Does Father Absence Place Daughters at Special Risk For Early Sexual Activity and Teenage Pregnancy?

Does Father Absence Place Daughters at Special Risk for Early Sexual Activity and Teenage
Pregnancy?
Author(s): Bruce J. Ellis, John E. Bates, Kenneth A. Dodge, David M. Fergusson, L. John
Horwood, Gregory S. Pettit and Lianne Woodward
Source: Child Development, Vol. 74, No. 3 (May - Jun., 2003), pp. 801-821
Published by: Wiley on behalf of the Society for Research in Child Development
Stable URL: http://www.jstor.org/stable/3696230
Accessed: 14-08-2016 16:10 UTC
REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/3696230?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact [email protected].
Wiley, Society for Research in Child Development are collaborating with JSTOR to digitize,
preserve and extend access to Child Development
This content downloaded from 134.117.10.200 on Sun, 14 Aug 2016 16:10:08 UTC
All use subject to http://about.jstor.org/terms
Child Development, May/June 2003, Volume 74, Number 3, Pages 801-821
Does Father Absence Place Daughters at Special Risk for Early Sexual Activity
and Teenage Pregnancy?
Bruce J. Ellis, John E. Bates, Kenneth A. Dodge, David M. Fergusson, L. John Horwood,
Gregory S. Pettit, and Lianne Woodward
The impact of father absence on early sexual activity and teenage pregnancy was investigated in longitudinal
studies in the United States (N = 242) and New Zealand (N = 520), in which community samples of girls were
followed prospectively from early in life (5 years) to approximately age 18. Greater exposure to father absence was
strongly associated with elevated risk for early sexual activity and adolescent pregnancy. This elevated risk was
either not explained (in the U.S. study) or only partly explained (in the New Zealand study) by familial,
ecological, and personal disadvantages associated with father absence. After controlling for covariates, there was
stronger and more consistent evidence of effects of father absence on early sexual activity and teenage pregnancy
than on other behavioral or mental health problems or academic achievement. Effects of father absence are
discussed in terms of life-course adversity, evolutionary psychology, social learning, and behavior genetic models.
In modern Western societies, adolescent girls face a trialized countries. Approximately 10% of girls in
biosocial dilemma. On the one hand, the biological the United States and 7% of girls in New Zealand
capacity to reproduce ordinarily develops in early between the ages of 15 and 19 years become
adolescence; on the other hand, girls who realize this pregnant each year, with around half of these
capacity before adulthood often experience a variety pregnancies culminating in a live birth (Chees-
of negative life outcomes. Specifically, adolescent brough, Ingham, & Massey, 1999; Dickson, Sporle,
childbearing is associated with lower educational Rimene, & Paul, 2000). Given these costs to
and occupational attainment, more mental and adolescents and their children, it is critical to identify
physical health problems, inadequate social support life experiences and pathways that place girls at
networks for parenting, and increased risk of abuse increased risk for early sexual activity and adoles-
and neglect for children born to teen mothers (e.g., cent pregnancy.
Furstenberg, Brooks-Gunn, & Chase-Lansdale, 1989; Many studies have identified the absence of the
Konner & Shostak, 1986; Woodward & Fergusson, biological father from the home as a major risk factor
1999). Despite these consequences, the United States for both early sexual activity (e.g., Day, 1992;
and New Zealand have the first and second highest Kiernan & Hobcraft, 1997; Newcomber & Udry,
rates of teenage pregnancy among Western indus- 1987) and teenage pregnancy (e.g., Geronimus &
Korenman, 1992; Hogan & Kitagawa, 1985; McLa-
nahan, 1999). This finding is consistent with life-
Bruce J. Ellis, Department of Psychology, University of Canter-
course
adversity models of early sexual activity and
bury; John E. Bates, Department of Psychology, Indiana Uni-
teenage pregnancy, which posit that a life history of
versity; Kenneth A. Dodge, Center for Child and Family Policy,
Duke University; David M. Fergusson and L. John Horwood,familial and ecological stress provokes earlier onset
of sexual
Department of Psychological Medicine, Christchurch School of activity and reproduction (e.g., Belsky,
Steinberg, & Draper, 1991; Coley & Chase-Lansdale,
Medicine; Gregory S. Pettit, Department of Human Development
and Family Studies, Auburn University; Lianne Woodward,
1998; Fergusson & Woodward, 2000a; Robbins,
Department of Education, University of Canterbury.
In the United States, this work was supported by National
Kaplan, & Martin, 1985; Scaramella, Conger, Simons,
Institute of Mental Health grants MH28018 and MH42498 & and
Whitbeck, 1998). Life-course adversity models,
National Institute of Child Health and Human however,
Development grant do not attribute any special causal sig-
nificance
HD30572. In New Zealand, this work was supported by the to father absence. Instead, these models
Health Research Council, National Child Health Research
conceptualize father absence as just one of many
Foundation, the Canterbury Medical Research Foundation, and
factors that can undermine the quality of family
the New Zealand Lottery Grants Board. We thank Jay Belsky,
Ronald Dahl, and Satoshi Kanazawa for comments on earlier environments. According to life-course adversity
drafts of this article. models, it is not father absence per se but various
Correspondence concerning this article should be addressed to
Bruce Ellis, Department of Psychology, University of Canterbury,
Private Bag 4800, Christchurch, New Zealand. Electronic mail may
? 2003 by the Society for Research in Child Development, Inc.
be sent to: [email protected]. All rights reserved. 0009-3920/2003/7403-0010
802 Ellis et al.
other stressors associated with father absence (e.g.,

