Mothers Living in Poverty:
How Uninvolved and Unreliable Men in their Lives May Drive them to Abortion and How Abortion Increases Risk for Adverse Outcomes

Association for Interdisciplinary Research in Values and Social Change
by Priscilla K. Coleman, Ph.D.
Volume 21, Number 1
Spring 2009
Reproduced with Permission

[Synopsis and highlights from: Coleman, P. K., Maxey, D. C., Spence, M. & Nixon, C. (Published in Online First, 12-31-08). The choice to abort among mothers living under ecologically deprived conditions: Predictors and consequences. International Journal of Mental Health and Addiction.]

Research has not revealed a single reliable profile of women who obtain abortions in the U.S. The reality is that women with widely ranging demographic and personal characteristics choose abortion for many reasons and under widely varying circumstances (Allanson & Astbury, 1995; Henshaw & Silverman, 1998; Kero, Hoegberg, Jacobsson, & Lalos, 2001; Patterson, Hill, & Maloy, 1995). However, when women are surveyed they will often describe common factors that figured into their decisions to terminate a pregnancy. Concerns with becoming a single parent and partner relationship problems are among the common motives for undergoing an abortion (Allanson, & Astbury, 2001; Soderberg, Andersson, Janzon, & Slosberg, 1997; Soderberg, Janzon, & Sjoberg, 1998). Fear that carrying a pregnancy to term will interfere with the continuation of one's current intimate relationship, future plans (Allanson & Astbury, 1995; Faria, Barrett, & Goodman, 1985; Patterson et al., 1995), young age (Faria et al., 1985), inadequate finances (Faria et al.,1985; Glander, Moore, Michielutte, & Parsons, 1998), and feeling one lacks the time and energy for another child (Kero et al., 2001) are other frequently cited reasons for aborting.

The literature devoted to identifying the reasons women decide to abort a pregnancy has unfortunately neglected to consider women's expectations regarding paternal involvement (material and psychological) in childcare. Such expectations are likely to play a vital role in abortion decision-making, particularly when the woman is not married to the father and may be unsure of his commitment to the relationship. Maternal expectations of paternal involvement are also inclined to factor strongly into women's abortion decisions when they are living in poverty and paternal financial support is viewed as essential for basic childcare needs. Information regarding paternal commitment that is available during pregnancy includes the father's verbalized interest in the pregnancy and his verbalized commitment to supporting the child after birth. Other sources of information regarding the father's willingness to become involved in childcare include the extent to which the father has taken on paternal responsibilities with another child in addition to personal assessments of the quality of the relationship between the woman and her partner.

Women's decisions to abort vs. deliver may also be associated with their perceptions of previous or current difficulties with their children. Women who are contemplating an abortion weigh the pros and cons of continuing the pregnancy (Allanson & Astbury, 1996) and without active support from the child's father and with difficulties handling existing children, the likelihood that a woman will decide to abort may increase. No earlier studies have examined aspects of child difficulty associated with previously born children as predictors of the choice to abort.

In this newsletter, the results of a study examining maternal assessments of various indicators of paternal involvement and maternal perceptions of difficulties incurred with raising a recently born child as predictors of the choice to abort a subsequent pregnancy are summarized. The examination of paternal and child factors associated with abortion represents a significant addition to the abortion decision-making literature, as no previous studies have focused directly on family members' behavior and relationships to the woman in attempts to understand why women choose abortion.

This study was specifically designed to identify predictors of the choice to abort or deliver a child within 18 months of a previous birth and to compare mothers who chose to abort or deliver relative to substance use and adverse partner behavior. Data from the Fragile Families and Well-Being Study were employed, with the sample including participants from hospitals with high rates of unmarried births in 16 large cities across the U.S. The women were predominantly young, low-income, and not married. The data analysis process involved two steps. First, predictors of the choice to abort were identified statistically. Second, mothers who chose to abort and deliver were compared relative to heavy drinking, cigarette smoking, and adverse partner behavior after controlling for statistically significant predictors identified in step one. Controls were therefore instituted for sociodemographic, mother, child, father/partner relationship variables found to be associated with the decision to abort. No previous studies have explored an extensive list of potential predictors of the choice to abort vs. deliver, nor have the few available relative risk studies comparing the effects of abortion and delivery on the subsequent well-being of women incorporated a wide array of potential covariates.

