: 19 February 2021

ELECTRONIC SUPPLEMENTARY MATERIAL The online ver sion of this ar ti cle (https: / /doi .org /10 .1215 /00703370 -8993840) con tains sup ple men tary ma te ri al.

Black-White Differences in Pregnancy Desire During the Transition to Adulthood

Jennifer S. Barber, Ka ren Ben ja min Guzzo, Jamie Budnick, Yasamin Kusunoki, Sa rah R. Hayford, and Warren Miller

ABSTRACT This ar ti cle ex plores race dif fer ences in the de sire to avoid preg nancy or be come preg nant us ing sur vey data from a ran dom sam ple of 914 young women (ages 18–22) liv ing in a Michigan county and semi-struc tured in ter views with a sub sam ple of 60 of the wom en. In the sur vey da ta, de sire for preg nan cy, in dif fer ence, and am biv-a lence are very rare but are more prev a lent among Black women than White wom en. In the semi-struc tured in ter views, al though few women de scribed fa tal is tic be liefs or lack of plan ning for fu ture preg nan cies, Black and White women did so equally of ten. Women more of ten de scribed fa tal is tic be liefs and lack of plan ning when ret ro spec-tively de scrib ing their past than when pro spec tively de scrib ing their fu ture. Using the sur vey data to com pare pro spec tive de sires for a fu ture preg nancy with wom en’s rec-ol lec tions of those de sires af ter they con ceived, more Black women shifted pos i tive than shifted neg a tive, and Black women were more likely to shift pos i tive than White wom en—that is, Black women do not dif fer en tially ret ro spec tively over re port pro spec-tively de sired preg nan cies as hav ing been un de sired be fore con cep tion. Young wom en’s con sis tent (over re peated in ter views) pro spec tive ex pres sion of strong de sire to avoid preg nancy and cor re spond ingly weak de sire for preg nan cy, along with the sim i lar ity of Black and White wom en’s preg nancy plans, lead us to con clude that a “plan ning par a digm”—in which young women are en cour aged and supported in implementing their preg nancy de sires—is prob a bly ap pro pri ate for the vast ma jor ity of young women and, most im por tant ly, is sim i larly ap pro pri ate for Black and White young wom en.

KEYWORDS Racial in equal i ty • Racial disparities • Unintended preg nan cy • Undesired preg nan cy • Unplanned preg nan cy

Introduction

According to na tion ally rep re sen ta tive sur vey da ta, Black-White disparities in un in-tended preg nancy in the United States are large. Black women ret ro spec tively re port that 64% of their preg nan cies are un in tend ed; the cor re spond ing per cent age for White women is only 38% (Finer and Zolna 2016). The un in tended preg nancy rate is nearly 2.5 times higher for Black women (79 per 1,000) than for White women (33 per 1,000) (Finer and Zolna 2016). These large race disparities, as well as high over all lev els of

Dow

nloaded from http://read.dukeupress.edu/dem

ography/article-pdf/58/2/603/909821/603barber.pdf by guest on 10 February 2022

https://doi.org/10.1215/00703370-8993840

https://doi.org/10.1215/00703370-8993840

604 J. S. Barber et al.

un in tended preg nan cy, have prompted on go ing pub lic health ef forts to re duce un in-tended preg nan cies as well as grow ing ques tions about whether women ac tu ally form preg nancy in ten tions and plans (Aiken et al. 2016; Gómez et al. 2019). In par tic u lar, per sis tent race dif fer ences in un in tended preg nancy rates mo ti vate a closer ex am i na-tion of how the con cept and its mea sure ment dif fer for Black and White wom en.

If Black women are less able to ful fill their child bear ing de sires1 than White wom-en, this is an im por tant re pro duc tive jus tice is sue. However, some re search ers have ar gued that the ap par ent Black-White dis par ity in un de sired preg nancy arises from mis un der stand ing Black wom en’s preg nancy de sires. For ex am ple, it may be that young Black women want to get preg nant but tend to ret ro spec tively re port those preg nan cies as un de sired be cause they are re luc tant to ad mit that they wanted a child (Dash 2003; Kearney and Levine 2012). Alternatively, Black women may be more likely than White women to have preg nancy de sires that fall some where be tween clearly want ing or not want ing, and di chot o mous mea sures tend to cat e go rize those in-be tween preg nan cies as un de sired (Borrero et al. 2015; Kemet et al. 2018). Con-ceptually dis tinct, Black women may have sim i lar de sires as White women for preg-nancy or to avoid preg nancy but may be less likely to make plans to im ple ment their de sires. Finally, a pos si bil ity that has largely been ig nored is that Black wom en’s feel ings about their preg nan cies may shift to be come more neg a tive over the course of their preg nancy or af ter their child is born, rel a tive to White wom en’s more sta ble or pos i tively shifting feel ings about their preg nan cies, par tic u larly if they ex pe ri ence dis crim i na tion or other neg a tive re sponses to their preg nan cies. Any of these pos si bil-i ties could lead to an over es ti mate of ra cial in equal ity in un de sired preg nan cy.

To ad dress this gap in knowl edge, we con sider four re search ques tions. First, we ad dress whether young Black women pro spec tively (be fore con cep tion) ex press more de sire for a preg nancy than White wom en. Second, we con sider whether Black wom en’s pro spec tive preg nancy de sire is more fre quently am biv a lent or in dif fer ent to ward preg nancy than White wom en’s. Third, we ask whether Black wom en’s preg-nancy plans are qual i ta tively dif fer ent than White wom en’s preg nancy plans. Fourth, we test whether preg nant Black wom en’s pro spec tive de sires for preg nancy are more likely than preg nant White wom en’s de sires to shift neg a tive af ter they con ceive.

We draw on two types of data to mea sure preg nancy de sires and plans: (1) unique sur vey mea sures of pro spec tive de sire for preg nancy and de sire to avoid preg nan cy, and (2) in-per son semi-struc tured in ter views about wom en’s feel ings surrounding past and fu ture preg nan cies. We fo cus on a par tic u larly im por tant point in the life course: the tran si tion to adult hood, at ages 18–22, when un de sired preg nancy rates are the highest (Finer and Zolna 2016).

Pregnancy Desires and Intentions: The TDIB Framework

Our con cep tu al i za tion of preg nancy de sires and plans is based on the Traits-Desires- Intentions-Behavior (TDIB) frame work (Miller 1994). The TDIB frame work in cor po rates

1 The most com monly used mea sures of “un in tend ed” child bear ing (in the National Survey of Family Growth) ask whether women wanted to get preg nant (preg nancy de sire, not in ten tion). Our use of “un de-sired” is con sis tent with other re search (e. g., see Kost and Zolna 2019; Kost et al. 2018). We dis cuss this lan guage fur ther in the next sec tion.

Dow



605Black-White Differences in Pregnancy Desire

el e ments of long-stand ing the o ries of child bear ing be hav ior, in clud ing de mo graphic de mand-for-chil dren mod els (e. g., Bulatao and Lee 1983), mi cro eco nomic mod els of fer-til ity (e. g., Bagozzi and van Loo 1978), sub jec tive-expected util ity mod els (e. g., Townes et al. 1977), and the the ory of rea soned ac tion (Fishbein and Azjen 1975).

This frame work de scribes a four-part mo ti va tional se quence. The first part in volves largely un con scious child bear ing mo ti va tional dis po si tions—name ly, traits. (They are not our fo cus in this ar ti cle.) Second, these mo ti va tional dis po si tions in flu ence con scious de sires for and/or against get ting preg nant and hav ing a ba by. Third, child bear ing de sires in flu ence in ten tions (plans) to try to get preg nant or to avoid get ting preg nant. Fourth, in ten tions lead to spe cific be hav iors that are designed to achieve or avoid preg nan cy.

Desires take many forms, from deep urges or gut feel ings to ex plicit for mu la tions of wants. Pregnancy de sires, the cen tral con cept in this ar ti cle, re flect whether and how much a woman wants to get preg nant in a spe cific timeframe. They are driven in part by traits but also by life cy cle fac tors (e. g., age, mar i tal sta tus) and other spe cific fa cil i tat ing or com pet ing de sires (e. g., for a lov ing spouse, a col lege de gree, a fulfilling ca reer). Extensive psy cho log i cal re search has supported the ex is tence of two pri mary di men sions of de sires: pos i tive (per ceived re wards) and neg a tive (per ceived threats) (Cacioppo et al. 1999; Miller 1994; Stanley and Meyer 2009). The TDIB model draws on this re search in con cep tu al iz ing am biv a lence as si mul ta neous strong pos i tive and strong neg a tive de sires for child bear ing, and in dif fer ence as si mul ta neous weak pos-i tive and weak neg a tive de sires for child bear ing. Both of these conflicting states are strong pre dic tors of sub se quent in con sis tent con tra cep tive use and preg nancy among young women (Miller et al. 2013; Moreau et al. 2012).

In con trast to de sires, in ten tions are fully con scious de ci sions or plans about how to be have to achieve a spe cific out come. Desires must be trans lated into in ten tions be fore any rel e vant ac tion is tak en.2 Of course, in di vid u als vary in their will ing ness and abil ity to con vert de sires into plans. First, conflicting de sires (e. g., am biv a lent or in dif fer ent preg nancy de sire), or de sires for other be hav iors that con flict with (e. g., col lege) or sup port (e. g., mar riage) preg nan cy, may im pede the trans la tion of ei ther de sire into plans. Second, par tic u larly for be hav iors that re quire a part ner, in ten tions in cor po rate what oth ers de sire (e. g., the in ti mate part ner). Third, in ten tions are con-strained by what an in di vid ual thinks is ac tu ally pos si ble. For ex am ple, a woman may want to de lay child bear ing but have fa tal is tic views about preg nancy plan ning, be liev ing that God or other forces de ter mine when preg nancy will oc cur.

We dis tin guish be tween de sires and in ten tions for two rea sons. First, de sire is what re search ers com monly mea sure—ask ing what a woman wants or wanted rather than ask ing about what she in tends or plans to do. We join other re search ers in their call for us ing ac cu rate lan guage to de scribe the con cept be ing mea sured (Kost and Zolna 2019), par tic u larly be cause re fer ring to a preg nancy that a woman did not want as “un in tend ed” rather than “un de sired” im plic itly and er ro ne ously at tri butes a lack of plan ning or de ci sion-mak ing to the preg nant woman (Potter et al. 2019). Black wom en’s higher lev els of “un in tend ed” (sic) child bear ing is likely one rea son why re search ers have suggested that Black women are less planful than White women in terms of preg nan cy, de spite the many other rea sons that Black women may not get

2 Miller noted that an in ten tion is formed even with im pul sive ac tions. What dif fer en ti ates im pul sive ac tions is that the cor re spond ing de sire arises sud denly and force ful ly, over whelm ing prior in ten tions (Miller 1994:231).

Dow




what they want. Second, from a re pro duc tive jus tice per spec tive, it is im por tant to ask whether Black women get what they want as fre quently as White wom en, be fore their in ten tions, de ci sions, plans, and be hav ior are shaped by dif fer en tial ac cess to op por tu-ni ties as a re sult of struc tural rac ism (Bloome 2014; Broman 2005; Pager et al. 2009; Raley et al. 2015; Western et al. 2012; Wilson 2012).