with Hetherington's (1972) work on the effects of
divorce, poverty, conflictual family relationships,
early father absence on personality development in
erosion of parental monitoring and control) that
adolescent daughters, the evolutionary model sug-
foster early sexual activity and pregnancy gests that in
girls detect and internally encode informa-
tion about parental reproductive strategies during
daughters (see Belsky et al., 1991, p. 658; Chisholm,
1999, p. 162; McLanahan, 1999, p. 119; Robbins et al.,
approximately the first 5 years of life as a basis for
1985, p. 568; Silverstein & Auerbach, 1999,calibrating
p. 403). the development of motivational sys-
tems, which make certain types of sexual behavior
In addition to the effects of life-course adversity,
underlying personality traits may accountmore for or
theless likely in adolescence. The model thus
relation between father absence and early sexual
posits a direct effect of quality of early paternal
outcomes in daughters. Specifically, certain person-(e.g., father presence vs. absence, quality
investment
ality traits that predispose girls toward early sexual
of paternal care giving, father-mother relationships)
activity and teenage pregnancy may covary on earlywith
onset of sexual and reproductive behavior.
father absence. Differences between children in In light of these theoretical considerations, the
current research examined the following set of
externalizing behavior problems-those behaviors
considered to be aggressive, disruptive, or opposi-
questions:
tional-derive in part from individual differencesGoals
in of the Current Research
temperamental characteristics such as negative 1. Is earlier onset of biological father absence
associated with increasing risk of early sexual
emotionality and resistance to control (Bates, Pettit,
Dodge, & Ridge, 1998; Rothbart & Bates, 1998). activity and teenage pregnancy in daughters?
Children who display externalizing behavioral pro- Despite voluminous research on father absence,
very
blems early in life are at elevated risk for a variety of few studies have examined the relation between
negative psychosocial outcomes in adolescence,
timing of onset of father absence and daughters'
including early sexual activity and teenage sexual
preg- outcomes. In a small observational study,
nancy (e.g., Bardone, Moffitt, Caspi, Dickson, &
Hetherington (1972) found that adolescent girls from
Silva, 1996; Quinton, Pickles, Maughan, & Rutter,
early father-absent homes (divorced before age 5)
1993; Woodward & Fergusson, 1999). Moreover,tended to initiate more contact with, and seek more
individuals who have a history of externalizing
attention from, adult males than did girls from late
disorders are not only at increased risk of becoming
father-absent homes (divorced after age 5). In a large
single parents or absent parents (e.g., Emery, retrospective survey, however, McLanahan (1999) did
Waldron, Kitzmann, & Aaron, 1999; Sampson not &find statistically significant relations between
Laub, 1990) but also may transmit a genetic disposi-
timing of onset of father absence and rates of teenage
tion toward externalizing behavioral problemschildbearing
and in daughters. The current research is the
associated personality characteristics to their chil-
first to measure prospectively the timing of onset of
dren (Rhee & Waldman, 2002; personality character-
father absence throughout early and middle child-
istics associated with both sexual risk taking hood
and and then test for its effects on early sexual
other forms of delinquent behavior in adolescence activity and pregnancy in adolescence.
are discussed in Kotchick, Shaffer, Forehand, & 2. Does earlier onset of biological father absence
Miller, 2001). Thus, girls from father-absent homes uniquely increase risk for early sexual activity and
may be at elevated risk for early sexual activity and adolescent pregnancy in daughters, independent of
teenage pregnancy because of higher genetic loading both early externalizing behavior problems and
for externalizing behavior problems. familial and ecological stressors that covary with
In contrast to the life-course adversity andfather absence? That is, does more exposure to father
personality trait models, evolutionary models sug- absence place daughters at special risk for early
gest that early onset of father absence placessexual outcomes-regardless of whether girls are
daughters at special risk for early sexual activity rich or poor, Black or White, cooperative or defiant
and adolescent pregnancy. Specifically, evolutionary in kindergarten, born to teenage or adult mothers,
psychologists have hypothesized that the develop-grow up in violent or safe neighborhoods, experi-
mental pathways underlying variation in daughters' ence many or few stressful life events, have warm-
reproductive strategies are especially sensitive to the supportive or harsh-rejecting parents, are exposed to
father's role in the family and the mothers' sexual functional or dysfunctional marriages, are closely or
attitudes and behavior in early childhood (Draper & loosely monitored by parents, and so forth?
Harpending, 1982, 1988; see also Ellis, McFadyen- A number of studies have found that father
Ketchum, Dodge, Pettit, & Bates, 1999). Consistent absence uniquely predicts early sexual activi
Father Absence 803
(Day, 1992; Devine, Long, & Forehand, 1993; Miller Method: United States
et al., 1997; Upchurch, Aneshensel, Sucoff, & Levy-
Storms, 1999) and adolescent pregnancy or child-
Participants and Overview
bearing (Hogan & Kitigawa, 1985; Robbins et al.,
1985), after controlling for such confounding The United States data were collected as part of
vari-
the ongoing Child Development Project, a multisite
ables as race, socioeconomic status (SES), neighbor-
hood danger, and parental monitoring and longitudinal
control. study of socialization factors in chil-
All of these studies, however, began when daughters dren's and adolescents' adjustment (see Dodge,
were already in early to late adolescence Bates, & Pettit, 1990; Pettit, Bates, & Dodge, 1997).
and thus
were unable to assess familial and ecological Participating families were initially recruited from
stressors before daughters' risk for involvement three geographical
in areas (Nashville and Knoxville,
sexual activity. The current research is theTennessee, first to and Bloomington, Indiana). At the time
assess prospectively life-course adversity through- of kindergarten preregistration in the summers of
out early and middle childhood, and control 1987
for(Cohort
its 1) and 1988 (Cohort 2), parents of
effects when testing for the relation between matriculating
timing children were solicited at random (in
of father absence and rates of early sexual person at the child's school or by mail) to become
activity
and adolescent pregnancy. involved in the study. About 75% agreed. A total of
3. Does earlier onset of biological father 585 families agreed to participate in the study. Of
absence discriminantly increase risk for early onset these 585 families, 281 of the children were girls. The
of sexual activity and teenage pregnancy-but analyses reported in this article are based on this
female subsample, which was demographically
not for adolescent behavioral and mental health
problems more generally-independent of early and representative of the geographic regions
diverse
(81% White, 17% African American, 2% other; 28%
externalizing problems and life-course adversity?
In other words, is greater exposure to lived
father
with a single mother at the beginning of the
absence a general risk factor for the develop-The Hollingshead (1975) Four-Factor Index
study).
ment of psychopathology, or is it specific toofsexual
Social Status was computed from demographic
development? information provided by the parents of the girls. The
To our knowledge, only Newcomer and Udry mean family score on the index at the beginning of
the study
(1987) have explicitly addressed this question. In a was 38.85 (SD=14.0), indicating a
predominantly
short-term longitudinal study of White adolescents, middle-class sample. Data on girls'
early
Newcomer and Udry found that the effect of father externalizing behavioral problems and on
familial and ecological stressors were collected in
absence on a composite measure of age-graded
minor delinquencies (e.g., smoking, drinking Years 1 through 9 of the study (ages 5-13). Data on
alco-
adolescent
hol, cheating on a test) was statistically significant sexual activity, pregnancy, internalizing
and
and about equal in magnitude to the effect of fatherexternalizing behavioral problems, academic
performance,
absence on onset of first sexual intercourse in girls. and violence were collected in Years 10
through
Newcomber and Udry, however, did not control for 13 of the study (ages 14-17). At the
potentially confounding third variables (e.g.,completion
race, of the study in Year 13, the average age
of the girls was 17.3 years (SD = .34). Of the original
SES, mother's age at first birth) that could account
281 girls, 242 (86%) participated in the Years 10
for the correlation between father absence and
delinquency. The current research examined through
the 13 data collections. This subset was gen-
unique effects of timing of father absence erally
on representative
a of the original sample (16%
African American; 25% from single-mother homes;
variety of psychosocial and educational outcomes,
after controlling for the effects of child mean SES = 39.45). Other analyses have shown that
conduct
problems and familial and ecological stressors
attrition has not significantly biased the sample on
during childhood. either initial child adjustment or family socialization
This set of questions was investigated in variables
two (see Pettit et al., 1997; Pettit, Bates, Dodge,
independent longitudinal studies in the United 1999). Nonetheless, there was a slight but
& Meece,
States and New Zealand. In the U.S. study, a statistically nonsignificant trend for the 242 girls in
community sample of girls was followed prospec-the current analyses to underrepresent girls from
tively from the summer before kindergarten through socially disadvantaged backgrounds (low SES, Afri-
to the 12th grade. In the New Zealand study, a birth can American, single-mother homes).
cohort of girls was followed prospectively from Following recruitment, mothers were interviewed
infancy through to age 18. at home in the summer before daughters' entry into
804 Ellis et al.
as thewhen
kindergarten (see Dodge, Pettit, & Bates, 1994), cutoff for late father absence to complete
most children were 5 years of age. The measurement
90-min of father absence before the onset of
audiorecorded interview included both open-ended
first pregnancy in daughters. Father presence was
and structured questions about each of two eras as
defined inbirth father presence in the home through
age
the child's life (a period from 12 months of 13.up
age Classification
to of girls into the father-present
or father-absent groups was based solely on birth
12 months ago, and the past 12 months). Questions
father status and did not take stepfathers into
concerned the child's development and child-care
account
history, family stressors, parental behavior, (33% = early father absent, 12% = late father
exposure
to socializing factors, and current functioning.
absent, 55% = father present).
Reliability was assessed through independent rat-
ings of 41 randomly selected families made bySexual
Adolescent a Outcomes
second coder who sat in with the interviewer.
Additional home interviews with the mothers were Early sexual activity. In Year 12 (age 16), girls were
conducted in Years 7 and 9 of the study (whenasked whether they had ever had sexual intercourse.
daughters were approximately ages 11 and 13).Girls who responded "no" were coded as 0 for early
Questions concerned family changes and adjust-
sexual activity (60%); girls who responded "yes"
ment, child's involvement in after-school care were coded as 1 for early sexual activity (40%). The
age 16 cutoff has been commonly used in past
settings, parenting practices, and neighborhood
characteristics over the past year. studies to demarcate early onset of sexual activity
(e.g., Fergusson & Woodward, 2000b; Kiernan &
In addition, mothers annually completed child
Hobcraft, 1997; Paul, Fitzjohn, Herbison, & Dickson,
behavior-problem questionnaires and provided fam-
2000).
ily demographic data. Behavior-problem question-
naires were also completed by daughters in Years 11
Adolescent pregnancy. In Years 10 through 13 (ages
through 13 of the study (approximate ages 15-17).
14-17), girls were asked annually whether they had
Daughters answered questions about sexual behav-
become pregnant in the last year. Girls who reported
ior and pregnancy at this time. Also at this time,
no pregnancies over this period were coded as 0 for
research staff requested permission to view the
adolescent pregnancy (85%); girls who reported at
participants' academic records. least one pregnancy over this period were coded as 1
for adolescent pregnancy (15%).
Timing of Onset of Father Absence
Covariate Factors
To determine timing of onset of father absence,
To assess the extent to which associations between
household composition data were collected during
timing of father absence and adolescent sexual
Years 1 through 9 of the study (ages 5-13). Because
Hetherington (1972) and Draper and Harpending outcomes could be explained by the effects of early
(1982) suggest that the first 5 years of life constituteexternalizing
a problems and familial and ecological
sensitive period for the effects of father absencestressors,
on the following 10 variables were included
as covariates in the analysis. The measures of
daughters' sexual development, early onset of father
familial and ecological stress were chosen as
absence was defined in this study as absence of the
"birth father" (either the biological father orcovariates
an on the basis of past research indicating
(a) covariation with father absence and (b) predic-
adoptive father present from birth) from the home at
tion to early sexual activity and adolescent preg-
or before age 5. This cutoff was also chosen to allow
comparison with past studies, which have com- nancy (see reviews by Kotchick et al., 2001; Miller,
monly defined early father absence as occurringBenson,
in & Galbraith, 2001). The covariates were
measured repeatedly and prospectively from the
the first 5 years (e.g., Bereczkei & Csanaky, 1996;
Blain & Barkow, 1988; Hetherington, 1972). Girlsbeginning of each study through age 13.
were thus classified as early father absent if they
Externalizing behavior problems (early childhood).
During Years 1 and 2 of the study (ages 5-6),
were either born into single-mother families or born
into intact two-parent families but subsequently
mothers completed the Child Behavior Checklist
experienced birth father absence at or before age(CBCL;
5. Achenbach, 1991). The 33-item externalizing
problems score, which has been reported to have
Late onset of father absence was defined as birth father
excellent psychometric properties (Achenbach,
presence in the home through age 5 but subsequent
1991), was used to index daughters' early externaliz-
absence of the birth father from the home beginning
sometime during ages 6 through 13. We chose ageing
13 problems. A composite externalizing behavioral
Father Absence 805
of discipline
problems score was computed by averaging over practices and whether the child had
Years 1 and 2 (A = .81, M = 10.63, SD = 6.47).
ever been harmed by an adult during each era (see
Dodgehow
Mother's age at first birth. Mothers reported et al., 1994). Interviewers then completed
old they were when they first gave birth ratings
to a child
of the degree of restrictive discipline received
(M = 23.23, SD = 4.82). by the child (1 =nonrestrictive, mostly prosocial
Race. Race was coded as a dummy variable:
guidance; 5 = severe, strict, often physical) and whether
0 = Caucasian (83%), 1 = non-Caucasian (17%). Of the target child had been severely harmed (1 = defi-
the 42 non-Caucasian participants, 38 were Africannitely not, 5 = authorities involved). These four ratings
American. (two ratings for each of two life eras) were averaged
SES. SES was computed on the basis of mothers' to derive the early childhood harshness of discipline
and fathers' occupation and years of education score (c = .81, inter-rater agreement = .78, M = 2.05,
SD et
(Hollingshead, 1975; full description in Dodge = .67).
Harshness of discipline (preadolescence). Harshness
al., 1994). Because the rank-ordering of SES between
of discipline was also assessed during the Years 7
families was highly stable over time, a composite
childhood SES score was computed by averaging and 9 interviews. Using a 4-point scale (1 = never,
SES scores from Year 1 (age 5) and Year 9 (age4 13;
=frequently), mothers rated how often they used
a = .84, M = 38.11, SD = 12.78). each of six harsh disciplinary tactics (e.g., scold, slap
or hit with hand, use belt/paddle). A composite
Family life stress (early childhood). Family life stress
was assessed during the Year 1 interview on harshness
the of discipline measure was computed by
basis of questions concerning changes and adjust-
averaging the Year 7 (a = .67) and Year 9 (x = .67)
ments in the home and their perceived impact measures (c across the two measures = .77, M = 2.06,
on the child during each era (see Dodge etSD = .42).
al.,
1994). Interviewers completed ratings of the extent
Parental monitoring (preadolescence). Parental mon-
of stressful, challenging events faced by itoring
the was assessed during the Years 7 and 9 home
interviews with the mothers. Although the two
child and family (1 = minimum challenge, 5 = severe
frequent challenges). The rating from the two eras
measures had slightly different content, both em-
were averaged to yield a score for family ployed
life 5-point frequency scales and focused on
stressors (ac= .64, proportion agreement between
parents' awareness of their children's activities and
companions. A composite measure of parental
independent raters of the same protocol = .79,
M = 3.04, SD = .94). monitoring was computed by standardizing and
Dyadic adjustment (early childhood). During thethen averaging the Year 7 (c = .73, M=4.65,
SD = .34; see Pettit et al., 1999) and Year 9 (c = .67,
Year 1 interview, mothers were asked to recall each
era and answer questions concerning the kinds M of= 4.32, SD = .45; see Pettit, Laird, Dodge, Bates, &
family strife and violence the child was exposedCriss,to 2001) measures (c across the two mea-
sures
(see Ellis et al., 1999). Interviewers then completed = .66).
ratings of the severity of conflict within the parental Neighborhood danger (preadolescence). Neighbor-
hood danger was assessed during the Years 7
dyad (1 = rarely even shout; 5 = physical fights, more
than once). The rating from the two eras were and 9 home interviews with the mother. During
the Year 7 interview, mothers responded to a set
averaged to yield an overall score (c = .74, inter-
rater agreement = .80, M = 2.19, SD = 1.03). Mothers of six items (adapted from the Self-Care Checklist;
were also asked questions concerning levels of help see Posner & Vandell, 1994) describing their
and emotional support from their partners during general appraisal of neighborhood and family safety.
each era (see Ellis et al., 1999). Interviewers then Items were rated on a 6-point scale (very safe to very
completed ratings of level of supportiveness in the unsafe) and averaged to form an overall neighbor-
parental dyad, and the ratings from the two eras hood safety score (a=.90, M=2.01, SD=.86).
were averaged to yield an overall score (a = .88, In addition, immediately following the Year 7
inter-rater agreement = .86, M = 2.37, SD = .57). and Year 9 interviews, the interviewer completed
A composite measure of dyadic adjustment was a 4-point rating of overall neighborhood safety
computed by standardizing and then averaging the (very safe to very unsafe; Ms= 1.82 and 1.71,
measures of "severity of conflict within the parental SDs = .85 and .77, respectively). A composite mea-
dyad" (reverse-scored) and "supportiveness in the sure of neighborhood danger was computed by
parental dyad" (a across the two measures = .55). standardizing and then averaging the mother-report
Harshness of discipline (early childhood). During the and two interviewer-report measures (A across the
Year 1 interview, mothers were asked about their use three measures = .78).
806 Ellis et al.
facilitate comparison with rates of early sexual