Study Details

The Fragile Families and Child Wellbeing Study served as the data source for this archival study. The study is a collaborative effort by principal investigators associated with Princeton University and Columbia University and the study is funded by the National Institute of Child Health and Development, the National Science Foundation, the U.S. Department of Health and Human Services, and the Office of Population Research at Princeton University. The study was designed to address non-marital childbearing, welfare reform, and paternal behavior. The public use data are available free for download from the Princeton University Office of Population Research data archive.

The baseline and one-year follow-up interviews were conducted from February 1998 to September 2000 and from June 1999 to March 2002 respectively. Initial interviewing was done in hospitals as opposed to using birth records and interviewing in homes in order to increase the response rate and maintain costs. Approximately one third of the one-year follow-up interviews were administered over the phone with the remaining two-thirds conducted in-person.

Although the study contains both paternal and maternal selfreport data, this report only used the maternal data. Exactly 3,119 women were included in the national sample and completed both the baseline (soon after giving birth) and the one-year followup surveys. Baseline data for the subs-sample used for the present study were collected within days of delivering a child and the follow-up data were gathered 12-18 months later. The first child of the women was between the ages of 12 and 18 months during the second wave of data collection. The women who terminated a second pregnancy did so at any time during the interval between baseline and followup; whereas those who chose to continue their pregnancies had a second child who was 8 months or younger at the time of follow-up. The sub-sample for the current study was comprised only of women who had another pregnancy and assumed care of the child (n=130) or aborted the pregnancy (n=112) prior to the follow-up assessment. For 76.9% of the women who delivered a second child, the same man had fathered the child born at baseline. Table 1 provides the basic demographic information for the national sample and for the subsample used for the purpose of this study. As indicated by the data presented, a majority of the participants had low incomes (under $35,000), were Black, unmarried, and had 12 or fewer years of education. Those not reporting income information (26%) were likely in the lower income groups as the original sampling design identified families considered at risk.

Table 1
Demographic characteristics of the participants in the Fragile Families and Wellbeing Study and for the current sub-sample
Characteristic National Sample
Marital Status
Married 24% 03%
Not Married 76% 87%
White 45% 24%
Black 28% 56%
Hispanic 14% 15%
Asian 6% 1%
American Indian 5% 4%
less than 12 yrs 33% 41%
12 yrs or GED 30% 31%
12 or more yrs 37% 28%
less than 20 18% 27%
20-29 59% 64%
30-39 21% 8%
40 or older 2% 1%
Annual Income
less than $15,000 22% 32%
$15,000-24,999 13% 13%
$25,000-34,999 9% 7%
$35,000-49,999 13% 12%
$50,000-74,999 14% 6%
$75,000 or more 11% 4%
Refused, did not know, or missing. 18% 26%


Predictors of the Choice to Abort

The results of analyses conducted to identify significant predictors of the choice to abort are provided in Table 2. Single parenthood, having considered an abortion for the first pregnancy, and the mother-father relationship not having improved since the pregnancy was confirmed were associated with a significantly increased likelihood of abortion. Low levels of father-child involvement of various forms measured at the one-year follow-up were also found to be associated with a significant increase in the likelihood of abortion. There were 12 dichotomous father involvement variables explored as predictors of the choice to abort and only one was not identified as a significant predictor.

None of the child variables were significant predictors of the choice to abort. In other words, being the mother of a young child who was behaviorally or physically challenging did not play a role in the decision to abort. The following socio-demographic variables were also examined as predictors of the choice to abort: age, race, employment, education, and income. None of these variables were found to discriminate between the choice to abort and deliver. Finally, maternal substance use variables and paternal aggressive behavior variables were explored as predictors of the choice to abort, because significant associations would indicate the need to control them when exploring possible effects of abortion. No significant relations were observed between these variables and the choice to abort.