Potential Race Differences in Desire for Pregnancy

Our first re search ques tion ex am ines the ex tent to which young Black women have more or less de sire for preg nancy than their White coun ter parts, draw ing on Arline Geronimus’ in flu en tial ideas about weath er ing, a biopsychosocial frame work for un der stand ing early health de te ri o ra tion among Black Amer i cans due to dis crim i na-tion and stress (Geronimus 1992, 2003; Geronimus et al. 2006). Geronimus used this frame work to ex plain why older Black women have less healthy births (e. g., lower birth weight and higher in fant mor tal i ty) than Black women who en ter moth er hood at youn ger ages, as well as why dif fer ent ra cial/eth nic mi nor ity groups tend to be come par ents at ages that min i mize their group-spe cific health risks (Geronimus 1987, 1992). If young Black women are aware of the po ten tial neg a tive con se quences of delaying child bear ing, they may want to en ter moth er hood while young to max i mize their chances of a healthy preg nancy and birth. In ad di tion, given higher mor bid ity and mor tal ity rates ear lier in the life course among Black rel a tive to White peo ple, a youn ger age at first birth may also max i mize the chances that grand par ents and other fam ily mem bers are avail able to help care for and in ter act with the ba by.

Even young Black women with high ed u ca tional as pi ra tions may pre fer youn-ger first births if they are aware that highly ed u cated Black women also ex pe ri ence weath er ing (Geronimus et al. 2006; Schoendorf et al. 1992). Although young child-bear ing might re duce their ed u ca tional at tain ment, Black women face more lim ited op por tu ni ties for ed u ca tion than White wom en. Further, most causal an a ly ses have dem on strated only small neg a tive con se quences of teen child bear ing on ed u ca tional out comes, and few, if any, neg a tive con se quences for par ent ing qual ity (Fletcher and Wolfe 2009; Geronimus and Korenman 1992; Hotz et al. 2005; Kane et al. 2013; Lee 2010). Consequently, what econ o mists call “op por tu nity costs” of young child bear-ing are likely lower for Black wom en.

A sec ond, re lated rea son that young Black women (as well as older Black wom en) may have a stron ger de sire for moth er hood than White women is that they dis pro-por tion ately live in impoverished neigh bor hoods (Lichter et al. 2012), and un cer-tainty and in sta bil ity are en demic to this con cen trated pov er ty. Burton and Tucker (2009) and Levine (2013) de scribed the in sta bil ity and in se cu rity that per vade the lives of poor Black wom en: em ploy ment op por tu ni ties that are lim ited to in ter mit tent and low-wage jobs, few al ter na tives to re duce their bread win ner bur den (e. g., sta bly employed hus bands), tran sient liv ing con di tions, anx i ety about se ri ous re la tion ships, fear of death, and gen eral mis trust. Because women view chil dren as an avail able path to sta bil ity for them selves and hope fully for the fa thers as well, moth er hood is val ued and sought af ter (Burton 1990; Edin and Kefalas 2005). This is also con sis-tent with de mog ra phers’ “un cer tainty re duc tion” the ory that hav ing chil dren is a key source of sta bil ity for in di vid u als whose other op tions for mak ing life seem more pre-

Dow




dict able and se cure (e. g., mar riage, ca reers, re tire ment sav ings) are lim ited (Friedman et al. 1994).3 Thus, women liv ing in un cer tain con di tions, such as those ex pe ri enced by many Black women in the United States, may de sire preg nancy at a youn ger age than other wom en. Because of res i den tial seg re ga tion and dis crim i na tion, even at high lev els of in come or ed u ca tion, these race dif fer ences may ex ist re gard less of so cio eco nomic char ac ter is tics.

Weathering and un cer tain ty/in sta bil ity form the ba sis of our first hy poth e sis:

Hypothesis 1: Black women have more de sire for preg nancy (and cor re spond-ingly less de sire to avoid preg nan cy) dur ing young adult hood than their White coun ter parts.

However, wom en’s feel ings about preg nancy are com plex, and there has been con sid-er able de bate about their ap pro pri ate mea sure ment (Klerman 2000; Kost and Lindberg 2015; Rackin and Mor gan 2018; Santelli et al. 2003, 2009), which leads to our sec ond re search ques tion—whether the con cept of preg nancy de sire itself is equally com plex across groups of women (Borrero et al. 2015; Foster et al. 2008; Moos et al. 1997; Stones et al. 2017). For ex am ple, Kemet and col leagues (2018:314) re cently wrote that “preg nancy in ten tion (sic) may not be en tirely rep re sen ta tive of the mul ti di men sional and intersecting so cial, emo tion al, cog ni tive and con tex tual as pects of preg nancy that Black and His panic wom en face,” and that “tra di tional mea sures of preg nancy in ten tion (sic) may of fer an in com plete rep re sen ta tion of Black and His panic wom en in par tic u-lar” (em pha ses added). They ar gued that ra cial/eth nic dif fer ences in the so cial ac cept-abil ity of and ex pec ta tions for young preg nancy ren der at tempts to mea sure preg nancy de sires or in ten tions less mean ing ful for Black and His panic women than for White wom en, pre sum ably be cause their so cial con texts are so dif fer ent.

In this per spec tive, Black com mu ni ties’ sup port of young par ent hood could en cour-age young Black women to de sire early births, es pe cially be cause par ent hood norms are but tressed by higher rates of re li gi os ity and re li gious at ten dance in Black com mu ni ties and the cor re spond ing pro-fam ily and pro-child bear ing ori en ta tion of re li gious groups (Chatters et al. 2009; Lin coln and Mamiya 1990; Mollborn 2017; Steensland et al. 2000). Thus, if young Black women si mul ta neously in ter nal ize these lo cal norms and conflicting so ci e tal norms against young par ent hood, they may have am biv a lence—pos i tive and neg a tive feel ings—about young preg nancy (Mollborn 2017; Sennott and Yeatman 2018). Indeed, re search ers have de scribed high lev els of am biv a lence among ur ban mi nor ity women (Aiken and Potter 2013; Cutler et al. 2018; Yoo et al. 2014). Alternatively, if the conflicting mes sages cause them to in ter nal ize nei ther set of norms, in dif fer ence about young preg nancy may be the re sult. Thus, our sec ond hy poth e sis is as fol lows:

Hypothesis 2: Young Black women are more in dif fer ent and/or am biv a lent about preg nancy than their White coun ter parts.

3 There is also a long his tory of mac ro-level hy poth e ses about tem po ral, rather than geo graph ic, var i a tion—that fer til ity in creases dur ing sta ble prosperous eco nomic pe ri ods and de creases dur ing the un cer tain/un sta ble pe ri ods of eco nomic down turns (for a re view, see Sobotka et al. 2011). However, con sis tent with our hy poth-e sis and the un cer tainty re duc tion as sump tion’s in di vid u al-level fo cus, other re search ers have found an in ter-ac tion ef fect with ed u ca tion: highly ed u cated women post pone par ent hood in times of un cer tain ly, whereas those with less ed u ca tion re spond to un cer tainty by en ter ing par ent hood (Kreyenfeld 2010).

Dow




Additionally, there is gen eral con cern that the con cepts of in ten tions or plan ning ap ply pri mar ily to White wom en, with their cor re spond ing so cio eco nomic ad van tage. In a re cent the o ret i cal cri tique of what they call the “plan ning par a digm,” Aiken and col leagues (Aiken et al. 2016) ar gued that the en tire con cept of preg nancy plan ning, and thus at tempts to mea sure un in tended or un planned preg nan cy, are largely in ap pli ca ble for some groups. Similar to the ar gu ments about mixed mes sages de scribed in the pre-vi ous par a graph, they ar gued that com plex ity and flu id ity of preg nancy de sires—along with dif fer ing cul tural norms, stig ma, and lev els of fa tal ism—make some groups of women want to let things un fold nat u ral ly, or de cide not to de cide. Empirical re search has documented high lev els of preg nancy fa tal ism or lack of plan ning among some mi nor ity pop u la tions (Borrero et al. 2015; Jones et al. 2015, 2016; Rocca and Harper 2012; Woodsong et al. 2004). This leads to our third hy poth e sis:

Hypothesis 3: Young Black wom en’s preg nancy plans are weaker or more fa tal-is tic than their White coun ter parts’ preg nancy plans.

The Dynamics of Pregnancy Desire

Finally, we also con sider whether re gard less of their pro spec tively mea sured pre-con cep tion de sire for preg nan cy, Black women are more likely than White women to ex pe ri ence a neg a tive shift in their feel ings about preg nancy af ter they con ceive. There are at least two rea sons this may be the case: the ma te rial con di tions in which young Black women ex pe ri ence their preg nan cies, and the cul tural stigma at tached to young Black preg nan cies.

First, young Black wom en—and their part ners—have less ac cess to sta ble high-pay ing jobs be cause of dis crim i na tion and op por tu nity de nial (Bloome 2014; Pager et al. 2009; Western et al. 2012). Black women also have less ac cess to part ners more gen er al ly, rel a tive to White wom en, given the higher rates of mor tal ity and in car-cer a tion among Black men rel a tive to White men (Raley et al. 2015; Wilson 2012). They are less likely to be mar ried when they con ceive, com pared with White wom en, and their in ti mate re la tion ships may be more con flic tual or part ner-dom i nated than White wom en’s (Broman 2005). Many young women hope to change these cir cum stances—their own em ploy ment, their part ner’s em ploy ment, or the qual ity of their re la tion ship—be fore be com ing preg nant or be tween con cep tion and the ba by’s birth (Edin and Kefalas 2005). If young Black women are less able to im prove these cir cum stances than White wom en, their feel ings may be more likely than White wom en’s feel ings to shift neg a tive as they come to grips with these cir cum stances.

Second, al though young Black women have likely ex pe ri enced rac ism, they may be un pre pared for the in ter sec tional stereotyping—based on their iden ti-ties as young, Black, and (prob a bly) un mar ried—that they ex pe ri ence as a re sult of their preg nan cies (Cole 2009; Rosenthal and Lobel 2016). Negative at ti tudes to ward young Black moth ers are fueled by per ni cious ste reo types about pro mis cu-ity (“Jez e bel”) and pub lic as sis tance (“wel fare queen”) (West 2008; Woodard and Mastin 2005). As a re sult, they may ex pe ri ence dis crim i na tion from their healthcare pro vid ers (Shavers et al. 2012), em ployers (Kennelly 1999), peers (Rosenthal and Lobel 2016), and oth ers.

Dow




Although ret ro spec tive mea sures of pre con cep tion preg nancy de sire ask preg nant women or moth ers to re call their feel ings be fore they be came preg nant, this is a cog-ni tively dif fi cult task if their feel ings have changed. People tend to be lieve that how they feel now is how they have al ways felt, a phe nom e non called “con sis tency bi as” (Schacter 1999).

Thus, based on ma te rial con di tions, stig ma, and con sis tency bi as, we hy poth e size the fol low ing:

Hypothesis 4: Young preg nant Black wom en’s feel ings are more likely than young preg nant White wom en’s feel ings to shift in a neg a tive di rec tion be tween their pro spec tively mea sured feel ings about a po ten tial preg nancy and their ret-ro spec tively mea sured pre con cep tion feel ings about their ac tual preg nan cy.

Data and Methods

Study Design

The Relationship Dynamics and Social Life (RDSL) study was based on a sim ple ran dom sam ple of the pop u la tion of young wom en, ages 18–19, re sid ing in Gene-see County, Michigan. The sam ple of 1,003 young women was drawn from driv er’s license and per sonal ID card re cords. A 60-min ute face-to-face base line sur vey in ter-view was conducted be tween March 2008 and July 2009 to as sess sociodemographic char ac ter is tics, fam ily back ground, at ti tudes, and early ex pe ri ences re lated to sex and re pro duc tive health. At the con clu sion of this base line in ter view, re spon dents were in vited to par tic i pate in a 2.5-year fol low-up study with weekly online or tele phone sur veys assessing in ti mate re la tion ships, sex, con tra cep tive use, preg nancy de sire, and preg nan cy. Details about sur vey in cen tives and re sponse rates are presented else-where (Barber et al. 2011, 2016). The fol low-up study con cluded in Jan u ary 2012 and yielded 58,594 weekly in ter views. Sample char ac ter is tics are reported in Table 1.