Measures of Psychosocial Adjustment and Educational
Achievement (Adolescence) activity and teenage pregnancy, both self-reports
and mother reports of both externalizing behavior
To assess the extent to which timing of father
problems were dichotomized (bottom 85% = 0, top
15%activity
absence discriminantly predicted early sexual = 1).
and adolescent pregnancy (but not other behavioral
Internalizing behavior problems (adolescence). Self-
and mental health problems), the following report
educa- and mother reports of internalizing be-
tional and psychosocial outcome variables havior problems-those behaviors considered to
were
investigated. These outcomes were measured be anxious,
con-withdrawn, or depressed-were also
currently with assessment of timing of sexualassessed in Years 11 through 13 using the YSR
activity
and adolescent pregnancy from ages 14 to 18.and CBCL (Achenbach, 1991). The highly reliable
internalizing
High school grade point average (GPA). Data on high problems score (32 items in both
school GPA were drawn from archival school the YSR and CBCL) was used to index daughters'
records (Grades 9-11). Staff members examined each
adolescent internalizing problems. A composite
child's file and noted the grades earned in math, internalizing behavioral problems score
self-report
was computed by averaging self-reports over
language, science, and social studies. Conventional
grade conversions were used (i.e., A =4, B=113,
Years through 13 (a across the three scores = .86,
C = 2, D = 1, E = 0). A composite GPA was
M =calcu-
11.39, SD = 7.40) and a composite mother-report
lated for each child by averaging the grades internalizing
received behavioral problems score was
across the four subjects across the threecomputed
years by averaging mother reports over Years
(0 = .89, M = 2.50, SD = .96). 11 through 13 (a across the three scores = .84,
Violent acts (adolescence). Data on violent acts were
M = 7.18, SD=5.98). The composite self-report
collected in Years 12 and 13 (approximate ages 16- and mother-report internalizing scores were
17). Girls in each year reported how often they had moderately correlated, r (241) = .46, p<.001. Again,
performed each of seven violent acts in the last 12 to facilitate comparison with rates of early sexual
months (e.g., "How many times have you been activity and teenage pregnancy, both self-reports and
physically cruel to someone else [causing harm]?" mother reports of both internalizing behavior
"How many times have you started a fight with problems were dichotomized (bottom 85% = 0,
someone else, where you hurt that person?" "How top 15% = 1).
many times have you used a weapon that can cause
serious physical harm to others [like a bat, brick, Method: New Zealand
broken bottle, knife, or gun]?"). Girls who reported
no violent acts in either year were coded as 0 for
violent acts (76%); girls who reported at least one
Participants and Overview
violent act in either year were coded as 1 for violent
acts (24%). The New Zealand data were collected as part o
Externalizing behavior problems (adolescence). Self- the Christchurch Health and Development Stud
report and mother reports of externalizing behavior (CHDS). The CHDS is an ongoing longitudinal study
problems were assessed in Years 11 through 13 (ages of an unselected birth cohort of 1,265 children (63
15-17) using the Youth Self-Report (YSR) and CBCL, males, 630 females) born in the Christchurch, New
respectively (Achenbach, 1991). The highly reliable Zealand, urban region during a 4-month period
externalizing problems score (30 and 33 items in the mid-1977 (Fergusson & Horwood, 2001; Fergusson
YSR and CBCL, respectively) was used to index Horwood, Shannon, & Lawton, 1989). The current
daughters' adolescent externalizing problems. A research is based on this female subsample, which
composite self-report externalizing behavioral pro- was demographically diverse and representative of
blems score was computed by averaging self-reports the geographic region (13% Maori/Polynesian, 25%
over Years 11 through 13 (a across the three father unemployed or in low-skill occupation, 8%
scores = .87, M = 10.72, SD = 6.29) and a composite living with a single mother at birth). The girls and
mother-report externalizing behavioral problems their families have been studied at birth, 4 months, 1
score was computed by averaging mother reports year, and at annual intervals to age 16 years, and
over Years 11 through 13 (L across the three again at ages 18 and 21 years. In the vast majority of
scores = .90, M = 7.91, SD= 7.39). The composite cases (typically>95%) follow-up assessments have
self-report and mother-report externalizing scores been conducted within 4 weeks of the sample
were moderately correlated, r (241) = .52, p < .001. To member's birthday. Data have been collected from
Father Absence 807
a combination of sources including: parentalcoding

inter- procedures used in the U.S. sample
views (birth-16 years), self-report (8-21 (16%
years),
= early father absent, 11% = late father absent,
psychometric testing (8-13 years), teacher73%reports
= father present).
(6-13 years), medical records (birth-21 years), and
police records (13-21 years). In general terms theSexual
Adolescent Outcomes
aims of the study have been to build up a running
record of the life history, social circumstances,
Early sexual activity. At each assessment from ages
health, and development of a large cohort14 oftoNew
16, sample members were questioned concern-
Zealand children growing up in the 1980s and
ing1990s.
their sexual behavior, including their experience
of consensual
In particular, the study has gathered a wealth of sexual intercourse since the previous
information on family composition, social and
assessment. At age 18 sample members were again
family functioning in childhood, and psychosocial
questioned concerning their previous experience of
outcomes in adolescence. sexual intercourse, and those who reported such
The present analyses are based on the sampleexperience
of were asked to report their age at first
520 female cohort members for whom information experience of consensual intercourse. Young women
were classified as having engaged in early sexual
on the timing of father absence and adolescent
outcome measures was available. This sampleactivity if they had ever reported involvement in
consensual sexual intercourse before age 16. Overall,
represented 83% of the original cohort of 630 females
and was generally representative of the original33% of the sample reported early sexual activity.
sample (13% Maori/Polynesian, 23% father unem- Adolescent pregnancy. At age 14, the mothers of
ployed or in low-skill occupation, and 7% living female sample members were asked whether their
with a single mother at birth). Comparison ofdaughterthe had ever been pregnant. From age 15
analysis sample of 520 females with the remaining onwards sample members themselves were ques-
110 sample members from the original female cohort tioned about any pregnancies since the previous
on a range of sociodemographic measures collected assessment and, in particular, the timing and out-
come of these pregnancies. Young women were
at birth suggested slight but statistically significant
classified as having an adolescent pregnancy if they
(p < .05) tendencies for the analysis sample to under-
represent girls from socially disadvantaged back- had ever been reported as being pregnant before age
grounds (low paternal occupational status, 18. lowOverall, 8% of young women had been pregnant
maternal education). This raises the issue of the before age 18.
extent to which study findings could be influenced
by the effects of sample-selection bias. To examineCovariate Factors
this issue, all analyses were repeated using the data-
weighting method described by Carlin, Wolfe,To assess the extent to which associations between
Coffey, and Patton (1999) to adjust for possible timing of father absence and adolescent sexual
outcomes could be explained by the effects of child
selection effects resulting from the pattern of sample
conduct problems and familial and ecological
attrition. These analyses produced essentially iden-
stressors, we included the following 10 variables as
tical results to those based on the unweighted data,
suggesting that the small biases detected in covariates
the in the analysis.
sample are unlikely to affect study conclusions. Early conduct problems (6 years). When sample
Because the two sets of results were mutually members were age 6, maternal and teacher reports of
the child's tendencies to conduct disordered and
consistent, in the interests of simplicity, the results
reported here are based on the unweighted sample oppositional behaviors were obtained using the 9-
data. item mother- and teacher-report versions of the
Rutter Behavior Rating Scale (Rutter, Tizard, &
Whitmore, 1970). For the present analysis the
Timing of Onset of Father Absence
maternal and teacher reports were summed to
Comprehensive data were gathered on family produce an overall scale measure reflecting the
composition at annual intervals to age 13, including extent to which the child was reported to be
information on the relationship between the daugh- exhibiting conduct problems at age 6 (a= .83,
ter and any adult males in the home. Classification M = 20.44, SD = 3.21).
of girls into the three father-absent and father- Maternal age at first childbirth. The mother's age at
present groups (early father absent, late father first childbirth was assessed during the initial
absent, and father present) was based on the same parental interview at the time of the survey child's
808 Ellis et al.
birth. The mean age at first childbirth was12 months.

23.7 yearsFor each item, a count of the number of
(SD = 4.2). positive reports over the 10-year period was calcu-
Race. The sample member's ethnicity was codedlated, and the resulting count measures were
as a dummy variable: 0 = European New Zealandercombined to produce a scale measure of the extent
(87%), 1 = Maori/Polynesian (13%). to which sample members were exposed to parental
Maternal education. The mother's education level conflict from birth to age 10 years (Fergusson,
Horwood, & Lynskey, 1992; a = .66, M = 4.24,
was assessed at the time of the survey child's birth
and coded into a three-level classification: no formal SD = 8.98).
educational qualifications (50.0% of the sample),Early mother-child interaction (3 years). To provide
high school qualifications (28.3%), and postsecond-
an assessment of the quality of early mother-child
ary certificate or degree (21.7%). Higher scores interactions, when sample members were age 3,
mothers were assessed on the 10-item Maternal
indicated higher levels of educational achievement.
Emotional Responsiveness and 5-item Maternal
Father's occupational status. Father's occupational
Punitiveness subscales of the Home Observation
status was classified at the time of the survey child's
for Measurement of the Environment (HOME) Inven
birth using the Elley-Irving (1976) scale of occupa-
tional status for New Zealand. This scale classifies
tory (Bradley & Caldwell, 1977; Elardo, Bradley, &
Caldwell, 1977). Each item is scored 0 or 1 to indicate
families into six groups on the basis of paternal
occupation. In the present analysis, the Elley-Irving
the absence or presence of the target behavior. The
coding was reduced to a three-level classification Maternal
as Emotional Responsiveness subscale pro-
vides an index of the frequency with which the
follows: Levels 1, 2 (professional, managerial: 22.5%
mother makes positive emotional responses to her
of the sample); Levels 3, 4 (clerical, technical, skilled:
54.4%); and Levels 5, 6 (semiskilled, unskilled, child and was scored so that a high score indicates
unemployed: 23.1%). This variable was reverse-more positive responses (a = .69, M = 8.44, SD =
scored so that higher scores represent higher 1.41). The Maternal Punitiveness subscale provides
occupational status. an index of the frequency with the which the mother
Family living standards (0-10 years). At eachis observed to make punitive responses to her child's
assessment from ages 1 to 10 years, a measure ofbehavior and was scored so that a high score implies
the quality of the family's standard of living was
more punitive responses (a = .71, M = .82, SD = .80).
obtained on the basis of an interviewer rating of
family living standards. Ratings were made on a 5-
Measures of Psychosocial Adjustment and Educational
point scale (1 = family obviously poor/very poor, 5 = fa-
Achievement (14-18 years)
mily obviously affluent and well-to-do). These ratings
were averaged over the 10-year period to provide an At ages 15 and 16, sample members were
overall measure of the quality of family living
interviewed by trained survey interviewers on a
standards during this period (a across the 10 comprehensive mental health interview that exam-
ratings = .92, M = 2.16, SD = .45). ined various aspects of the young person's psycho-
Family life stress (0-10 years). At each assessment
social adjustment over the preceding 12 months. A
up to the child's age 10, parents were questioned parallel interview was administered to parents. At
about the occurrence of adverse family life eventsage 18, a similar interview was administered to
during the preceding year using a 20-item life eventssample members that assessed the individual's
inventory based on the Holmes and Rahe (1967) mental health, psychosocial adjustment, and educa-
Social Readjustment Rating Scale. For each year,tional a achievement from 16 to 18 years. Using this
life events score was calculated for the family basedinformation, the following additional outcome mea-
on a count of the number of adverse events reported. sures were constructed.
To provide an overall measure of the family's School qualifications. School Certificate is a na-
exposure to adverse life stress from birth to tional
10 series of examinations that is undertaken by
years, the annual life events scores were summed
most New Zealand students in their third year of
high school. Students may sit examinations in any
over the 10-year period (c across the 10 ratings = .80,
mean number of adverse life events = 20.80, number of subjects (typically four or five), and
SD = 12.22). performance in each subject is graded from A to E,
Marital conflict (0-10 years). At annual intervals with a grade of C or better implying a pass in that
up until the children were age 10, parents were subject. For the present analysis, a young woman
questioned using three items that described the was classified as having left school without qualifi-
quality of the marital relationship over the previous cations if she had left school by age 18 years without
Father Absence 809
at least one pass grade in School Certificate: This