Comparisons Between the Abortion and Delivery Groups

As indicated by the data presented in Table 3 below, women who terminated a second pregnancy when compared to women who delivered a second time, were over three times more likely to report recent heavy use of alcohol (consumption of 5 or more drinks on one day in the past 30 days) and they were nearly twice as likely to report recent cigarette smoking (in the past 30 days). These results are with employment of statistical controls for the 14 significant predictors of the choice to abort.

Table 2
Baseline and follow-up variable as predictors of the choice to abort
Variable Odds Ratio, p, 95% CI* Abortion Delivery
Marital Status
Mother considered an abortion during initial pregnancy 2.00, p=.016, 1.14-3.51 40.0% 25.0%
Relationship with father had not improved since pregnancy confirmed 1.80, p=.03, 1.06-3.07 65.5% 51.3%
Mother and father married at birth .425, p=.043, .19-.87 8.1% 17.2%
Mother's age (25 or younger) Ns
Father's age (25 or younger) Ns
Mother's race (non-white) Ns
Father's race (non-white) Ns
Mother's level of formal education Ns
Father's level of formal education Ns
Mother has other biological children Ns
Father has other biological children Ns
Father not currently working Ns
Mother not currently working Ns
Income below $15,000 Ns
Mother used alcohol during pregnancy Ns
Mother smoked cigarettes during pregnancy Ns
Mother was ever treated for substance use Ns
Father did not give money to buy things for infant Ns
Father did not help in any way to prepare for infant Ns
Father did not say he would help support baby Ns
Father is never fair and willing to compromise Ns
Father hits/slap mother when angry Ns
Father never expresses affection Ns
Father insults and criticizes motherŐs ideas Ns
Father is never fair and willing to compromise Ns
Father hits/slaps mother when angry Ns
Follow-up variables
Father cannot be counted on to watch child for a week 7.58, p=.002, 2.15-26.74 17.5% 2.7%
Mother cannot trust father to take good care of child 6.08, p=.022, 1.30-28.47 8.9% 1.8%
Father does not act like the type of father mother wants for her child 5.89, p=.007, 1.64-21.17 14.3% 2.8%
Father does not support motherŐs way of raising child 4.13, p=.004, 1.58-10.82 19.4% 5.5%
Mother and father do not talk about problems raising child 3.51, p=.007, 1.40-8.75 19.6% 6.4%
Father cannot be counted on to watch child for a few hours 3.39, p=.003, 1.53-7.49 25.5% 9.2%
Father does not respect the schedule/rules mother establishes for child 2.71, p=.021, 1.16-6.31 19.6% 8.3%
Father does not watch child when mother needs to do things 2.67, p=.003, 1.41-5.06 35.3% 7.0%
Father does not run errands for mother 1.90, p=.037, 1.04-3.47 35.3% 22.3%
Father does not take child to daycare, doctor, etc 1.88, p=.027, 1.08-3.28 46.1% 31.3%
Father does not make improvements around the house 1.86, p=.038, 1.03-3.33 38.2% 25.0%
Mother has asked father to spend more time with child Ns
3 or more child visit(s) to professional for illness since birth Ns
1 or more child visit(s) to professional for injury since birth Ns
1 or more child visit(s) to emergency room since birth Ns
1 or more child hospital stay(s) since birth Ns
Child is shy Ns
Child is fussy Ns
Child gets upset easily Ns
Child reacts strongly when upset Ns
(*** Ns = Statistically non-significant predictor)

Table 3
Five Recent Representative Key Studies Linking Induced Abortion and Preterm Birth
Variable Percentage reporting "yes" by form
of pregnancy resolution
Odd Ratio p
Alcohol (5 or more alcoholic beverages consumed on one day in the last month) Delivery 17.2
Abortion 54.5
*3.39 p=.001
Smoked cigarettes in the last 30 days Delivery 23.0
Abortion 42.0
*1.99 p=.050
*Delivery group served as the reference group, with controls for the following variables: mother and father married at baseline, mother considered an abortion during initial pregnancy, and relationship with father got worse or remained the same after initial pregnancy was confirmed, and 11 variables related to paternal involvement in the care of the child born at baseline.