The RDSL Principal Investigator and re search as sis tants conducted 60- to 90-min ute semi-struc tured in ter views with two sub sets of RDSL re spon dents: those who ex pe ri-enced a preg nancy dur ing the study pe riod (n = 45), and those with high pro pen sity for preg nan cy4 but no preg nancy dur ing the study pe riod (n = 32).5 To en sure breadth, we strat i fied the sam ple along two ax es: poor ver sus non poor (based on re ceipt of pub lic as sis tance) and Black ver sus non-Black (based on self-reported race). Respondents were paid $40 for the semi-struc tured in ter view.

Two re spon dents did not con sent to be au dio recorded, and the re corder malfunc-tioned for an other in ter view. Two ad di tional in ter views did not re sult in us able da ta:

4 Respondents with high pro pen sity for preg nancy were se lected based on a haz ard model in clud ing the con trol var i ables listed in Tables 1 and 5, as well as time-vary ing preg nancy de sire, pro por tion of the study pe riod with an in ti mate part ner, and pro por tion of weeks with con sis tent con tra cep tive use (used a method ev ery time they had sex). Using the haz ard model co ef fi cients, the RDSL team se lected the non preg nant re spon dents with the highest predicted prob a bil ity of preg nan cy—that is, the non preg nant re spon dents who were most sim i lar to those who be came preg nant.5 By the time of the in ter view, four re spon dents se lected for preg nancy in ter views were not preg nant, and one re spon dent for a nonpregnancy in ter view was preg nant.

Dow




one re spon dent seemed to be fab ri cat ing or dra mat i cally embellishing her stories, and the other was non par tic i pa tory and dis tract ed. For this ar ti cle, we omit data from two ad di tional re spon dents who iden ti fied as nei ther Black nor White. In all , the el i gi ble sam ple of semi-struc tured in ter views in cluded 38 Black women and 32 White wom en.

All in ter views were tran scribed ver ba tim. Before an a lyz ing the semi-struc tured da ta, the re search team de vel oped a list of codes to cat e go rize each seg ment of text in the 2,343 pages of tran scribed in ter views, us ing the prox i mate de ter mi nants of preg nancy (sex ual be hav ior, con tra cep tive use) and char ac ter is tics of in ti mate re la-tion ships. We used a hy brid in duc tive-de duc tive ap proach, allowing un fore seen cate-gories to arise (Miles and Huberman 1984). Two trained re search as sis tants ap plied at least one code to all seg ments of text, and the re search team met fre quently to dis cuss discrepancies and de velop intercoder re li abil i ty. We used NVivo to at tach codes to the tex tual data to fa cil i tate tex tual an a ly ses.

Table 1 Characteristics of the re la tion ship dy nam ics and so cial life sam ple

Total SampleBlack Only

White Only

n = 914 women n = 317 n = 597

Mean SD Minimum Maximum

% Multiply Imputed Mean Mean

Pregnancy During Study Period .22 0 1 0 .24 .18Demographics Black .35 0 1 0 Age at base line 19.19 .57 18.12 20.34 0 19.18 19.20 Highly re li gious .58 0 1 0 .83 .44Childhood Disadvantage Mother had a teen birth .37 0 1 3 .56 .27 Mother’s ed u ca tion less than

high school .09 0 1 4 .12 .07 Grew up in a non-two-par ent

fam ily .46 0 1 0 .70 .34 Received pub lic as sis tance

dur ing child hood .36 0 1 0 .53 .27Current Socioeconomic

Characteristics High school GPA 3.16 .60 0 4.17 6 3.05 3.21 Receiving pub lic as sis tance .26 0 1 0 .41 .18Adolescent Experiences With Sex

and Reproductive Health (in dex sum ming the fol low ing ex pe ri ences be fore the RDSL study be gan) 1.84 1.49 0 4 4 2.31 1.60

One or more preg nan cies .25 0 1 <1 .39 .18 Age at first sex 16 years or

youn ger .52 0 1 <1 .62 .47 Two or more sex part ners .60 0 1 3 .70 .54 Ever had sex with out birth con trol .48 0 1 1 .61 .41

Dow




Dependent Variable Measures

Prospective Survey Measures of Pregnancy Desire

In each weekly sur vey when they were not preg nant, young women were asked mul-ti ple ques tions about their pro spec tive preg nancy de sire. We use the fol low ing two ques tions:

Desire for preg nan cy: How much do you want to get preg nant dur ing the next month? (0 = not at all want through 5 = re ally want)

Desire to avoid preg nan cy: How much do you want to avoid get ting preg nant dur ing the next month? (0 = not at all want to avoid through 5 = re ally want to avoid)

We use these ques tions to cre ate two di chot o mous mea sures of preg nancy de sire. First, be cause re spon dents rarely gave non zero re sponses to the ques tion about de sire for preg nan cy, and be cause any non zero re sponse sim i larly and strongly pre dicts sub-se quent preg nancy (Miller et al. 2013), we code scores of 1, 2, 3, 4, or 5 (any thing but 0) as any de sire for preg nan cy. Second, we code a score of 5 in re sponse to the ques tion about de sire to avoid preg nancy as stron gest de sire to avoid preg nan cy.

We also cre ate a cat e gor i cal com bined mea sure of preg nancy de sire based on pre vi ous re search us ing these ques tions (Miller et al. 2013). First, de sire for preg-nancy and de sire to avoid preg nancy are di chot o mized into strong (the top half of the re sponse categories: 3, 4, 5) and weak (the bot tom half of the re sponse cate-gories: 0, 1, 2). Next, we com bine these two di chot o mies into the fol low ing four categories: pronatal = strong de sire for preg nancy + weak de sire to avoid preg-nan cy; am biv a lent = strong de sire for preg nancy + strong de sire to avoid preg nan cy; in dif fer ent = weak de sire for preg nancy + weak de sire to avoid preg nan cy; and antinatal = weak de sire for preg nancy + strong de sire to avoid preg nan cy. We fur ther di vide antinatal into two categories. Strong antinatal is the spe cial case in which the de sire for preg nancy was the weakest (0) and the de sire to avoid preg nancy was the stron gest (5). The re main der of the antinatal cat e gory is called mod er ate antinatal.

Retrospective Survey Measures of Pregnancy Desire

In each weekly sur vey, re spon dents were asked, “Do you think there might be a chance that you are preg nant right now?” Respondents who an swered “yes” were asked, “Has a preg nancy test in di cated that you are preg nant?” When they reported a preg-nan cy, women were asked, “Before you found out you were preg nant, did you want to be come preg nant at some time in the fu ture?” For those who said no, preg nan cies are coded un de sired. Those who said yes were asked, “Did you be come preg nant at about the right time, ear lier than you wanted, or later than you wanted?” For those who responded “at about the right time,” preg nan cies are coded de sired. Those that were “ear lier than wanted” are coded un de sired. Only three re spon dents an swered “later than wanted” about their preg nan cy, which pre cludes cod ing them as a sep a rate cat e-go ry; we code them as de sired.

Dow




Survey Measures of Change in Pregnancy Desire

We com pare wom en’s ret ro spec tive rec ol lec tion of their pre con cep tion preg nancy de sire with their pro spec tive cat e gor i cal com bined mea sure of preg nancy de sire. The pro spec tive de sire is taken from the week prior to con cep tion, es ti mated based on the week the preg nancy was reported, the due date (updated dur ing the weekly in ter-views), the weeks in which she had sex with the fa ther, and/or the birth date. Change in preg nancy de sire is coded in two ways, as shown in Figure 1. Method A is con-ser va tive in regard to change: only switches from antinatal to de sired (shifted pos i-tive) and from pronatal to un de sired (shifted neg a tive) are coded as change. Method B also codes switches from am biv a lent/in dif fer ent to de sired (shifted pos i tive) and from am biv a lent/in dif fer ent to un de sired (shifted neg a tive).

Semi-Structured Interview Measures of Pregnancy Desire and Plans

Semi-struc tured in ter views with all women fo cused on de sires and plans for the fu ture, and also fo cused on cur rent preg nan cies among those who were preg nant. Because the in ter views also fo cused on cur rent and prior in ti mate re la tion ships, past preg nan cies were of ten part of the dis cus sion as well. The ex act ques tions depended on the flow of con ver sa tion and the vo cab u lary used by the re spon dents.

a. Method A of coding shift in pregnancy desire, including ambivalent/indifferent as change

Desired UndesiredPr

ospe

ctiv

eM

easu

re o

fPr

econ

cept

ion

Des

ire fo

rPr

egna

ncy Pronatal same shifted negative

Ambivalent/Indifferent same same

Antinatal shifted positive same

b. Method B of coding shift in pregnancy desire, not including ambivalent/indifferent as change

Pronatal same shifted negative

Ambivalent/Indifferent shifted positive shifted negative

Antinatal shifted positive same

Retrospective Measure ofPreconception Desire for Pregnancy

Pros

pect

ive

Mea

sure

of

Prec

once

ptio

nD

esire

for

Preg

nanc

y

Fig. 1 Coding of shift between prospective and retrospective measures of preconception pregnancy desire among pregnant women

Dow




Although de sires and in ten tions are con cep tu ally dif fer ent, when asked what they wanted in terms of child bear ing, these con ver sa tions about de sires of ten evolved into dis cus sions about spe cific plans.

Independent Variable Measures

Demographics

In the base line sur vey, all re spon dents were asked, “Which of the fol low ing groups de scribe your ra cial back ground? Please se lect one or more groups: Amer i can In dian or Alaska Native, Asian, Native Ha wai ian or Other Pacific Islander, Black or Af ri can Amer i can, or White.” Those who chose mul ti ple groups were asked which of the groups “best de scribes your ra cial back ground?” Our mea sure is coded 1 for Black and 0 for White.6 (The 8% of re spon dents who in di cated La tina eth nic ity in a pre ced ing ques tion are coded according to their re sponse to the ques tion about race.) Two per cent of the re spon dents reported an other race (Asian, Pacific Islander, or Native Amer i can) or did not iden tify a race; they are not in cluded in our an a ly ses. To cre ate age at base line (con-tin u ous in ex act years), we use the re spon dent’s birthdate from the driv er’s license and per sonal ID card re cords. Respondents who in di cated that re li gion was “very im por-tant” or “more im por tant than any thing else” are coded highly re li gious.

Childhood Disadvantage

We use four di chot o mous in di ca tors of child hood dis ad van tage: mother had a teen birth; moth er’s ed u ca tion was less than high school; re spon dent grew up in a non-two-par ent fam i ly (grew up with only one bi o log i cal par ent or in an other ar range ment, such as with grand par ents or an aunt); and re spon dent re ceived pub lic as sis tance dur-ing child hood.