behavior or other mental health problems were
criterion was met by 16.5% of the sample. offered assistance in obtaining a referral to an
Mood disorder. At ages 15 and 16, information on treatment service.
appropriate
the young person's experience of depressive symp-
Violent offending. At ages 15 and 16, the young
tomatology was obtained using items from person's
the childinvolvement in criminal offending over the
and parent versions of the Diagnostic Interview
previous year was assessed using the Self Report
Schedule for Children (DISC; Costello, Edelbrock, Early Delinquency inventory (SRED; Moffitt & Silva,
Kalas, Kessler, & Klaric, 1982). This information was 1988). Similar questioning was conducted at age 18
used to classify young people according to the using the Self Report Delinquency Inventory (SRDI;
Diagnostic and Statistical Manual of Mental Disorders Elliott & Huizinga, 1989). Using these data, young
(3rd ed., rev. [DSM-III-R], American Psychiatric women were classified as being violent offenders if
Association, 1987) symptom criteria for major de- they reported committing any violent offence (in-
pression (Fergusson, Horwood, & Lynskey, 1993). cluding physical assault, getting into fights, using a
At age 18 years, the assessment of depressive weapon or strong-arm tactics to commit a robbery,
symptomatology was based on the Diagnostic and threatening behavior, and related offenses) over the
Statistical Manual of Mental Disorders (4th ed. [DSM- 4-year period: This criterion was met by 13.7% of the
IV], American Psychiatric Association, 1994) criteria sample.
for major depression assessed using items from the Conduct disorder. At ages 15 and 16, sample
Composite International Diagnostic Interview members were assessed on DSM-III-R symptom
(CIDI; World Health Organization, 1993). For the criteria for conduct disorder based on self-reports
present analysis, young women were classified as and parent reports on the SRED (Fergusson et al.,
having a mood disorder from 14 to 18 years if they 1993). At age 18, DSM-IV criteria for conduct
met the relevant DSM criteria for major depression disorder were derived from items in the SRDI.
on the basis of self- or parent-report at any time Young women were classified as conduct disordered
during the 4-year period: This criterion was met byif they met DSM criteria for conduct disorder on the
37.3% of the sample. basis of self-report or parental report at any time
Anxiety disorder. Parallel to the assessment of during the 4-year period: This criterion was met by
major depression, at ages 15 and 16 sample members 7.5% of the sample.
and their parents were also questioned about the
young person's history of anxiety symptomatology Results
in the previous 12 months using items from the
DISC. This information was used to classify young
Statistical Analyses
people on DSM-III-R criteria for the following
anxiety disorders: separation anxiety, overanxious As described previously, there were 16 dependent
disorder, generalized anxiety disorder, social phobia, variables to be analyzed: early sexual activity,
simple phobia, agoraphobia, and panic disorder. As teenage pregnancy, and six other measures of
part of the age 18 interview, items from the CIDI psychosocial adjustment and educational achieve-
were used to assess DSM-IV symptom criteria for ment in each of the two samples. With one exception
the following anxiety disorders: generalized anxiety (GPA in the U.S. sample), all outcomes were
disorder, social phobia, specific phobia, agorapho- dichotomous. Analysis of the associations between
bia, and panic disorder. For the present analysis, father absence and the dependent variables was
young women were classified as having an anxiety conducted in several stages.
disorder if they met DSM criteria for any of the Before conducting the primary data analysis,
preceding disorders over the 4-year period: This preliminary analyses were carried out to test the
criterion was met by 44.6% of the sample. linearity of the associations between the three-level
Suicide attempts. At ages 15, 16, and 18, sample timing of onset of father absence measure and the
members were questioned about their experience of dependent variables. For the 15 dichotomous de-
suicidal thoughts since the previous assessment. pendent variables, these tests were conducted using
Those reporting suicidal thoughts were further the Mantel-Haenszel chi-square test of linearity.
questioned about any suicide attempts and the Comparison of the Mantel-Haenszel results with
frequency, nature, and outcome of any such at- the alternative Pearson's chi-square test of indepen-
tempt(s). Overall, 7.1% of the sample reported dence showed that, in all cases, the linear model
making at least one suicide attempt during the 4- appeared to provide the best fitting and most
year period. All respondents who reported suicidal parsimonious representation of the association. For
810 Ellis et al.
standardized
the measure of GPA, similar tests of linearity were regression coefficient (beta) for the
conducted within an ANOVA framework. Thesetiming of father absence measure.
tests also suggested that a linear model mostTo illustrate the extent of the association between
accurately represented the association. Wethe
thus
timing of father absence and the binary outcome
measures
concluded that the relations between timing of onset after adjustment for covariates, estimates
of father absence and all outcome measures were of the adjusted rates for each outcome were
essentially linear. In all subsequent analyses, there-
computed using the parameters of the fitted logistic
fore, father absence was treated as a continuous regression models. The adjusted rates were com-
(linear) variable, which was coded so that higherputed using the method described by Lee (1981) and
scores indicated earlier onset of father absencecan be interpreted as the hypothetical rates of each
outcome that would have been observed had all
(0 = father presence, 1 = late onset of father absence,
2 = early onset of father absence). sample members experienced their existing mix of
covariate factors but varied in their exposure to
Treating father absence in this manner is concep-
tually similar to analyzing age at onset of father
father absence.
absence. Although age at onset might be a more
appropriate metric for analysis, detailed information
Rates of Early Sexual Activity and Adolescent Pregnancy
on this variable was available only in the New
by Timing of Father Absence
Zealand sample. Thus, for consistency we have used
Do rates of early sexual activity and adolescent
the same three-level classification of timing of onset
pregnancy differ according to timing of onset of
of father absence across the two samples. However,
father absence? We expected a dose-response rela-
further analysis of the New Zealand data indicated
that age at onset of father absence correlated in in which early father-absent girls would
tionship
excess of .97 with the current three-level measure. have the highest rates of early sexual activity and
This suggests that similar conclusions would be
teenage pregnancy, followed by late father-absent
drawn if more accurate assessments of the timing girls,
of followed by father-present girls.
father absence were available in both samples. Figure 1 shows rates of early sexual activity and
teenage pregnancy in both the U.S. and New
The principal data analyses were based on a series
of regression analyses examining the relations
Zealand samples according to timing of father
between the timing of father absence and the 16
absence: Early father absence (beginning ages 0-5),
late father absence (beginning ages 6-13), and father
dependent variables before and after adjustment for
child, family, and ecological factors. For binary
presence (ages 0-13). For each father-absent and
dependent variables, these analyses were conductedfather-present group, the solid lines in the figure
using logistic regression methods in which the logshow the percentage of girls who had sexual
odds of the dependent variable was modeled asintercourse
a by age 16 and the percentage of girls
linear function of the timing of father absence andwho experienced an adolescent pregnancy. Logistic
covariates (where applicable). The full covariate regression of the data in Figure 1 showed that earlier
adjusted model fitted to the data was of the form: onset of father absence was associated with a
corresponding increase in girls' rates of both early
sexual activity and adolescent pregnancy in bot
logit [pr(Yi)] =BOi + B1iX1 + ,BjiZj
samples. For early sexual activity in the U.S. sample
where logit[pr(Yi)] was the log odds of the ith N = 227, B(SE = .16) = .70, X2 = 20.51, p<.0001, odds
dependent variable, X1 was the continuous measure ratio = 2.01; and for early sexual activity in the Ne
of timing of father absence, and Zj was the set ofZealand sample: N = 520, B(SE = .12) = .76,
child, family, and ecological covariates. The para-X2 = 38.04, p<.0001, odds ratio = 2.14. For adolescent
pregnancy in the U.S. sample: N = 242,
meter Bli represents the effect of father absence on
the log odds of the ith dependent variable. B(SE
A = .23) = 1.15, X2=24.97, p<.0001, odds ra-
tio = 3.15; and for adolescent pregnancy in the
measure of effect size is provided by the odds ratio
between the timing of father absence and the
New Zealand sample: N= 520, B(SE = .19)= 1.16,
dependent variable. The odds ratio represents the
X2= 38.28, p<.0001, odds ratio = 3.19. As expected,
multiplicative effect of a one-unit shift in the three-
early father-absent girls had the highest rates of both
early sexual activity and adolescent pregnancy,
level father absence variable. The corresponding
analyses for the continuous dependent variablefollowed by late father-absent girls, followed by
(GPA) were based on standard linear regression,father-present girls (Figure 1). For example, adoles-
and the measure of effect size is provided by the
cent pregnancy rates were approximately 7 times
Father Absence 811
United States New Zealand

65% 65%
60% -.--------..------ -----------------------------------0--- 60%0 -
55% ........ ............... ................................ 55% ........ .... .............................