In a final analysis, women with a history of abortion were also more likely to report having been slapped or kicked by the child's father compared to women who delivered a second pregnancy, F (7,126) = 3.90, p=.050. This result was after using all the same controls employed in the previous analysis.

Analysis of Findings and Conclusions

The results of this study clearly suggest that women who felt their first child's father had not assumed enough paternal responsibility and/ or lacked the ability to contribute to their efforts to raise the child, were reluctant to bear another child. Professionals who work with low-income mothers confronting a pregnancy decision soon after giving birth should be encouraged to explore strategies for helping fathers assume a more active role in the lives of their children when such involvement is deficient and when greater involvement would be in the family's best interest. If the father is psychologically and/or physically unavailable, counselors can assist women in identifying other sources of support within and outside the family. Women who are successfully assisted in their efforts to secure more assistance are freer to make a decision regarding an unplanned pregnancy, which is consonant with their own desires. For a majority of the current sample (over 75%), the same man was involved in the two pregnancies resulting in births, indicating that most of the women were in relatively stable relationships. Additional research in this area might explore partner predictors of the choice to abort in a sample wherein the relationships are less stable.

The results also demonstrated that women who have recently undergone an abortion are more prone to substance use than women who have recently delivered a child, possibly due to negative emotions incurred from the abortion. Existing data do indicate that women with a history of abortion, compared to women without such a history, are more inclined to experience a wide range of negative emotions (Coleman et al. 2005; Gould, 1980; Franco et al., 1989; Moseley et al., 1981). This interpretation is supported by the fact that substance use during the initial pregnancy (alcohol, tobacco) as well as ever having been treated for substance use were not systematically related to the choice to abort. Birth may also serve as a protective experience against substance use, since many women who have recently had a child are likely to be in the habit of taking better care of their health. Conversely, women who terminated may feel freer to engage in substance use.

Because physical violence as measured at baseline was not found to be predictive of the choice to abort, the post-abortion enhanced risk for violence detected in this study may have been attributed to abortion-related stress introduced into the relationship and/or elevated psychological stress in one or both partners. For example, a man may become very upset that his "baby" was aborted against his will. Alternatively, the woman may suffer from depression if she chose abortion for situational as opposed to personal reasons and she may become withdrawn, triggering anger in a man prone to violent behavior. Another possibility is that for some women the physical violence began after the birth of the first child and prior to the second pregnancy, thereby factoring into the decision to abort.

The strengths of this study include the use of data from a carefully selected national sample of women, professional data collection, a longitudinal design, and the use of controls for several potentially confounding factors related to the choice to abort. Perhaps the most significant contribution is the new insight regarding the centrality of paternal behavior in women's abortion decision-making. Relational dimensions of abortion decisionmaking and adjustment may seem obvious, but up until this point in time, abortion has tended to be studied individually as opposed to relationally. Hesitation among social scientists to broaden the scope of attention to include partners and other family members is likely due to the framing of abortion as a private women's issue in contemporary society and/or the methodological complexities introduced by broadening the focus.

In conclusion, these results should challenge professionals to spend more time with women contemplating an abortion. Healthcare professionals should specifically assess pregnant women for perceptions of physical and emotional support, current domestic violence, and future risk in order to determine the safety inherent in the partnership while helping women arrive at a decision that will be most adaptive for them given their current life challenges. Inquiries about a history of prior or current substance abuse and education efforts regarding documented substance abuse risks associated with the choice to abort ought to be conveyed.

[Pricilla Coleman, Ph. D. is Associate Professor of Human Development and Family Studies at Bowling Green State University, Bowling Green, Ohio. She is also Co-director of the Alliance for Post-Abortion Research and Training.]