Current Socioeconomic Characteristics

High school GPA is a con tin u ous var i able representing ed u ca tional at tain ment up to the time of the base line in ter view as well as the po ten tial for fu ture at tain ment. Respondents were coded as re ceiv ing pub lic as sis tance at the time of the base line in ter view if they in di cated at least one of the fol low ing sources: Women, Infants and

6 We rec og nize the in her ent lim i ta tions in this di chot o mous sim pli fi ca tion of Black and White wom en’s race. Women’s con cep tu al i za tion of their race can be nu anced, and it varies over time and space (Alba et al. 2016; Saperstein and Penner 2012). We fo cus on this sim pli fied cat e go ri za tion for par si mo ny, with the hope that this re search will spur ad di tional re search on this com plex top ic. In ad di tion, we fo cus ex clu-sively on race and not eth nic i ty dif fer ences, again for par si mo ny. All re spon dents who in di cated La tina eth nic ity also in di cated that they were ei ther Black or White, and their small num bers pre clude a sep a rate anal y sis. Removing the La tina women from the anal y sis does not change the re sults or our con clu sions.

Dow




Children pro gram (WIC), Family Independence Program (FIP); tem po rary as sis tance to fam i lies with chil dren (TANF); cash wel fare; or food stamps.

Adolescent Experiences with Sex and Reproductive Health

Four di chot o mous var i ables rep re sent ad o les cent ex pe ri ences be fore the base line in ter view: age at first sex ual in ter course ≥ 16, two or more sex ual part ners, ever had in ter course with out us ing con tra cep tion, and had one or more preg nan cies. All ques-tions were asked at the base line in ter view and re ferred to the past.

Missing Data

Survey Data

Because the ques tions about preg nancy de sire re fer to the up com ing month, only weekly in ter views that were com pleted more than 30 days af ter the prior in ter view (4%) re sult in a gap in the con tin u ous re cord of preg nancy de sire. Women skipped the ques tions about preg nancy de sire very in fre quently (<1% of weeks), but they were not asked about their preg nancy de sire when they were preg nant or thought they might be preg nant (5% of non preg nant weeks). In each of these cases, we use the mea sure of preg nancy de sire from the prior in ter view. Missing data for all other var i ables is mul ti ply im puted (us ing mi in Stata) with 10 it er a tions (by de fault). The per cent age of cases mul ti ply im puted is presented in Table 1. Overall, our an a lytic sam ple is 53,063 weekly in ter views with 914 re spon dents, 597 White and 317 Black. These women reported 224 preg nan cies dur ing the study pe ri od, but 10 preg nan-cies are miss ing data on ret ro spec tive preg nancy de sire; those preg nan cies are not in cluded in our re gres sion mod el.

Semi-Structured Interview Data

Six Black re spon dents and four White re spon dents did not dis cuss preg nancy de-sires or plans. In some cases, the in ter views fo cused on other top ics the re spon dents wanted to dis cuss (e. g., re la tion ship with the ba by’s fa ther, ba by’s fa ther re cently get ting shot). In oth ers, the in ter viewer felt that the re spon dent was dis cour ag ing her from ask ing about fu ture child bear ing plans (e. g., re spon dent was very un happy about the cur rent preg nan cy, the re spon dent was never preg nant and not in a re la tion-ship). In all , we use 32 in ter views with Black women and 28 in ter views with White women in our an a ly ses.

Analytic Strategy

To test Hypotheses 1 and 2, we com pare the week-level and wom an-level sur vey mea-sures of de sire for preg nancy and de sire to avoid preg nancy for Black and White wom en.

Dow




We used t tests from un ad justed re gres sion mod els (lo gis tic for di chot o mous var i ables and or di nary least squares for con tin u ous var i ables, with clus tered stan dard er rors at the week lev el) to de ter mine whether dif fer ences were sta tis ti cally sig nifi cant by race. We used the two lev els to ad dress dif fer ent ques tions: (1) at the week lev el, whether there were dif fer ences in the over all pro por tion of weeks with any de sire for preg nancy and/or less than the stron gest de sire to avoid preg nancy and the strength of those de sires in those weeks; and (2) at the woman lev el, whether there were dif fer ences in the pro por-tion of women who ev er gave such re sponses and in the con sis tency or strength of those de sires among those wom en.

To test Hypothesis 3, we used NVivo to ex tract all text seg ments broadly pertaining to preg nancy de sires and plans. We read all this text, as well as the text be fore and af ter the seg ments, and reread most of the in ter views in their en tire ty. In ad di tion to dis cern-ing de sires and plans for child bear ing, we spe cifi cally looked for re sponses suggesting that it was im pos si ble or un de sir able to make such plans. We also noted whether the plans re ferred to po ten tial fu ture preg nan cies (pro spec tive) or re ferred to pre con cep tion feel ings about a cur rent or past preg nancy (ret ro spec tive). We dis tilled each seg ment down to a short ex cerpt, lightly edited for read abil i ty. We in duc tively de vel oped sev eral categories to fa cil i tate qual i ta tive com par i sons across race: stop ping/long de lay, child spac ing, ed u ca tion/ca reer, ma te rial con di tions, re la tion ship con di tions, age range/oth er, proception, and not plan ning. In the text, we de scribe the qual i ta tive dif fer ences, or lack there of, be tween the ex cerpts for Black women and those for White wom en, within cat e go ry, sep a rately for pro spec tive and ret ro spec tive de sires/plans. We pro vide rep re-sen ta tive ex am ples in the text and all ex cerpts in the online ap pen dix.

To test Hypothesis 4, we first pres ent cross-tab u la tions of pro spec tive and ret ro-spec tive preg nancy de sire, for preg nant wom en, strat i fied by race. We es ti mated two mul ti no mial lo gis tic re gres sion mod els of change over time in preg nancy de sire, which com pare the log odds of be ing in two categories—neg a tive shift and pos i tive shift—with the ref er ence cat e go ry, no change. We re port co ef fi cients, which rep re sent the es ti mated ad di tive ef fect of the in de pen dent var i able on the log odds of pos i tive shift ver sus no change and neg a tive shift ver sus no change.

Results

Desire for Pregnancy and Desire to Avoid Pregnancy (Hypothesis 1)

Table 2 shows com par i sons for the sur vey mea sures of preg nancy de sires. Overall, preg nancy de sire was low; women expressed any de sire for preg nancy in only 7% of their weekly in ter views. However, more than one-third (37%) of women expressed some non zero de sire for preg nancy dur ing at least one of their weekly in ter views. Among this group of women who ever had any preg nancy de sire, when they expressed such de sire, its strength was mod er ate; the mean across these women in their non zero weeks was 2.98. The mean de sire across all non zero weeks was 3.03. The con sis tency of their de sire was low: they expressed non zero de sire in only 22% of their weekly in ter views.

The only sig nifi cant race dif fer ence in de sire for preg nancy is that a larger pro por-tion of Black women ever had any (non ze ro) de sire for preg nancy than White women

Dow




Tabl

e 2

Prop

ortio

ns o

r mea

ns (r

ange

s in

pa re

n the

ses)

for m

ea su

res o

f pre

g nan

cy d

e sire

s (R

elat

ions

hip

Dyn

amic

s and

Soc

ial L

ife d

ata

set)

Wee

k-Le

vel M

easu

res

Wom

an-L

evel

Mea

sure

s

All

Wom

en

(n =

53,0

63)

Bla

ck

Wom

en

(n =

14,5

29)

p fo

r D

iff.

Whi

te

Wom

en

(n =

38,5

34)

All

Wom

en

(n =

914)

Bla

ck

Wom

en

(n =

317)

p fo

r D

iff.

Whi

te

Wom

en

(n =

597)

Des

ire fo

r Pre

gnan

cyD

esire

for P

regn

ancy

A

ny d

e sire

for p

reg n

ancy

(0, 1

).0

7.0

9.0

6

Ever

had

any

de s

ire fo

r pre

g nan

cy (0

,1)

.37

.46

***

.32

D

urin

g w

eeks

with

non

zero

pr

eg na

ncy

de si

re (n

= 3,

744a ):

A

mon

g w

omen

who

eve

r had

any

de s

ire fo

r pr

eg na

n cy:

St

reng

th: L

evel

of d

e sire

for

preg

nanc

y (1

–5)

3.03

3.20

2.94

Stre

ngth

: Mea

n de

sire

for p

reg n

ancy

dur

ing

w

om an

’s n

on ze

ro d

e sire

wee

ks (1

.00–

5.00

)

(n =

327a )

2.98

2.96

3.00

Con

sist

ency

: Pro

porti

on o

f wom

an’s

wee

ks

w

ith a

ny d

e sire

for p

reg n

ancy

(.01–

1.00

) (n =

328a )

.22

.21

.23

Des

ire to

Avo

id P

regn

ancy

Des

ire to

Avo

id P

regn

ancy

A

nyth

ing

less

than

the

stron

gest

de sir

e to

avo

id p

reg n

ancy

(1, 0

).0

8.1

1.0

7

Ever

had

any

thin

g le

ss th

an th

e st

ron g

est d

e sire

to

avo

id p

reg n

ancy

(1,0

).4

4.5

5**

*.3

9

Dur

ing

wee

ks w

ith le

ss th

an th

e st

ron g

est d

e sire

to a

void

pr

eg na

ncy

(n =

4,22

2a ):

A

mon

g w

omen

who

eve

r had

any

thin

g le

ss

than

the

stro

n ges

t de s

ire to

avo

id p

reg n

ancy

(n

= 40

1a ):

Stre

ngth

: Lev

el o

f de s

ire to

av

oid

preg

nanc

y (0

–4)

2.38

2.31

2.42

Stre

ngth

: Mea

n de

sire

to a

void

pre

g nan

cy

du

r ing

wee

ks w

ith le

ss th

an th

e st

ron g

est

de

sire

to a

void

pre

g nan

cy (.

01–4

.00)

1.90

1.92

1.88

Con

sist

ency

: Pro

porti

on o

f wee

ks w

ith le

ss

th

an th

e st

ron g

est d

e sire

to a

void

preg

nanc

y(.0

1–1.

00)

.22

.23

.22

Dow




Wee

k-Le

vel M

easu

res

Wom

an-L

evel

Mea

sure

s

All

Wom

en

(n =

53,0

63)

Bla

ck

Wom

en

(n =

14,5

29)

p fo

r D

iff.

Whi

te

Wom

en

(n =

38,5

34)

All

Wom

en

(n =

914)

Bla

ck

Wom

en

(n =

317)

p fo

r D

iff.

Whi

te

Wom

en

(n =

597)

Cat

egor

ical

Com

bine

d M

easu

re

of P

regn

ancy

Des

ireC

ateg

oric

al C

ombi

ned

Mea

sure

of P

regn

ancy

D

esire

St

rong

ant

inat

al.9

1.8

9.9

3

Mea

n pe

r cen

t age

of w

eeks

stro

ng a

ntin

atal

.89

.87

*.9

1

Mod

erat

e an

tinat

al.0

32.0

31.0

32

Mea

n pe

r cen

t age

of w

eeks

mod

er at

e an

tinat

al.0

38.0

44.0

35

Indi

ffere

nt.0

086

.015

***

.006

1

Mea

n pe

r cen

t age

of w

eeks

in di

f fer e

nt.0

16.0

25*

.011

A

mbi

vale

nt.0

22.0

38**

.016

M

ean

per c

ent a

ge o

f wee

ks a

m bi

v a le

nt.0

25.0

33.0

21

Pron

atal

.024

.028

.023

M

ean

per c

ent a

ge o

f wee

ks p

rona

tal

.028

.030

.028

Not

e: V

ery

smal

l pro

por ti

ons a

nd m

eans

of v

ery

smal

l pro

por ti

ons a

re p

rese

nted

with

thre

e or

four

dec

i mal

pla

ces i

n or

der t

o pr

es en

t tw

o si

g nifi

cant

(non

-lead

ing-

ze ro

) dig

its.

a Bec

ause

ther

e ar

e 10

im pu

ta tio

ns a

nd th

e sa

m pl

e si

ze v

arie

s ac

ross

im pu

ta tio

ns, t

hese

sub

sam

ples

do

not e

x act

ly m

atch

914

wom

en ti

mes

the

frac

tion

of w

omen

repo

rted

in

the

line

abov

e.