5 0 % ............... . ......... ........................... 5 0 % ............... . ... . .... ...............................
45% -------- ----- ------------..........- 45% ..................Rates of

44- early sexual
40% 40% ............... ................40%........... ........................ activitya
Rates of
35%......................--..---.------- 35% -............-E.- early sexual

2....activityb
30%.---........
% 40Rates of
25% - ...................... 25% - ................ ....... ...... teenage
pregnancya
20% - --------------------- - 20% -
Rates of
1. .. teenage
1 5% ----- ----------------------------- 15% . .... . G . ..... .. .. ... .. .. ... .. .. .. ........ b
pregnancy
10% . ............................%................................
o10%..... . .................................. ..................
5% - ----------------- ------------------------- ----- ---- - 5% -
0% I I 0% I I I
Early father- Late father- Father pre- Early father- Late father- Father pre-
absent girls absent girls sent girls absent girls absent girls sent girls
aRates prior to adjustment for covariates. bRates after adjustment for covariates.
Figure 1. Rates of early sexual activity and teenage pregnancy, before and after adjustment for covariates.
Table 1 displays mean levels of child conduct

higher in the U.S. sample and 8 times higher in the
New Zealand sample among early father-absent
problems and familial and ecological stressors in
relation to (a) the timing of father absence, (b)
girls than among father-present girls. In addition,
occurrence of early sexual activity, and (c) occur-
there was remarkable similarity between the U.S.
and New Zealand samples in both the orderingrence
of of an adolescent pregnancy. For ease of data
presentation, all measures (except for race and
results across groups and the base rates for early
sexual activity and teenage pregnancy within each mother's age at first birth) have been expressed in
group (despite the overall base rates' being higher standardized
in form. Mean differences were tested
the U.S. sample). using the F statistic.
Table 1 demonstrates the presence of a pervasive
relationship between earlier timing of father absence
Child, Family, and Ecological Factors Associated With
and more exposure to familial and ecological
Timing of Father Absence, Early Sexual Activity, and
stressors. Across both samples, girls whose birth
Adolescent Pregnancy
fathers were absent from an earlier age were more
likely to come from socially disadvantaged back-
Although the results in Figure 1 indicate that
earlier onset of father absence was associated with
grounds characterized by young motherhood, min-
increased risk of early sexual activity and adolescent
ority racial status, lower SES, more family life stress,
pregnancy, it is possible that these associationspoor
are parental relationships (i.e., low dyadic adjust-
due to contextual factors that correlate with both the
ment, high marital conflict), and low-quality par-
ental investment (i.e., harsh discipline, lack of
timing of father absence and early sexual activity
and adolescent pregnancy. To examine this issue, parental monitoring, low maternal emotional
812 Ellis et al.
Table 1
Mean Levels of Child Conduct Problems and Familial and Ecological Stressors by Timing of Father Absence, Early Sexual
Pregnancy: United States and New Zealand
Father absence status Sexual activity Pregnancy status
Early Late Early No early

father father Father sexual sexual Not
Variable absence absence presence F activity activity F Pregnant pregnant F
United States
Externalizing problems 0.20 - 0.24 - 0.08 2.86 0.22 - 0.13 6.66* 0.48 - 0.09 10.77***
(ages 4-6)
Mother's age at first birth 20.82 22.30 24.84 19.80*** 22.69 23.63 1.98 21.68 23.51 4.24*
Race (% other) 32% 21% 8% 19.28*** 24% 13% 4.33* 41% 13% 16.65***
SES (ages 4-13) -0.58 -0.07 0.40 28.78*** -0.19 0.18 7.71** - 0.59 0.15 17.48***
Family life stress 0.43 0.23 - 0.35 18.55"*** 0.17 - 0.13 5.30* 0.33 - 0.08 4.38*
(ages 1-5)
Dyadic adjustment - 0.79 0.09 0.42 46.26#* - 0.34 0.27 21.72*** - 0.67 0.15 21.72***
(ages 1-5)
Harsh discipline 0.38 - 0.21 - 0.19 9.00*** 0.22 - 0.14 7.52** 0.58 -0.11 15.76***
(ages 4-5)
Harsh discipline 0.25 - 0.25 - 0.08 3.69* 0.07 - 0.06 0.87 0.45 - 0.07 7.83**
(ages 10-13)
Parental monitoring -0.47 - 0.04 0.30 15.10*** - 0.22 0.21 10.14** - 0.66 0.13 18.67***
(ages 10-13)
Neighborhood danger 0.57 - 0.08 - 0.31 29.39*- 0.20 - 0.13 7.68** 0.55 -0.11 18.10***
(ages 10-13)
New Zealand
Conduct problems 0.38 0.20 -0.11 9.25*** 0.16 - 0.08 6.12* 0.52 - 0.05 12.17***
(age 6)
Mother's age at first birth 21.01 22.70 24.43 27.07m 22.29 24.38 30.47*** 21.67 23.88 11.03***
Race (% Maori/Polynesian) 28% 19% 8% 26.52*** 15% 12% .94 29% 11% 10.63***
Father's occupation (at birth) -0.54 -0.20 0.15 18.84*** -0.32 0.16 27.28*** - 0.63 0.05 18.50***
Mother's education (at birth) -0.46 -0.31 0.15 16.43#* -0.32 0.15 26.93*** -0.54 0.05 13.70***
Standard of living (ages 0-10) - 0.77 - 0.23 0.20 38.27*** - 0.24 0.12 15.43*** - 0.64 0.06 19.67***
Family life stress (ages 0-10) 0.73 0.58 -0.23 42.78*** 0.34 -0.16 27.72*** 0.79 -0.07 26.79***
Mom emotional responsiveness -0.49 -0.07 0.11 12.61m* - 0.16 0.08 6.20* -0.24 0.02 2.59
(age 3)
Mom punitiveness (age 3) 0.40 -0.19 -0.05 8.15*** 0.10 -0.05 2.32 0.48 -0.04 10.14**
Marital conflict (ages 0-10) 1.18 0.59 -0.32 111.10m 0.32 - 0.15 23.87*** 0.86 -0.07 31.71***
Note. All variables standardized, except race and mother's age at first birth. F statistic and p values for comparison
way ANOVA. Comparison of percentages by race are based on the x2 test. For the U.S. sample, Ns = 213-243; for th
Ns = 468-520.
*p <.05. **p <.01. ***p <.001.
responsiveness). The strong pattern of covariation

motherhood, minority racial status, lower SES, and
more family life stress; who were exposed to
between timing of father absence and girls' exposure
to familial and ecological stressors was similar
dysfunctional parental relationships; and who re-
across the two samples (Table 1). ceived low-quality parental investment were more
Table 1 also demonstrates, in both the U.S. and
likely to engage in early sexual activity and become
New Zealand samples, that early conduct problemspregnant as adolescents (Table 1). The overall
and exposure to familial and ecological stressors
pattern of relations between girls' early behavioral,
familial, and ecological characteristics and their
during childhood were associated with precocious
sexual outcomes. That is, girls who displayed early
subsequent involvement in early sexual and repro-
conduct problems; who were from socially disad-
ductive activity was again similar across the two
vantaged backgrounds characterized by youngsamples (Table 1).
Father Absence 813
followed by father-present girls (Figure 1). For

Rates of Early Sexual Activity and Adolescent Pregnancy
by Timing of Father Absence, After Adjustment example,
for after covariate adjustment, adolescent
Covariates pregnancy rates were approximately 5 times higher
in the U.S. sample and 3 times higher in the New
Next, we examined whether timing of fatherZealand sample among early father-absent girls than
absence contributed to subsequent risk of early among father-present girls (Figure 1).
sexual activity and teenage pregnancy, even after There was one notable difference between the U.S.
controlling for early child conduct problems andand New Zealand samples. Whereas the effects of
familial and ecological stressors. That is, we exam- father absence on sexual activity and adolescent
ined whether father absence constituted an inde- pregnancy remained largely unchanged after covari-
pendent path to early sexual and reproductive
ate adjustment in the U.S. sample, these effects were
activity. substantively reduced after covariate adjustment in
The results presented in Figure 1 and Table 1 the New Zealand sample (as shown in Figure 1). To
indicate that although father absence was associated examine which covariates caused this reduction,
with elevated risk of early sexual activity and additional logistic regression analyses were con-
adolescent pregnancy, the behavioral, familial, and ducted in the New Zealand sample in which father
ecological profiles of father-absent girls were com- absence was entered into the equation simulta-
paratively disadvantaged. Moreover, early conduct neously with each covariate. This enabled us to
problems and exposure to familial and ecological calculate the degree to which individual covariates
stressors consistently predicted early sexual activity caused a reduction in the effect of father absence (as
and adolescent pregnancy. Thus, girls' behavioral, indicated by change in the odds ratio) on early
familial, and ecological profiles could potentially sexual activity and adolescent pregnancy. For
account for the relations between timing of father early sexual activity, the following covariates each
absence and subsequent sexual outcomes. caused a reduction in the odds ratio at least 10%:
To address this issue, we conducted logistic mothers' age at first birth, family life stress, father's
regressions to estimate the strength of the associa- occupational status, maternal education, and marital
tion between timing of father absence and rates of conflict. Similarly, for adolescent pregnancy, reduc-
early sexual activity and adolescent pregnancy after tions in the odds ratio of at least 10% were caused by
adjustment for child conduct problems and familial family living standards, family life stress, father's
and ecological stressors. Ten covariates were simul- occupational status, maternal education, maternal
taneously controlled for in the analyses. These punitiveness, and marital conflict.
covariates are listed in the first column of Table 1 Finally, to examine which group of covariates
(see upper section of table for covariates in theuniquely
U.S. predicted early sexual activity and teenage
study and lower section of table for covariates in
pregnancy after controlling for timing of father
New Zealand study). absence, we again performed the logistic regression
As shown by the broken lines in Figure 1, afteranalyses using forward stepwise procedures, forcing
the entry of the father absence variable into the
statistical adjustment for all covariates, there con-
tinued to be a linear logistic association between
equation on the first step and then allowing free
earlier onset of father absence and higher ratesentry
of of all covariates into the equation on subse-
both early sexual activity and adolescent pregnancy
quent steps. In the U.S. sample, in prediction of both
in both samples. For early sexual activity in the early
U.S. sexual activity and adolescent pregnancy, only
sample: N = 197, B(SE = .23) = .72, X2 = 9.54, early childhood externalizing problems entered the
p=.002, odds ratio = 2.04; and for early sexual
equation after controlling for timing of father
activity in the New Zealand sample: N=466, absence. None of the measures of familial or
B(SE = .17)=.45, z2=6.75, p = .009, odds ratio
ecological stress, therefore, predicted early sex
1.57. For adolescent pregnancy in the U.S. sample: outcomes after controlling for timing of fath
N= 207, B(SE = .33) = .1.07, X2 = 10.45, p = .001, absence and early externalizing problems. In t
odds ratio = 2.91; and for adolescent pregnancy Newin Zealand sample, in prediction of both ear
the New Zealand sample: N = 466, B(SE = .26) = .74,sexual activity and adolescent pregnancy, bot
X2 = 7.89, p = .005, odds ratio = 2.09. Thus, even maternal
after education and family life stress enter
the equation after controlling for timing of fath
simultaneously controlling for all covariates, early
absence. In addition, father's occupational stat
father-absent girls continued to have the highest
entered the equation for predicting early sex
rates of both early sexual activity and adolescent
pregnancy, followed by late father-absent girls,
activity.
814 Ellis et al.
were no substantively meaningful linear relations