*p <

.05;

**p

< .0

1; *

**p <

.001

(tw

o-ta

iled

t tes

ts).

Tabl

e 2

(con

tinue

d)

Dow




(46% vs. 32%, re spec tive ly). However, the women who ever had any de sire for preg-nancy did not dif fer in terms of the strength or con sis tency of their de sire.

Correspondingly, de sire to avoid preg nancy was very high at these ages. Women expressed any thing less than the stron gest pos si ble de sire to avoid preg nancy (5, on a 0 to 5 scale) in only 8% of the weekly in ter views. However, at the woman lev el, nearly one-half (44%) had some thing less than the stron gest de sire to avoid preg nancy in at least one week dur ing the study. These women had a mean de sire to avoid preg nancy of 1.90 in the weeks when they did not re spond with a 5, and the mean across all weeks that were not coded 5 is 2.38. Consistency was low, at an av er age of 22% of weeks.

The only race dif fer ence in de sire to avoid preg nancy is that more Black than White women ever had some thing other than the stron gest de sire to avoid preg nancy (55% vs. 39%). However, those women did not dif fer in the strength of their de sire to avoid preg nancy or in the con sis tency of their de sire to avoid preg nan cy.

Ambivalence and Indifference (Hypothesis 2)

Table 2 also shows that wom en’s de sire for preg nancy and de sire to avoid preg nancy tend to align. Women reported zero de sire for preg nancy and the stron gest de sire to avoid preg nancy (strong antinatal) in the vast ma jor ity (91%) of their weekly in ter views. Moderate antinatal de sire was the next most com mon com bi na tion but oc curred in only 3.2% of in ter views. Pronatal de sire—strong de sire for preg nancy and cor re spond ingly weak de sire to avoid preg nan cy—was reported in 2.4% of in ter views. Overall, in con-sis tent re sponses to the two ques tions were quite rare: less than 1% (.86) of weeks for in dif fer ence (weak de sire for preg nancy and weak de sire to avoid preg nan cy) and 2.2% for am biv a lence (strong de sire for preg nancy and strong de sire to avoid preg nan cy).7 The cor re spond ing mea sures at the woman lev el, which were com puted as the pro por-tion of a wom an’s to tal weekly in ter views in each cat e go ry, are sim i lar.

Black and White women ex hibit sim i lar pat terns. At the week lev el, they dif fer in only the two in con sis tent categories: Black women more fre quently expressed am biv a-lence (3.8% vs. 1.6%) and in dif fer ence (1.5% vs. 0.61%) than White wom en. Although the two categories were very rare among both Black and White wom en, these race dif-fer ences are sta tis ti cally sig nifi cant.

At the woman lev el, the pro por tions are sim i lar, but there are two dif fer ences in sta tis ti cal sig nifi cance. First, the race dif fer ence in the mean pro por tion of weeks in the am biv a lent cat e gory is not sig nifi cant. The slightly smaller wom an-level dif fer-ence (.009 smaller than the week-level dif fer ence, not shown) and the smaller sam ple size for women rel a tive to weeks ren der it in sig nifi cant (p = .13, not shown). Second, the race dif fer ence in the mean pro por tion of weeks in the strongly antinatal cat e gory is sig nifi cant, al though the p val ues are not very dif fer ent across the two lev els (p = .04 for woman lev el; p = .06 for week lev el).

7 The preg nancy de sire ques tions were in tro duced as fol lows, to en cour age in con sis tent re sponses: “You know, get ting preg nant and hav ing a baby is a big event, one that has a lot of con se quences. Most peo ple your age have some pos i tive and some neg a tive feel ings about get ting preg nant and hav ing a child. For this rea son we are go ing to ask you first how much you want to get preg nant, us ing a scale from 0 to 5. Then we are go ing to ask you how much you want to avoid get ting preg nant, us ing a scale from 0 to 5.”

Dow




Planning (Hypothesis 3)

Black and White women sim i larly de scribed their plans for fu ture preg nan cies. The most com mon plan, for both groups, was to stop child bear ing al to gether or to have a long de lay. For ex am ple, “If I de cide to get off (in ject able con tra cep tion), I’ll be get-ting my tubes tied” (Black wom an), or “After I have this ba by, I am go ing to get my tubes tied” (White wom an). Another com mon plan was for a spe cific age gap be tween chil dren. For ex am ple, “I think I will have about three or four. Not back-to-back, ei ther. I won’t do that. I’ll wait a cou ple of years” (Black wom an), or “I’m not try ing to have them back-to-back, I want space” (White wom an).

Others de scribed ed u ca tion or ca reer plans as de ter mi nants of the tim ing of their next preg nan cy—for ex am ple, “af ter I get my de gree” and “I’d wait un til I had al ready got ten established and had a job” (Black wom en), and “I’d rather get done with school and ev ery thing first,” and “I want to start a ca reer, get some sta bil ity in my life be fore I even think about hav ing kids or any thing like that” (White wom en). Closely re lat ed, sev eral women had ma te rial con di tions they wanted be fore hav ing a(noth er) ba by. For ex am ple, “I don’t want to penny pinch. I want to be com fort able. I want to be in a po si tion to get my kids what ever they want” (Black wom an), and “. . . right now is not the best time fi nan cially to have a kid” (White wom an). Relationship con di tions were also com mon, such as “I defi nitely want to be mar ried first” (Black wom an) and “Clearly we don’t want to get preg nant yet. We want to wait un til (wed ding month) be fore we plan for an other kid” (White wom an).

A few Black and White women de scribed some what less-spe cific plans—for ex am-ple, a pos si ble age range, or a time frame—and did not pro vide a spe cific ra tio nale. For ex am ple, “I’m 23 now, so I’m hoping that by around 25, 26, I’ll have one by then” (Black wom an), and “I’m just not ready” and “I don’t have sex at all with out con doms and birth con trol” (White wom en).

Only one Black woman and two White women suggested fa tal ism or lack of plan-ning their fu ture preg nan cies. The Black woman said, “I mean, if it hap pens, it hap-pens, what can you do?” Consistent with other re search on re pro duc tion, how ev er, re sponses did not con form to a bi nary con cep tu al i za tion of fa tal ism (Bell and Het-terly 2014; Jones et al. 2016). That re spon dent im me di ately followed her state ment by suggesting her own agen cy: “But I’m not try ing to make it hap pen.” The two White women who expressed fa tal is tic feel ings sim i larly presented a non bi nary pic-ture. One White woman said, “I don’t want to put a time on it, be cause when it’s bound to hap pen . . . When it’s your sea son, it’s go ing to hap pen. I mean, when ever God has that per son for me, and me and the guy get mar ried,” and she followed that with, “But noth ing right now, noth ing in the next cou ple years.” The sec ond White woman said, “But if it were to hap pen, I would roll with the punches like I did with [child’s name],” but also “I don’t want more kids right now.”

Thus, we find only scant ev i dence for fa tal ism or lack of plan ning for fu ture preg-nan cies in this sam ple of young wom en. And there is no ev i dence of a race dif fer ence in terms of planfulness.

Women’s ret ro spec tive de scrip tions of past preg nan cies in cor po rated many of the same themes, but many more of these women de scribed a lack of plan ning: 6 of the 10 Black wom en, and 7 of the 13 White wom en. Black and White women used sim i lar ter mi nol o gy, such as “Whenever it hap pened,” “We weren’t try ing but we weren’t not

Dow




try ing,” “If it’s go ing to hap pen” (Black wom en), “I didn’t care ei ther way,” “There’s never a good time for a baby to come,” and “Everything hap pens for a rea son” (White wom en); and they used pas sive voice to de scribe what hap pened, such as “I ended up preg nant” (White wom an). Three re spon dents spe cifi cally men tioned God’s will as in stru men tal in their preg nan cy. Two Black women said, “God didn’t . . . it wasn’t time for me yet,” and “He wrote my life,” and one White woman de scribed praying to God that she wasn’t preg nant. Two of the six Black women were de scrib ing oth-ers’ fa tal is tic views about preg nan cy: one woman de scribed her boy friend’s mother as say ing “If it’s go ing to hap pen . . . ,” and an other de scribed her boy friend’s nonplan for her to get preg nant “when ever it hap pened” but also de scribed her own preference: “Even though I was hav ing un pro tected sex, I didn’t want a baby then.”

Overall, we do not find race dif fer ences in wom en’s de scrip tions of their plans for their preg nan cies. However, there is a strong dif fer ence be tween wom en’s pro spec-tive and ret ro spec tive de scrip tions of plan ning their preg nan cies, with the ret ro spec-tive ac counts of women who ac tu ally got preg nant—both Black and White—much more likely to de scribe a lack of plan ning.

Dynamic Change Over Time in Pregnancy Desire (Hypothesis 4)

Table 3 shows the cross-tab u la tion of pro spec tive and ret ro spec tive preg nancy de sire, sep a rately for preg nant Black and White wom en, us ing both meth ods of cod ing change (de scribed in the Measures sec tion). Note that in both meth ods, we com bine the am biv-a lent and in dif fer ent categories be cause they are so un com mon among the preg nant wom en.

As panel A shows, al though a sub stan tially smaller pro por tion of preg nant Black women were pro spec tively pronatal rel a tive to White women (4% vs. 19%; see to tals for “Pronatal” rows), sim i lar pro por tions ret ro spec tively reported their preg nan cies as de sired at the time of con cep tion (17% and 19%, re spec tive ly; see to tals for “Desired” col umns). In other words, a smaller pro por tion of Black than White women shifted neg a tive over time. In fact, 77% (3% + 74%, shaded ar eas) of preg nant Black women had sta ble preg nancy de sires, whereas 14% shifted pos i tive and only 9% shifted neg-a tive. Among White wom en, 72% (9% + 63%, shaded ar eas) remained sta ble, and the pat tern was re versed: 10% shifted pos i tive, and a larger per cent age (18%) shifted neg a tive.

Multinomial lo gis tic re gres sion mod els, shown in Table 4, con firm that Black women are sig nifi cantly less likely than are White women to shift neg a tive af ter con cep tion, and that Black and White women do not dif fer in their log odds of shifting pos i tive. Model 1 shows the un ad justed as so ci a tion be tween race and shifting preg nancy de sire, which is not sta tis ti cally sig nifi cant (the p value for a neg a tive shift is .067). However, Model 2 in di cates that once con trol var i ables are added to the mod el, the race dif fer ence is sig nifi cant and sub stan tial, with Black women hav ing 1.26 lower log odds (OR = .28, 72% lower odds) than White women of shifting neg a tive. This is be cause young women whose moth ers did not grad u ate from high school are par tic u larly un likely to shift neg a-tive and are in fact likely to shift pos i tive, and these women are over rep re sented among young Black moth ers.

We also reestimated the mul ti no mial lo gis tic re gres sion model us ing method B

Dow




Tabl

e 3

Tabu

latio

n an

d cr

oss-

tab u

la tio

n of

pro

spec

tive

and

ret ro

spec

tive

mea

sure

s of p

re co

n cep

tion

de si

re fo

r pre

g nan

cy a

mon

g pr

eg na

nt w

om en

, sep

a rat

e by

race

(n

= 21

4 pr

eg na

n cie

s, R

elat

ions

hip

Dyn

amic

s and

Soc

ial L

ife d

ata

set)

Ret

rosp

ectiv

e R

ecol

lect

ion

of P

reco

ncep

tion

Des

ire

Bla

ck W

omen

Whi

te W

omen

Pros

pect

ive

Prec

once

ptio

n D

esire

Des

ired

Und

esire

dTo

tal

Des

ired

Und

esire

dTo

tal

A. M

etho

d A

of C

odin

g C

hang

e in

Pre

gnan

cy D

esire

, In

clud

ing

Am

biva

lent

/Indi

ffere

nt a

s Cha

nge

Pr

onat

al3

(3%

)8

(9%

)4

(4%

)11

(9%

)22

(18%

)

23 (1

9%)

Am

biva

lent

/Indi

ffere

nt13

(14%

)10

(11%

)

12 (1

0%)

13 (1

1%)

Ant

inat

al70

(74%

)80

(85%

)75

(63%

)84

(70%

)

Tota

l16

(17%

)78

(83%

)94

(100

%)

23 (1

9%)

97 (8

1%)

120

(100

%)

B. M

etho

d B

of C

odin

g C

hang

e in

Pre

gnan

cy D

esire

, N

ot In

clud

ing

Am

biva

lent

/Indi

ffere

nt a

s Cha

nge

Pron

atal

6 (6

%)

1 (1

%)

4 (4

%)

14 (1

2%)

12 (1

0%)

23 (1

9%)

Am

biva

lent

/Indi

ffere

nt

77 (8

1%)

10 (1

1%)

85 (7

1%)

13 (1

1%)

Ant

inat

al10

(11%

)80

(85%

)9

(8%

)84

(70%

)

Tota

l16

(17%

)78

(83%

)94

(100

%)

23 (1

9%)

97 (8

1%)

120

(100

%)

Not

es: S

hade

d ar

eas

are

code

d as

“no

dif f

er en

ce”

be tw

een

pro s

pec t

ive

and

ret ro

spec

tive

mea

sure

s. W

hite

ar e

as in

di ca

te a

pos

i tive

or n

eg a t

ive

shift

from

pro

spec

tive

mea

sure

of

pre

con c

ep tio

n de

sire

to re

t ro sp

ec tiv

e m

ea su

re o

f pre

con c

ep tio

n de

sire

.

Dow




Tabl

e 4

Mul

tinom

ial l

o gis

tic re

gres

sion

mod

els p

redi

ctin

g sh

ift b

e tw

een

pro s

pec t

ive

and

ret ro

spec

tive

mea

sure

s of p

re co

n cep

tion

preg

nanc

y de

sire

am

ong

preg

nant

wom

en

(co e

f fi ci

ents

, with

stan

dard

er r

ors i

n pa

ren t

he se

s; n

= 21

4 pr

eg na

n cie

s, R

elat

ions

hip

Dyn

amic

s and

Soc

ial L

ife d

ata

set)

Mod

el 1

Mod

el 2

Neg

ativ

e Sh

ift

in D

esire

vs.

No

Cha

nge

Posi

tive

Shift

in

Des

ire v

s. N

o C

hang

eN

egat

ive

Shift

in D

esire

vs

. No

Cha

nge

Posi

tive

Shift

in D

esire

vs

. No

Cha

nge

Bla

ck–.

85(.4

5).2

4(.4

7)−1

.26

(.51)

*.3

7(.6

1)H

ighl

y R

elig

ious

.54

(.42)

–.18

(.61)

Mot

her H

ad a

Tee

n B

irth

.05

(.42)

.04

(.45)

Mot

her’s

Edu

catio

n Le

ss T

han

Hig

h Sc

hool

–.35

(.65)

***

1.16

(.65)

*G

rew

up

in a

Non

-Tw

o-Pa

rent

Fam

ily.2

8(.4

7).3

2(.5

1)R

ecei

ved

Publ

ic A

ssis

tanc

e D

urin

g C

hild

hood

.58

(.42)

.48

(.48)

Ado

lesc

ent E

xper

ienc

es W

ith S

ex a

nd

Rep

rodu

ctiv

e H

ealth

(in d

ex)

.23

(.17)

.29

(.25)

Con

stan

t−1

.36

−1.9

7−2

.64

−3.3

8(1

.02)

Pseu

do-R

2.0

1.0

6

*p <

.05;

***

p < .0

01 (t

wo-

taile

d te

sts)

Dow




(shown in Figure 1 [panel b] and Table 3 [panel B]), in stead con sid er ing am biv a-lent or in dif fer ent women as shifting pos i tive if they ret ro spec tively reported their preg nan cies as de sired and shifting neg a tive if they ret ro spec tively reported their preg nan cies as un de sired (rather than cod ing these two groups as “no change”). Under this sce nar io, only 1 Black woman (1%) shifted neg a tive, whereas 12 White women (10%) did so.

Discussion

Although our re sults could be interpreted as con sis tent with the idea that weath er-ing or in sta bil ity leads Black women to want preg nancy more than White women at these young ages, the race dif fer ences in our an a ly ses are not par tic u larly strong, be cause the over whelm ing ma jor ity of both Black and White young women wanted to avoid or de lay preg nancy in the near fu ture. Further, Black women were less likely than were White women to ret ro spec tively re call their pre con cep tion preg nancy de-sires as more neg a tive than they were. In other words, we find no ev i dence for the idea that young preg nant Black women ac tu ally wanted to be come preg nant more than young preg nant White wom en, but ret ro spec tively reported those preg nan cies as un de sired be cause of the stigma they as so ci ated with want ing a preg nancy at a young age (Aiken et al. 2016; Kearney and Levine 2012). On the con trary, per haps young women who wanted to de lay preg nancy ret ro spec tively reported those preg nan cies as de sired be cause of the stigma as so ci ated with be ing un able to avoid their preg-nan cies. However, the RDSL’s ret ro spec tive sur vey ques tions about pre con cep tion preg nancy de sires were asked while the women were still preg nant. It may be that Black women shift neg a tive about their preg nan cies over the long term, but the RDSL ques tions were asked be fore they ex pe ri enced the in ter sec tional stigma and dis crim i-na tion as so ci ated with be ing a young Black preg nant wom an. Future re search should fur ther ad dress the dy nam ics in her ent in these feel ings.

We find lim ited ev i dence that young Black wom en’s pro spec tive feel ings about preg-nancy are more am biv a lent and/or in dif fer ent than White wom en’s, and over all find very low lev els of am biv a lence and es pe cially in dif fer ence. Our ap proach to am biv a-lence dif fers from most other re search, which has de fined women as am biv a lent if they want to avoid preg nancy but would ac cept, wel come, or be happy about a preg nancy any way, and typ i cally has not de fined in dif fer ence (Aiken and Potter 2013; Higgins 2017; Yoo et al. 2014). Perhaps our stricter defi ni tion of am biv a lence—si mul ta neously want ing and not want ing preg nan cy—pro duced these lower prev a lence es ti ma tes. Oth-ers have ar gued for more clar ity in de fin ing am biv a lence and for sep a rat ing the no tions of preg nancy ac cept abil ity or hap pi ness from preg nancy plan ning and de sire (Gómez et al. 2019).

In con trast to many oth ers, we do not find race dif fer ences in young wom en’s preg-nancy plans, and we find low lev els of fa tal is tic be liefs and lack of plan ning for fu ture preg nan cies. Research dem on strat ing high lev els of fa tal ism has been based on ret ro-spec tive ques tions about past preg nan cies (Borrero et al. 2015; Hodgson et al. 2013) or gen eral ques tions that do not re fer to spe cific preg nan cies (Jones et al. 2015; Rocca and Harper 2012; Woodsong et al. 2004). Although our semi-struc tured in ter views were cross-sec tion al—that is, they could not as sess change over time—the women

Dow




who were pro spec tively de scrib ing feel ings about a po ten tial preg nancy were much less likely to be fa tal is tic or not plan ning their next preg nancy than the women who were ret ro spec tively recalling their pre con cep tion feel ings about a preg nancy that ac tu ally oc curred. Consistent with re search on in fer til ity (Bell and Hetterly 2014), we spec u late that some young women who are cur rently rais ing chil dren born from un de sired preg nan cies use fa tal ism about past preg nan cies as a means of cop ing with hav ing not got ten what they wanted.

Limitations of the RDSL Sample

The RDSL sam ple has im por tant lim i ta tions. The RDSL had a nar row geo graphic fo cus (a sin gle county in Michigan), and the sam ple was not na tion ally rep re sen ta tive; how ev er, Michigan falls around the na tional me dian in mea sures of co hab i ta tion, mar riage, age at first birth, com pleted fam ily size, non mar i tal child bear ing, and teen age child bear ing (see Ela and Budnick 2017 for a com par i son to the National Survey of Family Growth; also see Lesthaeghe and Neidert 2006). More im por tant, the county has a large Black pop u la-tion (about 35%), and the pro por tion of res i dents who are Black in the ma jor city within the county is even higher. The United States has 65 cit ies that are at least 25% Black, representing at least 10 mil lion of the 39 mil lion Black res i dents in the United States. Thus, the women in the RDSL sam ple live in ar eas with sim i lar ra cial com po si tion as the neigh bor hoods of many Black peo ple in the United States. On the other hand, the study in cluded only a small num ber of La ti nas, who were clas si fied as ei ther White or Black in our an a ly ses—a lim i ta tion that we hope mo ti vates fu ture re search ers to im ple ment sim i-lar stud ies on larger and more di verse pop u la tions.

Our semi-struc tured in ter view re spon dents rep re sent spe cific ex pe ri ences: women who were preg nant in their late teens or early 20s, and women who had a high pro-pen sity for preg nancy but avoided it dur ing the study pe ri od. Many women in both groups also al ready had chil dren. Thus, in ter views with these two sub groups may not gen er al ize to the views or ex pe ri ences of all young wom en. These women may have thought more care fully about their preg nancy de sires and plans than women who did not have young births. Given their young births, they may also be the women who were least likely to plan and most likely to have fa tal is tic feel ings about ret ro spec-tive preg nan cies. It is also pos si ble that Black and White women were dif fer en tially se lected into this group, and ra cial dif fer ences in the pop u la tion as a whole may be dif fer ent from those ob served in this sam ple.

Feelings about preg nancy are highly re lated to wom en’s age, and the age dis-tri bu tion of preg nan cies dif fers by race. Although 24% of the Black women in our sam ple be came preg nant dur ing the study pe ri od, com pared with 18% of White wom en, the per cent age of preg nan cies that Black and White women ret ro spec tively recalled as un de sired was sim i lar: 83% and 81%, re spec tive ly. This is con sis tent with other re search show ing that the Black-White dis par ity in (ret ro spec tive) un de-sired preg nancy is par tially explained by age: youn ger preg nan cies are more likely to be re mem bered as un de sired, and Black women have youn ger preg nan cies, on av er age (Kim et al. 2016). Thus, there may be larger race dif fer ences in preg nancy de sire or plan ning at ages even youn ger than the RDSL sam ple. Race disparities

Dow




are likely to be dif fer ent at older ages, as well. Future re search should con tinue to ex am ine this im por tant ra cial dis par i ty, at both youn ger and older ages.