Rates of Behavioral Problems and Academic Performance
between timing of father absence and any of the
by Timing of Father Absence, Before and After
Adjustment for Covariates measures of behavioral problems (all p values > .33)
in the U.S. sample, as indicated by both the low odds
Next, we examined whether father absence dis- ratios (range = 1.05-1.35) and relatively flat rates of
criminantly increased risk for adolescent sexual behavioral problems across the two father-absent
outcomes but not for behavioral and mental and one father-present groups. In addition, after
health problems in general. To address this question,
statistical adjustment for all covariates, there was not
we conducted the same regression analyses a substantively meaningful relation between father
that were conducted in the preceding section, absence and high school GPA (N= 177, = -.11,
but we substituted different outcome variables for
t = -1.43, p = .16).
early sexual activity and teenage pregnancy. The As noted in the Method section, the four
outcome measures examined in the U.S. sample measures of externalizing and internalizing behavior
included externalizing behavioral problems (agesproblems were dichotomized (to facilitate compar-
ison with other outcome variables). Because dichot-
15-17; mother report and child report), internalizing
behavior problems (ages 15-17; mother reportomization
and attenuates the power to detect relations
child report), violent acts (ages 16-17), and high
with other variables (MacCallum, Zhang, Preacher,
school GPA. The outcome measures examined in the & Rucker, 2002), we also performed the analyses
New Zealand sample included DSM-III-R diag- using standard linear regression with continuous
noses for conduct disorder, mood disorder, and measures of the four dependent variables (as
anxiety disorder (all ages 14-18); violent offendingdescribed in the Method section). After controlling
(ages 14-18); attempted suicide (ages 14-18); andfor the full set of covariates, the effects of timing of
onset of father absence on both mother- and
failure to attain at least one pass in School Certificate
before leaving high school. As in the previous daughter-reported externalizing and internalizing
analyses, the effect of timing of onset of father
behavior problems remained uniformly small and
absence on each outcome variable was examined statistically nonsignificant (N = 203; Ps range from
before and after adjustment for all covariates .01 listed
to .16, all ps > .05).
in Table 1. The pattern of results was different for the New
The key analysis concerns the effect of timing of Zealand sample. As shown in Table 3 (adjusted rates
father absence after adjustment for covariates. As in parentheses), after statistical adjustment for all
shown in Table 2 (adjusted rates in parentheses),covariates, there was a pattern of modest associa
after statistical adjustment for all covariates, there tions between father absence and the measures of
Table 2
Rates of Behavioral Problems and Academic Performance by Timing of Father Absence, Before and After Adjustment for Covariates: United States
Timing of onset of father absence
Variable Early onset of Late onset of Father B (SE) z2 p Odds ratio

father absence father absence presence
Externalizing problems
Mother report 25.6% 10.3% 9.8% .58 (.20) 8.55 .003 1.79
(15.8%) (13.3%) (11.1%) .30 (.36) 0.69 .41 1.35
Child report 15.6% 24.1% 11.3% .20 (.20) 1.02 .31 1.22
(17.5%) (14.7%) (12.3%) .28 (.36) 0.61 .44 1.32
Internalizing problems
Mother report 14.1% 24.1% 12.9% .08 (.20) 0.15 .70 1.08
(14.1%) (13.7%) (13.2%) .05 (.31) 0.02 .89 1.05
Child report 15.6% 27.6% 12.8% .14 (.19) 0.52 .47 1.15
(18.9%) (16.3%) (13.9%) .22 (.31) 0.49 .49 1.24
Violent acts 39.0% 29.6% 15.3% .63 (.17) 14.22 <.001 1.88
(28.1%) (23.8%) (20.1%) .25 (.26) 0.94 .33 1.28
Note. Percentages after covariate adjustment are shown in parenth

N = 239 and 202 (child report externalizing and internalizing), a
adjustment, respectively.
Father Absence 815
behavioral and mental health problems, as indicated studied throughout their entire childhoods-e
by both the odds ratios (range = 1.36-1.59) abled and the
us to examine child and family variables th
modest decline in rates of these outcome variables preceded risk for involvement in sexual activity a
across the two father-absent and one father-present pregnancy in adolescence. Finally, the use of mu
groups. Most of these associations obtained at leastple informants, in which antecedent child and fam
marginal statistical significance. data were collected from mothers and adolescent
In sum, in the U.S. sample, after statistically sexual outcome data were collected from daughters,
controlling for all covariates, timing of onset of makes it less likely that the current findings are an
artifact of method variance.
father absence remained strongly associated with
early sexual activity and teenage pregnancy but not
with other behavioral problems and academic
Does Father Absence Place Daughters at Special Risk for
performance. Although the direction of the effects
indicated that earlier onset of father absence was Early Sexual Activity and Teenage Pregnancy?
associated with more behavioral and academic Although the current research cannot demon-
problems in the U.S. sample, the size of the strate
effects
causation, three converging lines of evidence
were small and did not approach statistical signifi-
suggest that the answer to this question is yes. First,
cance. By contrast, in the New Zealand sample, in after
both the U.S. and New Zealand samples, there
statistically controlling for all covariates, there
was awas
dose-response relationship between timing of
still a pattern of at least trend associations between
onset of father absence and early sexual outcomes:
timing of father absence and the measures Earlyof
father-absent girls had the highest rates of both
early sexual activity and adolescent pregnancy,
adolescent adjustment, with odds ratios ranging
followed
from 1.36 to 2.09. Although early sexual activity and by late father-absent girls, followed by
teenage pregnancy occupied the upper endfather-present
of this girls. This dose-response relationship
range, and although the odds ratio for teenagesuggests that past research, which has consistently
pregnancy was substantially higher than treated
for anyfather absence as a dichotomous yes-no
variable,
other variable (+.50 or greater), there was not a clear has underestimated the impact of father
absence
divide between the effects of father absence on earlyon daughters' sexual outcomes. This issue
sexual activity and other behavioral and may mentalbe especially relevant to predicting rates of
health outcomes. Specifically, after covariateteenage
adjust-pregnancy, which were 7 to 8 times higher
ment, the odds ratio for early sexual activity (1.57)
among early father-absent girls, but only 2 to 3 times
was about the same as for conduct disorder (1.59),
higher among late father-absent girls, than among
violent offending (1.56), and no school qualifications
father-present girls.
(1.50). Second, in both the U.S. and New Zealand
samples, father absence constituted a unique and
Discussion independent path to early sexual activity and
adolescent pregnancy. Although measures of early
Does father absence uniquely and discriminantly
conduct problems and life-course adversity covaried
with both
increase daughters' risk for early sexual activity and timing of father absence and adolescent
sexual outcomes, these measures either did not
teenage pregnancy, independent of early externaliz-
account
ing behavior problems and exposure to familial for (in the U.S. sample) or only partially
and
ecological stressors during childhood? In addressingfor (in the New Zealand sample) the links
accounted
this question, the current research hadbetweenseveral father absence and early sexual activity and
important strengths. First, the use of a cross-national
teenage pregnancy. The relations between father
research design enabled us to replicate keyabsence
findingsand teenage pregnancy were particularly
across diverse samples in different countries.
robust.Sec-
For example, after controlling for all of the
ond, in conducting two studies, we were able toearly father-absent girls were still about 5
covariates,
times
carry out independent tests of the hypotheses more likely in the U.S. sample and 3 times
using
more likely
different measures and methods. The similarity in in the New Zealand sample to experi-
results across the U.S. and New Zealand ence an adolescent pregnancy than were father-
samples
underscores the robustness and generalizability of In total, these data suggest that father
present girls.
the findings. Nonetheless, it will be important
absence may
to affect daughters' sexual development
through
replicate these findings in non-Western samples processes that operate independently of life-
(see
Waynforth, 2002). Third, the longitudinal course
nature of
adversity and go beyond mere continuation
the research--in which girls were prospectively
of early conduct problems.
816 Ellis et al.
Third, in the U.S. sample, father absence was and most parsimonious representation of
best fitting
discriminantly associated with early sexual the associations between father absence and the
activity
and teenage pregnancy. This association was outcome variables. Power would have been low,
specific
to sexual outcomes and, after controlling for however,
early to detect nonlinearity in the U.S. sample
conduct problems and familial and ecological (given the use of dichotomous dependent variables
stressors, did not extend to academic, behavioral,
and the relatively small sample size in the late
or mental health problems more generally. In the group). The base rates shown in Table
father-absent
New Zealand sample, however, the picture2was indicate
less nonlinear trends in the U.S. data, with late
clear. After covariate adjustment, there was still a
father-absent girls displaying higher rates of inter-
pattern of at least trend associations between nalizing
timingproblems (both child and mother reports)
of father absence and the measures of adolescent and externalizing problems (child reports only) than
adjustment, with early sexual activity and adoles-
either early father-absent or father-present girls.
These nonlinear trends did not replicate in the
cent pregnancy occupying the upper end of this
New Zealand data (see Table 3). Nonetheless, the
range of associations. Considering the U.S. and New
Zealand findings together, after controlling possibility
for that late father absence places daughters
measures of early conduct problems and life-course
at special risk for some outcome variables deserves
adversity, the effects of father absence on sex further
and consideration in future research with larger
pregnancy (a) were generally stronger than weresample
the sizes.
effects of father absence on other outcome variables
and (b) clearly replicated across the two studies
whereas other effects of father absence were more Implications for the Life-Course Adversity Model
equivocal and replicated only in the sense of being In

in the literature on early sexual activity and
the same direction. In sum, after covariate adjust-
teenage pregnancy, the life-course adversity model
ment, there was stronger and more consistent occupies a dominant position. It proposes that a life
history of familial and ecological stress-poverty,
evidence of effects of father absence on early sexual
activity and teenage pregnancy than on other exposure to violence, inadequate parental guidance
behavioral or mental health problems or academic
and supervision, lack of educational and career
achievement. opportunities -makes early sexual activity and
adolescent pregnancy more likely (e.g., Coley &
It is worth reiterating that all of these conclusions
Chase-Lansdale, 1998; Rindfuss & St. John, 1983).
are based on the linear model, which provided the
Table 3
Rates of Behavioral and Mental Health Problems by Timing of Father Absence, Before and After Adjustment for Covariates: New Zealand
Timing of onset of father absence
Variable Early onset of Late onset of Father B (SE) p2 p Odds

father absence father absence presence ratio
Conduct disorder 16.9% 15.8% 4.2% .78 (.19) 17.85 <.001 2.19
(12.6%) (8.5%) (5.7%) .46 (.27) 3.03 .082 1.59
Mood disorder 54.2% 49.1% 31.8% .49 (.12) 17.04 <.001 1.64
(48.1%) (40.9%) (34.1%) .31 (.17) 3.29 .070 1.36
Anxiety disorder 59.0% 54.4% 40.0% .41 (.12) 11.72 .001 1.50
(56.5%) (48.8%) (41.0%) .33 (.17) 3.80 .051 1.39
Violent offending 31.3% 14.0% 9.7% .71 (.15) 23.12 <.001 2.03
(21.4%) (15.2%) (10.5%) .44 (.21) 4.28 .039 1.56
Suicide attempt 14.5% 8.8% 5.3% .56 (.19) 8.33 .004 1.74
(10.9%) (8.3%) (6.3%) .32 (.27) 1.40 .237 1.38
No school qualifications 35.8% 37.5% 9.3% .90 (.14) 41.09 <.001 2.45
(23.7%) (18.5%) (14.1%) .40 (.21) 3.62 .057 1.50
Note. Percentages after covariate adjustment are shown in pa

covariate adjustment, respectively; for all other variables, N = 5
Father Absence 817
The life-course adversity model has gained mechanisms