Conclusion

If un de sired births oc cur among women who can not or do not want to plan their preg nan cies, then im pos ing a “plan ning par a digm” on all women could be in ap pro-pri ate (Aiken et al. 2016). However, the vast ma jor ity of young women in our an a-ly ses were quite spe cific and con sis tent about their fu ture child bear ing de sires and plans. Although Bachrach and Mor gan (2013: ab stract) spec u lated that peo ple “do not nec es sar ily have fer til ity in ten tions,” but rather “form them only when prompted by spe cific sit u a tions,” few Black or White young women had not thought about their fu ture child bear ing plans be fore we talked to them. On the con trary, most of them read ily pro vided spe cific plans for delaying their next birth or stop ping child bear ing al to geth er, and their pro spec tive de sires in the weekly sur veys were re mark ably con-sis tent over time. But we did find a small num ber of women who did not have spe-cific plans for the fu ture, or who held am biv a lent or in dif fer ent short-term de sires for preg nan cy. Thus, we agree with Aiken and col leagues (2016) that some women may sim ply not want to plan ahead for preg nan cy, or they may want to leave a ran dom el e ment to when they get preg nant. However, the “plan ning par a digm” ap pears to be ap pro pri ate for the vast ma jor ity of young women and does not ap pear to be dif fer en-tially ap pli ca ble to Black and White wom en.

If per cep tions about race dif fer ences in the ap pli ca bil ity of a plan ning par a digm per-sist, these per cep tions may bias re search, in ter ven tion, and clin i cal prac tice. Assuming that Black women form in ter nally in con sis tent de sires re lated to preg nancy or that they do not want to plan their preg nan cies is likely to ex ac er bate ra cial disparities in un de sired preg nancy by fa cil i tat ing White wom en’s child bear ing de sires and plans more than Black wom en’s child bear ing de sires and plans. Our re search chal lenges these po ten tial im plicit biases by ex plic itly dem on strat ing that there is lit tle, if any, dif-fer ence in the ap pli ca bil ity of such a par a digm in our sam ple. Future re search should fur ther in ves ti gate—in dif fer ent set tings, with dif fer ent sub pop u la tions, and among women of dif fer ent ages—whether the con cept of preg nancy de sire and the re lated “plan ning par a digm” are equally ap pro pri ate for all wom en. ■

Acknowledgments This re search was supported by Grants R03 HD077051 and R01 HD078412 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) as well as NICHD grants to the University of Michigan’s Population Studies Center (R24 HD041028), Ohio State University’s Institute for Population Research (P2C-HD058484), and Bowling Green State University’s Center for Family and Demographic Research (R24 HD050959). The data used in this anal y sis were col-lected with the sup port of grants R01 HD050329 and R01 HD050329-S1 from NICHD. The au thors grate fully ac knowl edge Robert Melendez for pro gram ming as sis tance. We are also grate ful to the Survey Research Operations (SRO) unit at the Survey Research Center of the Institute for Social Research for their help with the data col lec tion, par tic u larly Vivienne Outlaw, Sharon Parker, and Meg Stephenson. The au thors also ac knowl edge the in tel lec tual con tri bu tions of the other mem bers of the orig i nal RDSL pro ject team, Heather Gatny, William Axinn, Mick Couper, Steven Heeringa, as well as the Advisory Committee for the pro ject: Larry Bumpass, Elizabeth Cooksey, Kathie Harris, and Linda Waite.

Dow




References

Aiken, A. R. A., Borrero, S., Callegari, L. S., & Dehlendorf, C. (2016). Rethinking the preg nancy plan-ning par a digm: Unintended con cep tions or un rep re sen ta tive con cepts? Perspectives on Sexual and Reproductive Health, 48, 147–151. https: / /doi .org /10 .1363 /48e10316

Aiken, A. R. A., & Potter, J. E. (2013). Are La tina women am biv a lent about preg nan cies they are try-ing to pre vent? Evidence from the Border Contraceptive Access Study. Perspectives on Sexual and Reproductive Health, 45, 196–203. https: / /doi .org /10 .1363 /4519613

Alba, R. D., Insolera, N. E., & Lindeman, S. (2016). Is race re ally so flu id? Revisiting Saperstein and Penner’s em pir i cal claims. Amer i can Journal of Sociology, 122, 247–262.

Bachrach, C. A., & Mor gan, S. P. (2013). A cog ni tive-so cial model of fer til ity in ten tions. Population and Development Review, 39, 459–485.

Bagozzi, R. P., & van Loo, M. F. (1978). Fertility as con sump tion: Theories from the be hav ioral sci ences. Journal of Consumer Research, 4, 199–228.

Barber, J. S., Kusunoki, Y., & Gatny, H. H. (2011). Design and implementation of an online weekly jour nal to study un in tended preg nan cies. Vienna Yearbook of Population Research, 9, 327–334.

Barber, J. S., Kusunoki, Y., Gatny, H., & Schulz, P. (2016). Participation in an in ten sive lon gi tu di nal study with weekly web sur veys over 2.5 years. Journal of Medical In ter net Research, 18(6), e105. https: / /doi .org /10 .2196 /jmir .5422

Bell, A. V., & Hetterly, E. (2014). “There’s a higher pow er, but He gave us a free will”: Socioeconomic sta tus and the in ter sec tion of agency and fa tal ism in in fer til i ty. Social Science & Medicine, 114, 66–72.

Bloome, D. (2014). Racial in equal ity trends and the in ter gen er a tional per sis tence of in come and fam ily struc ture. Amer i can Sociological Review, 79, 1196–1225.

Borrero, S., Nikolajski, C., Steinberg, J. R., Freedman, L., Akers, A. Y., Ibrahim, S., & Schwarz, E. B. (2015). “It just hap pens”: A qual i ta tive study ex plor ing low-in come wom en’s per spec tives on preg-nancy in ten tion and plan ning. Contraception, 91, 150–156.

Broman, C. L. (2005). Marital qual ity in Black and White mar riages. Journal of Family Issues, 26, 431–441.Bulatao, R. A., & Lee, R. D. (1983). A frame work for the study of fer til ity de ter mi nants. In R. A. Bulatao,

R. D. Lee, P. Hollerbach, & J. Bongaarts (Eds.), Determinants of fer til ity in de vel op ing countries (Vol. 1, pp. 1–26). New York, NY: Academic Press.

Burton, L. M. (1990). Teenage child bear ing as an al ter na tive life-course strat egy in multigeneration Black fam i lies. Human Nature, 1, 123–143. https://doiorg/10.1007/BF02692149

Burton, L. M., & Tucker, M. B. (2009). Romantic unions in an era of un cer tain ty: A post-Moynihan per-spec tive on Af ri can Amer i can women and mar riage. ANNALS of the Amer i can Academy of Political and Social Science, 621, 132–148.

Cacioppo, J. T., Gardner, W. L., & Berntson, G. G. (1999). The af fect sys tem has par al lel and in te gra tive pro-cessing com po nents: Form fol lows func tion. Journal of Personality and Social Psychology, 76, 839–855.

Chatters, L. M., Taylor, R. J., Bullard, K. M., & Jackson, J. S. (2009). Race and eth nic dif fer ences in re li gious in volve ment: Af ri can Amer i cans, Ca rib bean Blacks and non-His panic Whites. Ethnic and Racial Studies, 32, 1143–1163.

Cole, E. R. (2009). Intersectionality and re search in psy chol o gy. Amer i can Psychologist, 64, 170–180.Cutler, A., McNamara, B., Qasba, N., Kennedy, H. P., Lundsberg, L., & Gariepy, A. (2018). “I just don’t

know”: An ex plo ra tion of wom en’s am biv a lence about a new preg nan cy. Women’s Health Issues, 28, 75–81.

Dash, L. (2003). When chil dren want chil dren: The ur ban cri sis of teen age child bear ing. Champaign: University of Il li nois Press.

Edin, K., & Kefalas, M. J. (2005). Promises I can keep: Why poor women put moth er hood be fore mar riage (1st ed.). Berkeley: University of California Press.

Ela, E. J., & Budnick, J. (2017). Non-het ero sex u al i ty, re la tion ships, and young wom en’s con tra cep tive be hav ior. Demography, 54, 887–909.

Finer, L. B., & Zolna, M. R. (2016). Declines in un in tended preg nancy in the United States, 2008–2011. New England Journal of Medicine, 374, 843–852.

Fishbein, M., & Ajzen, I. (1975). Belief, at ti tude, in ten tion, and be hav ior: An in tro duc tion to the ory and re search. Reading, MA: Addison-Wesley.

Fletcher, J. M., & Wolfe, B. L. (2009). Education and la bor mar ket con se quences of teen age child bear ing ev i dence us ing the tim ing of preg nancy out comes and com mu nity fixed ef fects. Journal of Human Resources, 44, 303–325.

Dow



https://doi.org/10.1363/48e10316

https://doi.org/10.1363/4519613

https://doi.org/10.2196/jmir.5422

https://doi.org/10.2196/jmir.5422


Foster, D. G., Biggs, M. A., Ralph, L. J., Arons, A., & Brindis, C. D. (2008). Family plan ning and life plan-ning: Reproductive in ten tions among in di vid u als seek ing re pro duc tive health care. Women’s Health Issues, 18, 351–359.

Friedman, D., Hechter, M., & Kanazawa, S. (1994). A the ory of the value of chil dren. Demography, 31, 375–401.

Geronimus, A. T. (1987). On teen age child bear ing and neo na tal mor tal ity in the United States. Population and Development Review, 13, 245–279.

Geronimus, A. T. (1992). The weath er ing hy poth e sis and the health of Af ri can-Amer i can women and in fants: Evidence and spec u la tions. Ethnicity & Disease, 2, 207–221.

Geronimus, A. T. (2003). Damned if you do: Culture, iden ti ty, priv i lege, and teen age child bear ing in the United States. Social Science & Medicine, 57, 881–893.

Geronimus, A. T., Hicken, M., Keene, D., & Bound, J. (2006). “Weathering” and age pat terns of allostatic load scores among Blacks and Whites in the United States. Amer i can Journal of Public Health, 96, 826–833.

Geronimus, A. T., & Korenman, S. (1992). The so cio eco nomic con se quences of teen child bear ing recon-sidered. Quarterly Journal of Economics, 107, 1187–1214.

Gómez, A. M., Arteaga, S., Villaseñor, E., Arcara, J., & Freihart, B. (2019). The mis clas si fi ca tion of am biv-a lence in preg nancy in ten tions: A mixed-meth ods anal y sis. Perspectives on Sexual and Reproductive Health, 51, 7–15. https: / /doi .org /10 .1363 /psrh .12088

Higgins, J. A. (2017). Pregnancy am biv a lence and long-act ing re vers ible con tra cep tive (LARC) use among young adult wom en: A qual i ta tive study. Perspectives on Sexual and Reproductive Health, 49, 149–156. https: / /doi .org /10 .1363 /psrh .12025

Hodgson, E. J., Collier, C., Hayes, L., Curry, L. A., & Fraenkel, L. (2013). Family plan ning and con tra-cep tive de ci sion-mak ing by eco nom i cally dis ad van taged, Af ri can-Amer i can wom en. Contraception, 88, 289–296.

Hotz, V. J., McElroy, S. W., & Sanders, S. G. (2005). Teenage child bear ing and its life cy cle con se quences exploiting a nat u ral ex per i ment. Journal of Human Resources, 40, 683–715.

Jones, R. K., Frohwirth, L. F., & Blades, N. M. (2016). “If I know I am on the pill and I get preg nant, it’s an act of God”: Women’s views on fa tal ism, agency and preg nan cy. Contraception, 93, 551–555.

Jones, R. K., Tapales, A., Lindberg, L. D., & Frost, J. (2015). Using lon gi tu di nal data to un der stand changes in con sis tent con tra cep tive use. Perspectives on Sexual and Reproductive Health, 47, 131–139. https: / /doi .org /10 .1363 /47e4615

Kane, J. B., Mor gan, S. P., Harris, K. M., & Guilkey, D. K. (2013). The ed u ca tional con se quences of teen child bear ing. Demography, 50, 2129–2150.

Kearney, M. S., & Levine, P. B. (2012). Why is the teen birth rate in the United States so high and why does it mat ter? Journal of Economic Perspectives, 26(2), 141–166.

Kemet, S., Lundsberg, L. S., & Gariepy, A. M. (2018). Race and eth nic ity may not be as so ci ated with risk of un in tended preg nan cy. Contraception, 97, 313–318.

Kennelly, I. (1999). “That sin gle mother el e ment”: How White em ployers typ ify Black wom en. Gender & Society, 13, 168–192.

Kim, T. Y., Dagher, R. K., & Chen, J. (2016). Racial/eth nic dif fer ences in un in tended preg nan cy: Evidence from a na tional sam ple of U.S. wom en. Amer i can Journal of Preventive Medicine, 50, 427–435.

Klerman, L. V. (2000). The intendedness of preg nan cy: A con cept in tran si tion. Maternal and Child Health Journal, 4, 155–162.

Kost, K., & Lindberg, L. (2015). Pregnancy in ten tions, ma ter nal be hav iors, and in fant health: Investigating re la tion ships with new mea sures and pro pen sity score anal y sis. Demography, 52, 83–111.

Kost, K., Maddow-Zimet, I., & Kochhar, S. (2018). Pregnancy de sires and preg nan cies at the state lev-el: Estimates for 2014 (Report). New York, NY: Guttmacher Institute. Retrieved from https: / /www .guttmacher .org /report /pregnancy -desires -and -pregnancies -state -level -estimates -2014

Kost, K., & Zolna, M. (2019). Challenging un in tended preg nancy as an in di ca tor of re pro duc tive au ton-o my: A re sponse. Contraception, 100, 5–9. https: / /doi .org.10.1016/j.contraception.2019.04.010

Kreyenfeld, M. (2010). Uncertainties in fe male em ploy ment ca reers and the post pone ment of par ent hood in Germany. Eu ro pean Sociological Review, 26, 351–366.

Lee, D. (2010). The early so cio eco nomic ef fects of teen age child bear ing: A pro pen sity score matching ap proach. Demographic Research, 23, 697–736. https://doi.org/10.4054/DemRes.2010.23.25

Lesthaeghe, R. J., & Neidert, L. (2006). The sec ond de mo graphic tran si tion in the United States: Exception or text book ex am ple? Population and Development Review, 32, 669–698.

Dow



https://doi.org/10.1363/psrh.12088

https://doi.org/10.1363/psrh.12025

https://doi.org/10.1363/47e4615

https://doi.org/10.1363/47e4615

https://www.guttmacher.org/report/pregnancy-desires-and-pregnancies-state-level-estimates-2014

https://www.guttmacher.org/report/pregnancy-desires-and-pregnancies-state-level-estimates-2014

https://doi.org.10.1016/j.contraception.2019.04.010

https://doi.org/10.4054/DemRes.2010.23.25


Levine, J. A. (2013). Ain’t no trust: How bosses, boy friends, and bu reau crats fail low-in come moth ers and why it mat ters. Berkeley: University of California Press.

Lichter, D. T., Parisi, D., & Taquino, M. C. (2012). The ge og ra phy of ex clu sion: Race, seg re ga tion, and con cen trated pov er ty. Social Problems, 59, 364–388.

Lin coln, C. E., & Mamiya, L. H. (1990). The Black church in the Af ri can Amer i can ex pe ri ence (1st ed.). Dur ham, NC: Duke University Press.

Miles, M. B., & Huberman, A. M. (1984). Qualitative data anal y sis: A source book of new meth ods. Thou-sand Oaks, CA: Sage.

Miller, W. B. (1994). Childbearing mo ti va tions, de sires, and in ten tions: A the o ret i cal frame work. Genetic, Social & General Psychology Monographs, 120, 225–258.

Miller, W. B., Barber, J. S., & Gatny, H. H. (2013). The ef fects of am biv a lent fer til ity de sires on preg nancy risk in young women in the USA. Population Studies, 67, 25–38.

Mollborn, S. (2017). Mixed mes sages: Norms and so cial con trol around teen sex and preg nan cy. New York, NY: Oxford University Press.

Moos, M.-K., Petersen, R., Meadows, K., Melvin, C. L., & Spitz, A. M. (1997). Pregnant wom en’s per-spec tives on intendedness of preg nan cy. Women’s Health Issues, 7, 385–392.

Moreau, C., Hall, K. S., Trussell, J., & Barber, J. S. (2012). Effect of pro spec tively mea sured preg nancy in ten-tions on the con sis tency of con tra cep tive use among young women in Michigan. Human Reproduction, 28, 642–650.

Pager, D., Bonikowski, B., & Western, B. (2009). Discrimination in a low-wage la bor mar ket: A field ex per i ment. Amer i can Sociological Review, 74, 777–799.

Potter, J. E., Stevenson, A. J., Coleman-Minahan, K., Hopkins, K., White, K., Baum, S. E., & Grossman, D. (2019). Challenging un in tended preg nancy as an in di ca tor of re pro duc tive au ton o my. Contraception, 100, 1–4. https: / /doi .org /10 .1016 /j .contraception .2019 .02 .005

Rackin, H., & Mor gan, S. P. (2018). Prospective ver sus ret ro spec tive mea sure ment of un wanted fer til-i ty: Strengths, weaknesses, and inconsistencies assessed for a co hort of US wom en. Demographic Research, 39, 61–94. https: / /doi .org /10 .4054 /DemRes .2018 .39 .3

Raley, R. K., Sweeney, M. M., & Wondra, D. (2015). The grow ing ra cial and eth nic di vide in U.S. mar riage pat terns. Future of Children, 25(2), 89–109.

Rocca, C. H., & Harper, C. C. (2012). Do ra cial and eth nic dif fer ences in con tra cep tive at ti tudes and knowl edge ex plain disparities in method use? Perspectives on Sexual and Reproductive Health, 44, 150–158. https: / /doi .org /10 .1363 /4415012

Rosenthal, L., & Lobel, M. (2016). Stereotypes of Black Amer i can women re lated to sex u al ity and moth-er hood. Psychology of Women Quarterly, 40, 414–427.

Santelli, J. S., Lindberg, L. D., Orr, M. G., Finer, L. B., & Speizer, I. (2009). Toward a mul ti di men sional mea sure of preg nancy in ten tions: Evidence from the United States. Studies in Family Planning, 40, 87–100.

Santelli, J. S., Rochat, R., Hatfield-Timajchy, K., Gilbert, B. C., Curtis, K., Cabral, R., . . . Unintended Preg-nancy Working Group. (2003). The mea sure ment and mean ing of un in tended preg nan cy. Perspectives on Sexual and Reproductive Health, 35, 94–101. https: / /doi .org /10 .1363 /3509403

Saperstein, A., & Penner, A. M. (2012). Racial flu id ity and in equal ity in the United States. Amer i can Journal of Sociology, 118, 676–727.

Schacter, D. L. (1999). The seven sins of mem o ry: Insights from psy chol ogy and cog ni tive neu ro sci ence. Amer i can Psychologist, 54, 182–203.

Schoendorf, K. C., Hogue, C. J. R., Kleinman, J. C., & Rowley, D. (1992). Mortality among in fants of Black as com pared with White col lege-ed u cated par ents. New England Journal of Medicine, 326, 1522–1526.

Sennott, C., & Yeatman, S. (2018). Conceptualizing child bear ing am biv a lence: A so cial and dy namic per-spec tive. Journal of Marriage and Family, 80, 888–901.

Shavers, V. L., Fagan, P., Jones, D., Klein, W. M. P., Boyington, J., Moten, C., & Rorie, E. (2012). The state of re search on ra cial/eth nic dis crim i na tion in the re ceipt of health care. Amer i can Journal of Public Health, 102, 953–966.

Sobotka, T., Skirbekk, V., & Philipov, D. (2011). Economic re ces sion and fer til ity in the de vel oped world. Population and Development Review, 37, 267–306.

Stanley, D. J., & Meyer, J. P. (2009). Two-di men sional af fec tive space: A new ap proach to orienting the ax es. Emotion, 9, 214–237.

Dow



https://doi.org/10.1016/j.contraception.2019.02.005

https://doi.org/10.4054/DemRes.2018.39.3

https://doi.org/10.1363/4415012

https://doi.org/10.1363/3509403


Steensland, B., Robinson, L. D., Wilcox, W. B., Park, J. Z., Regnerus, M. D., & Woodberry, R. D. (2000). The mea sure of Amer i can re li gion: Toward im prov ing the state of the art. Social Forces, 79, 291–318.

Stones, R. E., Stulberg, D. B., & Bello Kottenstette, J. K. (2017). Patient ex pe ri ences with preg nancy plan-ning and per spec tives on re pro duc tive care in com mu nity health cen ters: A qual i ta tive study of Af ri can Amer i can women in Chicago. Women’s Health Issues, 27, 322–328.

Townes, B. D., Beach, L. R., Campbell, F. L., & Martin, D. C. (1977). Birth plan ning val ues and de ci sions: The pre dic tion of fer til i ty. Journal of Applied Social Psychology, 7, 73–88.

West, C. M. (2008). Mammy, Jez e bel, Sapphire, and their home girls: Developing an “op po si tional gaze” to ward the im ages of Black wom en. In J. Chrisler, C. Golden, & P. Rozee (Eds.), Lectures on the psy-chol ogy of wom en (Vol. 32, pp. 286–299). New York, NY: McGraw Hill.

Western, B., Bloome, D., Sosnaud, B., & Tach, L. (2012). Economic se cu rity and so cial strat i fi ca tion. Annual Review of Sociology, 38, 341–359.

Wilson, W. J. (2012). The truly dis ad van taged: The in ner city, the un der class, and pub lic pol i cy (2nd ed.). Chicago, IL: University of Chicago Press.

Woodard, J. B., & Mastin, T. (2005). Black wom an hood: “Essence” and its treat ment of ste reo typ i cal im ages of Black wom en. Journal of Black Studies, 36, 264–281.

Woodsong, C., Shedlin, M., & Koo, H. (2004). The “nat u ral” body, God and con tra cep tive use in the south-east ern United States. Culture, Health & Sexuality, 6, 61–78.

Yoo, S. H., Guzzo, K. B., & Hayford, S. R. (2014). Understanding the com plex ity of am biv a lence to ward preg nan cy: Does it pre dict in con sis tent use of con tra cep tion? Biodemography and Social Biology, 60, 49–66.

Jennifer S. Barber (cor re spond ing au thor)jenbarb@iu .edu

Barber • Department of Sociology and Kinsey Institute, Indiana University, Bloomington, IN, USA

Guzzo • Department of Sociology, Bowling Green State University, Bowling Green, OH, USA

Budnick • Population Studies Center and Center for Sexuality and Health Disparities, University of Michigan, Ann Arbor, MI, USA

Kusunoki • Department of Systems, Populations, and Leadership, School of Nursing, University of Michigan, Ann Arbor, MI, USA

Hayford • Department of Sociology, The Ohio State University, Co lum bus, OH, USA

Miller • Transnational Family Research Institute, Aptos, CA, USA

Dow



mailto:[email protected]

Documents

Black-White Differences in Pregnancy Desire During the