wide and processes that account for th
acceptance through consistent empiricalrelations support. between increasing exposure to fathe
Rates of teenage pregnancy have been found absencetoand greater risk for early sexual activit
covary positively with family stress, conflict, and adolescent
and pregnancy? From a social learnin
disruptions (e.g., Fergusson & Woodward, 2000a;
perspective, increasing duration of father absence
Hanson, Myers, & Ginsburg, 1987; Robbins et al., with increasing exposure of daughters t
associated
1985); with low parental warmth or support, their
lackmothers'
of dating and repartnering behavior
parental control and monitoring, and maternal and these exposures may encourage earlier onset of
punitive behavior (e.g., Fergusson & Woodward, sexual behavior in daughters, with consequen
2000a; Hansen et al., 1987; Scaramella et al., 1998; increased risk of teenage pregnancy. As Thornto
reviewed in Miller et al., 2001); with low SES (e.g., and Camburn (1987, p. 325) suggest, "We expect th
Fergusson & Woodward, 2000a; Geronimus & many children know whether their parents ar
Korenman, 1992; Robbins et al., 1985); with high sexually active after a marital dissolution and that
neighborhood mortality rates (Geronimus, 1996; formerly married parents who continue to
Wilson & Daly, 1997); and with minority racial or sexually active serve as behavioral models for their
ethnic status (Cheesbrough et al., 1999; Dickson et maturing children, thus increasing the children
al., 2000). The results presented in Table 1 are levels of permissiveness." The social learning mode
consistent with this body of research. thus posits that the effect of father absence o
As discussed in the Introduction, the life-course
daughters' sexual outcomes will be mediated by
adversity model has incorporated father absence as mothers' dating and repartnering behaviors. Th
one of many stressors that can influence sexual hypothesis deserves careful consideration in futur
outcomes. Indeed, as shown in Table 1, timing of research.
father absence significantly covaried with all of the Another possibility is that mothers' dating and
measures of familial and ecological stress in both the repartnering behaviors do not fully mediate the
U.S. and New Zealand studies. Proponents of the relation between father absence and precocious
life-course adversity model have recurrently stated sexual outcomes in daughters. Rather, as discussed
that father absence predicts early sexual outcomes earlier, quality of paternal investment may have a
because it covaries with these stressors (Belsky, et al., direct effect on daughters' sexuality. The current
1991, p. 658; Chisholm, 1999, p. 162; McLanahan, evolutionary model posits that the motivational
1999, p. 119; Robbins et al., 1985, p. 568; Silverstein & systems underlying variation in timing of sexual
Averbach, 1999, p. 403). and reproductive behavior are especially sensitive to
The current research suggests that the opposite the father's role in the family in early childhood.
interpretation is equally plausible: Measures of life- According to Draper and Harpending (1982, 1988),
course adversity may predict early sexual outcomes girls whose early family experiences are character-
primarily because they covary with timing of father ized by father absence tend to develop sexual
absence. In the U.S. sample, father absence predicted psychologies that are consistent with the expectation
early sexual activity and adolescent pregnancy after that male parental investment is unreliable and
controlling for early conduct problems and all of the unimportant; these girls are hypothesized to
measures of familial and ecological stress; however, develop in a manner that accelerates onset of sexual
none of the measures of familial and ecological stress activity and reproduction, reduces reticence in
predicted either early sexual activity or adolescent forming sexual relationships, and orients the
pregnancy after controlling for timing of father individual toward relatively unstable pair-bonds
absence and early conduct problems. The results in (see also Ellis & Garber, 2000; Ellis et al., 1999).
the New Zealand sample were more equivocal: Both This evolutionary model posits an early sensitive
father absence and some measures of familial and period (approximately the first 5 years of life) for the
effects of father absence on daughters' sexual
ecological stress (i.e., maternal education and family
life stress) independently predicted early sexual
development. Although the current results--that
outcomes. earlier onset of father absence was associated with
greater risk for early sexual activity and teenage
pregnancy--are consistent with the sensitive period
Evolutionary and Social Learning Models
hypothesis, they do not clearly support it because
Given that the life-course adversity modeltiming
does of father absence was confounded with
not appear to explain the current results, the of father absence in the current research. In
length
total, the current results are equally consistent with
question then becomes: What are the psychological
818 Ellis et al.
either a sensitive period or linear dose-response

away from a social address perspective, in which
interpretation. variables such as father absence and social class
were studied without explicitly considering how
they influenced child functioning, to a developmen-
Alternative Behavior Genetic Explanations
tal process perspective, in which intervening path-
Perhaps the major weakness of theways current
and mechanisms have become of fundamental
research design was that it was not genetically interest (discussed in Bronfenbrenner & Crouter,
informative. As noted in the Introduction, one 1983). Critiques of the father absence literature
plausible behavior-genetic explanation for the cur- (reviewed in Phares, 1996) partly motivated this
rent findings is that, through genetic transmission, change. A widely held assumption is that it is not
mothers and fathers who have a history of externa- father absence per se that is harmful to children but
lizing disorders not only tend to have daughtersthe stress associated with divorce, family conflict,
who experience externalizing behavioral problemsloss of a second parent, loss of an adult male income,
(including increased rates of early sexual activityand so on. The current research suggests that, in
and teen pregnancy) but also tend to disproportion-relation to daughters' sexual development, the social
ately expose their daughters to father absence and address of father absence is important in its own
accompanying maternal dating and repartneringright and not just as a proxy for its many correlates.
behaviors because externalizing disorders predictThis does not imply that process is unimportant, but
divorce. A second plausible behavior-genetic expla- rather that relevant processes are likely to be father
nation is that mothers who experience early age of driven (e.g., father-daughter processes, father-
first sex and pregnancy not only tend to have mother relationships, exposure to stepfathers; see
daughters who experience early age of first sexEllis et al., 1999).
and pregnancy (through genetic transmission; see In conclusion, father absence was an overriding
Dunne et al., 1997; Rodgers, Rowe, & Buster, 1999)risk factor for early sexual activity and adolescent
but also tend to expose disproportionately theirpregnancy. Conversely, father presence was a major
daughters to father absence and maternal dating andprotective factor against early sexual outcomes, even
repartnering because young mothers are less likely if other risk factors were present. These findings may
to form stable relationships with the fathers of theirsupport social policies that encourage fathers to
children (e.g., Amato, 1996; Bennett, Bloom, & form and remain in families with their children
Miller, 1995). (unless the marriage is highly conflictual or violent;
Consistent with these behavior genetic models, in Amato & Booth, 1997).
the current research both early childhood conduct
problems in daughters and earlier age at first birth in
mothers generally predicted early sexual activity
and adolescent pregnancy in daughters. It is References
important, though, that controlling for both early
Achenbach, T. M. (1991). Integrative guide for the 1991 CBCL
conduct problems and mothers' age at first birth
14-18, YSR, and TRF profiles. Burlington: University of
(along with the other covariates) either did not
Vermont, Department of Psychiatry.
account for (in the U.S. sample) or only partiallyAmato, P. (1996). Explaining the intergenerational trans-
accounted for (in the New Zealand sample) the mission of divorce. Journal of Marriage and the Family, 58,
relations between father absence and elevated rates 628-640.
of early sexual activity and adolescent pregnancy.
Amato, P., & Booth, A. (1997). A generation at risk: Growing
Although these results do not rule out the possibilityup in an era of family upheaval. Cambridge, MA: Harvard
that common genetic influences underlie the covar-University Press.
iation between father absence and precocious sexual American Psychiatric Association. (1987). Diagnostic and
outcomes (see especially Comings, Muhleman,Statistical Manual of Mental Disorders (3rd ed., Rev.).
Johnson, & MacMurray, 2002), they do make it less Washington, DC: American Psychiatric Association.
American Psychiatric Association. (1994). Diagnostic and
likely that the current findings can be accounted for
Statistical Manual of Mental Disorders (4th ed). Washing-
by the specific genetic pathways outlined above.
ton, DC: American Psychiatric Association.
Bardone, A. M., Moffitt, T. E., Caspi, A., Dickson, N., &
Conclusion Silva, P. A. (1996). Adult mental health and social
outcomes of adolescent girls with depression and
Over the last 25 years the field of developmental conduct disorder. Development and Psychopathology, 8,
psychology has experienced a fundamental shift811-829.
Father Absence 819
Bates, J. E., Pettit, G. S., Dodge, K. A., & Ridge, B. correlates,

(1998). and predictors. Advances in Behavioral Re-
The interaction of temperamental resistance to search controlTherapy, 15, 185-209.
and restrictive parenting in the development of Dickson,
externa-N., Sporle, A., Rimene, C., & Paul, C. (2000).
lizing behavior. Developmental Psychology, 34, 982-995. Pregnancies among New Zealand teenagers: Trends,
Belsky, J., Steinberg, L., & Draper, P. (1991). Childhood current status and international comparisons. New
experience, interpersonal development, and reproduc- Zealand Medical Journal, 113, 241-245.
tive strategy: An evolutionary theory of socialization. Dodge, K. A., Bates, J. E., & Pettit, G. S. (1990). Mechanisms
Child Development, 62, 647-670. in the cycle of violence. Science, 250, 1678-1683.
Bennett, N. G., Bloom, D. E., & Miller, C. K. (1995). The Dodge, K. A., Pettit, G. S., & Bates, J. E. (1994).
influence of nonmarital childbearing on the formation of Socialization mediators of the relation between socio-
first marriages. Demography, 32, 47-62. economic status and child conduct problems. Child
Bereczkei, T., & Csanaky, A. (1996). Evolutionary pathways Development, 65, 649-665.
of child development: Lifestyles of adolescents and Draper, P. & Harpending, H. (1982). Father absence and
adults from father-absent families. Human Nature, 7, reproductive strategy: An evolutionary perspective.
257-280. Journal of Anthropological Research, 38, 255-273.
Blain, J., & Barkow, J. H. (1988). Father involvement, Draper, P. & Harpending, H. (1988). A sociobiological
reproductive strategies, and the sensitive period. In K. perspective on the development of human reproductive
B. MacDonald (Ed.), Sociobiological perspectives on human strategies. In K. B. MacDonald (Ed.), Sociobiological
development (pp. 376-396). New York: Springer-Verlag. perspectives on human development (pp. 340-372). New
Bradley, R. H., & Caldwell, B. M. (1977). Home observation York: Springer-Verlag.
for measurement of the environment: A validation study Dunne, M. P., Martin, N. G., Statham, D. J., Slutske, W. S.,
of screen efficiency. American Journal of Mental Deficiency, Dinwiddie, S. H., Bucholz, K. K., et al. (1997). Genetic
81, 417-424. and environmental contributions to variance in age at
Bronfenbrenner, U., & Crouter, A. C. (1983). The evolution first sexual intercourse. Psychological Science, 8, 211-216.
of environmental models in developmental research. In Elardo, R., & Bradley, R. M. C., & Caldwell, B. M. (1977). A
W. Kessen (Ed.), Handbook of child psychology, Vol. 1: longitudinal study of the relation of infant home
History, theory, and methods (pp. 357-414). New York: environments to language development at age three.
Wiley. Child Development, 48, 595-603.
Carlin, J. B., Wolfe, R., Coffey, C., & Patton, G. C. (1999). Elley, W. B., & Irving, J. C. (1976). Revised socio-economic
Tutorial in biostatistics. Analysis of binary outcomes in index for New Zealand. New Zealand Journal of Educa-
longitudinal studies using weighted estimating equa- tional Studies, 11, 25-36.
tions and discrete-time survival methods: Prevalence
Elliott, D. S., & Huizinga, D. (1989). Improving self-
and incidence of smoking in an adolescent cohort.
reported measures of delinquency. In M. W. Klein (Ed.),
Statistics in Medicine, 18, 2655-2679. Cross-national research in self-reported crime and delin-
Cheesbrough, S., Ingham, R., & Massey, D. (1999). A review quency (pp. 155-186). Dordrecht, Netherlands: Kluwer.
Ellis, B. J., & Garber, J. (2000). Psychosocial antecedents of
of the international evidence on preventing and reducing
variation in girls' pubertal timing: Maternal depression,
teenage conceptions: The United States, Canada, Australia,
and New Zealand. London: Health Education Authority. stepfather presence, and martial and family stress. Child
Chisholm, J. S. (1999). Death, hope and sex: Steps to Development,
an 71, 485-501.
evolutionary ecology of mind. Cambridge, England: Cam-Ellis, B. J., McFadyen-Ketchum, S., Dodge, K. A., Pettit, G.
bridge University Press. S., & Bates, J. E. (1999). Quality of early family
Coley, R. L., & Chase-Lansdale, P. L. (1998). Adolescent relationships and individual differences in the timing
pregnancy and parenthood: Recent evidence and future of pubertal maturation in girls: A longitudinal test of an
directions. American Psychologist, 53, 152-166. evolutionary model. Journal of Personality and Social
Comings, D. E., Muhleman, D., Johnson, J. P., & MacMur- Psychology, 77, 387-401.
ray, J. P. (2002). Parent-daughter transmission of Emery,
the R. E., Waldron, M., Kitzmann, K. M., & Aaron, J.
androgen receptor gene as an explanation of the effect of(1999). Delinquent behavior, future divorce or nonmar-
father absence on age of menarche. Child Development, ital childbearing, and externalizing behavior among
73, 1046-1051. offspring: A 14-year prospective study. Journal of Family
Costello, A., Edelbrock, C., Kalas, R., Kessler, M., & Klaric,
Psychology, 13, 568-579.
Fergusson, D. M., & Horwood, L. J. (2001). The Christch-
S. A. (1982). Diagnostic Interview Schedule for Children
(DISC). Bethesda, MD: National Institute of Mental urch Health and Development Study: Review of
Health. findings on child and adolescent mental health. Aus-
Day, R. D. (1992). The transition to first intercourse among tralian and New Zealand Journal of Psychiatry, 35, 285-296.
racially and culturally diverse youth, Journal of MarriageFergusson, D. M., Horwood, L. J., & Lynskey, M. T. (1992).
and the Family, 54, 749-762. Family change, parental discord and early offending.
Devine, D., Long, P., & Forehand, R. (1993). A prospec- Journal of Child Psychology & Psychiatry & Allied
tive study of adolescent sexual activity: Description, Disciplines, 33, 1059-1075.
820 Ellis et al.
Fergusson, D. M., Horwood, L. J., & Lynskey, McLanahan,

M. T. (1993).S. S. (1999). Father absence and the welfare of
children.
Prevalence and comorbidity of DSM-III-R diagnoses inIn E. M. Hetherington (Ed.), Coping with
a birth cohort of 15 year olds. Journal of the divorce,
Americansingle parenting, and remarriage. Mahwah, NJ:
Erlbaum.
Academy of Child & Adolescent Psychiatry, 32, 1127-1134.
Fergusson, D. M., Horwood, L. J., Shannon, F. T.,
Miller, B. &C., Benson, B., & Galbraith, K. A. (2001). Family
Lawton, J. M. (1989). The Christchurch Child relationships
Develop- and adolescent pregnancy risk: A research
ment Study: A review of epidemiological findings.synthesis. Developmental Review, 21, 1-38.
Miller, B. C., Norton, M. C., Curtis, T., Hill, E. J.,
Paediatric & Perinatal Epidemiology, 3, 278-301.
Fergusson, D. M., & Woodward, L. J. (2000a).Schvaneveldt,
Teenage P., & Young, M. H. (1997). The timing of
pregnancy and female educational underachievement:
sexual intercourse among adolescents: Family, peer, and
A prospective study of a New Zealand birth cohort.
other antecedents. Youth and Society, 29, 54-83.
Moffitt, T. R., & Silva, P. A. (1988). Self-reported delin-
Journal of Marriage and the Family, 62, 147-161.
Fergusson, D. M., & Woodward, L. J. (2000b). Educational,
quency: Results from an instrument for New Zealand.
psychosocial, and sexual outcomes of girls with Australian
conductand New Zealand Journal of Criminology, 21,
problems in early adolescence. Journal of Child227-240.
Psychol-
ogy and Psychiatry, 41, 779-792. Newcomber, S., & Udry, J. R. (1987). Parental marital status
Furstenberg, E F., Jr., Brooks-Gunn, J., & Chase-Lansdale,
effects on adolescent sexual behavior. Journal of Marriage
L. (1989). Teenaged pregnancy and childbearing. and Amer-
the Family, 49, 235-240.
ican Psychologist, 44, 313-320. Paul, C., Fitzjohn, J., Herbison, P., & Dickson, N. (2000).
Geronimus, A. T. (1996). What teen mothers know.TheHuman
determinants of sexual intercourse before age 16.
Nature, 7, 323-352. Journal of Adolescent Health, 27, 136-147.
Geronimus, A. T., & Korenman, S. (1992). The socio- Pettit, G. S., Bates, J. E., & Dodge, K. A. (1997). Supportive
economic consequences of teen childbearing reconsid- parenting, ecological context, and children's adjustment:
ered. Quarterly Journal of Economics, 107, 1187-1214. A seven year longitudinal study. Child Development, 68,
Hanson, S. L., Myers, D. E., & Ginsburg, A. L. (1987). The 908-923.
role of responsibility and knowledge in reducing teen- Pettit, G. S., Bates, J. E., & Dodge, K. A., & Meece, D. W.
age out-of-wedlock childbearing. Journal of Marriage and (1999). The impact of after-school peer contact on early
the Family, 49, 241-256. adolescent externalizing problems is moderated by
Hetherington, E. M. (1972). Effects of father absence on parental monitoring, perceived neighborhood safety,
personality development in adolescent daughters. De- and prior adjustment. Child Development, 70, 768-778.
velopmental Psychology, 7, 313-326. Pettit, G. S, Laird, R. D, Dodge, K. A., Bates, J. E, & Criss,
Hogan, D. P., & Kitagawa, E. M. (1985). The impact of M. M. (2001). Antecedents and behavior-problem out-
social status, family structure, and neighborhood on the comes of parental monitoring and psychological control
fertility of black adolescents. American Journal of Sociol- in early adolescence. Child Development, 72, 583-598.
ogy, 90, 825-855. Phares, V. (1996). Fathers and developmental psychopathology.
Hollingshead, A. B. (1975). Four factor index of social status. New York: Wiley.
New Haven, CT: Yale University, Department of Posner, J. K., & Vandell, D. L. (1994). Low-income
Sociology. children's after-school care: Are there beneficial effects
Holmes, S. J., & Rahe, R. (1967). The social readjustment of after-school programs? Child Development, 65, 440-
rating scale. Journal of Psychosomatic Research, 11, 213- 456.
218. Quinton, D., Pickles, A., Maughan, B., & Rutter, M. (1993).
Kiernan, K. E., & Hobcraft, J. (1997). Parental divorce Partners, peers, and pathways: Assortative pairing and
during childhood: Age at first intercourse, partnership continuities in conduct disorder. Development and Psy-
and parenthood. Population Studies, 51, 41-55. chopathology, 5, 763-783.
Konner, M., & Shostak, M. (1986). Adolescent pregnancy Rhee, S. H., & Waldman, I. D. (2002). Genetic and
and childbearing: An anthropological perspective. In J. environmental influences on antisocial behavior: A
B. Lancaster & B. A. Hamburg (Eds.), School-age meta-analysis of twin and adoption studies. Psychologi-
pregnancy and parenthood: Biosocial dimensions (pp. 325- cal Bulletin, 128, 490-529.
346). New York: Aldine de Gruyter. Rindfuss, P. R., & St. John, C. (1983). Social determinants of
Kotchick, B. A., Shaffer, A., Forehand, R., & Miller, K. S. age at first birth. Journal of Marriage and the Family, 45,
(2001). Adolescent sexual risk behavior: A multi-system 553-565.
perspective. Clinical Psychology Review, 21, 493-519. Robbins, C., Kaplan, H. B., & Martin, S. S. (1985).
Lee, J. (1981). Covariance adjustment of rates based on the Antecedents of pregnancy among unmarried adoles-
multiple logistic regression model. Journal of Chronic cents. Journal of Marriage and the Family, 47, 567-583.
Disease, 34, 415-426. Rodgers, J. L., Rowe, D. C., & Buster, M. (1999). Nature,
MacCallum, R. C., Zhang, S., Preacher, K. J., & Rucker, D. nurture, and first sexual intercourse in the USA: Fitting
D. (2002). On the practice of dichotomization of behavioural genetic models to NLSY kinship data.
quantitative variables. Psychological Methods, 7, 19-40. Journal of Biosocial Science, 31, 29-41.
Father Absence 821
adolescent
Rothbart, M. K., & Bates, J. E. (1998). Temperament. In W. sexual activity. Journal of Marriage and the
Damon (Ed. in Chief) & N. Eisenberg (Vol. Ed.),61, 920-933.
Family,
Handbook of child psychology (pp. 105-176). NewWaynforth,
York: D. (2002). Evolutionary theory and reproduc-
Wiley. tive responses to father absence: Implications of kin
Rutter, M., Tizard, J., & Whitmore, K. (1970). Education, selection and the reproductive returns to mating and
health and behaviour. London: Longmans. parenting effort. In C. Tamis-LeMonda & N. Cabrera
Sampson, R. J., & Laub, J. H. (1990). Crime and deviance (Eds.), Handbook of father involvement (pp. 337-357).
over the life course: The salience of adult social bonds. Mahwah, NJ: Erlbaum.
American Sociological Review, 55, 609-627. Wilson, M., & Daly, M. (1997). Life expectancy, economic
Scaramella, L. V., Conger, R. D., Simons, L., & Whitbeck, L.inequality, homicide, and reproductive timing in Chi-
B. (1998). Predicting risk for pregnancy by late adoles- cago neighbourhoods. British Medical Journal, 314, 1271-
cence: A social contextual perspective. Developmental 1274.
Psychology, 34, 1233-1245. Woodward, L. J., & Fergusson, D. M. (1999). Early conduct
Silverstein, L. B., & Auerbach, C. E (1999). Deconstructing problems and later risk of teenage pregnancy in girls.
the essential father. American Psychologist, 54, 397-407. Development and Psychopathology, 11, 127-141.
Thornton, A., & Camburn, D. (1987). The influence of the World Health Organization. (1993). Composite International
family on premarital sexual attitudes and behavior.Diagnostic Interview (CIDI). Geneva, Switzerland: World
Demography, 24, 323-340. Health Organization.
Upchurch, D. M., Aneshensel, C. S., Sucoff, C. A., & Levy-
Storms, L. (1999). Neighborhood and family contexts of

Does Father Absence Place Daughters at Special Risk For Early Sexual Activity and Teenage Pregnancy?

Uploaded by

Copyright:

Available Formats

Does Father Absence Place Daughters at Special Risk For Early Sexual Activity and Teenage Pregnancy?

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Does Father Absence Place Daughters at Special Risk For Early Sexual Activity and Teenage Pregnancy?

Uploaded by

Copyright:

Available Formats

Does Father Absence Place Daughters at Special Risk for Early Sexual Activity and Teenage

other stressors associated with father absence (e.g.,

facilitate comparison with rates of early sexual

a combination of sources including: parentalcoding

birth. The mean age at first childbirth was12 months.

at least one pass grade in School Certificate: This

United States New Zealand

60% -.--------..------ -----------------------------------0--- 60%0 -

55% ........ ............... ................................ 55% ........ .... .............................

45% -------- ----- ------------..........- 45% ..................Rates of

35%......................--..---.------- 35% -............-E.- early sexual

Table 1 displays mean levels of child conduct

Father absence status Sexual activity Pregnancy status

Early Late Early No early

Variable absence absence presence F activity activity F Pregnant pregnant F

responsiveness). The strong pattern of covariation

followed by father-present girls (Figure 1). For

were no substantively meaningful linear relations

Timing of onset of father absence

Variable Early onset of Late onset of Father B (SE) z2 p Odds ratio

Note. Percentages after covariate adjustment are shown in parenth

equivocal and replicated only in the sense of being In

Timing of onset of father absence

Variable Early onset of Late onset of Father B (SE) p2 p Odds

Note. Percentages after covariate adjustment are shown in pa

The life-course adversity model has gained mechanisms

either a sensitive period or linear dose-response

Bates, J. E., Pettit, G. S., Dodge, K. A., & Ridge, B. correlates,

Fergusson, D. M., Horwood, L. J., & Lynskey, McLanahan,

You might also like