Journal of Pediatric Psychology, 42(4), 2017, 445–456

doi: 10.1093/jpepsy/jsw085
Advance Access Publication Date: 1 October 2016
Original Research Article

A Longitudinal Study of Maternal and Child

Internalizing Symptoms Predicting Early
Adolescent Emotional Eating
Katherine M. Kidwell,1 MA, Timothy D. Nelson,1 PHD,

Downloaded from https://academic.oup.com/jpepsy/article/42/4/445/2670347 by guest on 08 September 2023

Jennifer Mize Nelson,1,2 PHD, and Kimberly Andrews Espy,1,3 PHD
1
Department of Psychology, University of Nebraska-Lincoln, 2Office of Research, University of Nebraska-Lincoln,
3
Department of Psychology, University of Arizona
All correspondence concerning this article should be addressed to Katherine Kidwell, University of Nebraska-
Lincoln, 238 Burnett Hall, Lincoln, NE 68588-0308, USA. E-mail: kkidwell@huskers.unl.edu
Received June 1, 2016; revisions received September 5, 2016; accepted September 6, 2016

Abstract
Objective To examine maternal and child internalizing symptoms as predictors of early adoles-
cent emotional eating in a longitudinal framework spanning three critical developmental periods
(preschool, elementary school, and early adolescence). Methods Participants were 170 children
recruited at preschool age for a longitudinal study. When children were 5.25 years, their mothers
completed ratings of their own internalizing symptoms. During the spring of 4th grade, children
completed measures of internalizing symptoms. In early adolescence, youth completed a measure
of emotional eating. Results Maternal and child internalizing symptoms predicted adolescent
emotional eating. The results indicated that child psychopathology moderated the association be-
tween maternal psychopathology (except for maternal anxiety) and early adolescent emotional eat-
ing. There was no evidence of mediation. Conclusions Pediatric psychologists are encouraged to
provide early screening of, and interventions for, maternal and child internalizing symptoms to pre-
vent children’s emotional eating.

Key words: anxiety; child internalizing; depression; emotional eating; maternal internalizing.

Emotional eating is defined as the consumption of a time of increased stress/emotionality and a time of
food in response to emotional cues rather than in re- greater independence with food choice (Bassett,
sponse to physical hunger cues (Evers, Stok, & de Chapman, & Beagan, 2008). Identifying the factors
Ridder, 2010). High emotional eating in youth has that lead to the development of emotional eating in
been correlated with overeating and consumption of youth is important for understanding the contribu-
high-fat foods, both of which have been associated tors to this potentially obesogenic behavior and in-
with risk for childhood obesity (Croker, Cooke, & forming pediatric obesity prevention efforts. In
Wardle, 2011; van Strien, Engels, van Leeuwe, & addition to specific child predictors (i.e., child inter-
Snoek, 2005). Emotional eating in youth is con- nalizing symptoms), the literature has examined three
sidered a learned behavior, based, in part, on the facets of parental influences on children’s emotional
finding that young children lost their appetites when eating: feeding styles, parenting styles, and parental
facing stressors, while older children and adolescents internalizing psychopathology. Although a number of
were more likely to overeat in this context (van child and parent factors have been explored, longitu-
Strien & Ouwens, 2007; van Strien, van der dinal studies examining how factors early in develop-
Zwaluw, & Engels, 2010). In fact, adolescence is a ment affect emotional eating in adolescence are
key period for studying emotional eating because it is lacking.

C The Author 2016. Published by Oxford University Press on behalf of the Society of Pediatric Psychology.
V
All rights reserved. For permissions, please e-mail: journals.permissions@oup.com 445
446 Kidwell, Nelson, Nelson, and Espy

Theoretical Framework satiety cues (as proposed by Farrow et al., 2015).

Emotional eating in children can be understood in a Other research has found that when parents minimized
contingent-reinforcement learning model, which ex- children’s negative emotions, children were more likely
plained children’s behaviors as being shaped by the to engage in emotional eating (Schuetzmann, Richter-
parental environment (Rodgers et al., 2014). The the- Appelt, Schulte-Markwort, & Schimmelmann, 2008;
ory proposed by Rodgers and colleagues (2014) was Topham et al., 2011). Finally, two studies by
that parents with greater internalizing problems Vandewalle and colleagues (2014, 2016) demonstrated
would use emotional eating to modulate their own that maternal rejection led to emotional eating in chil-
emotions (Spoor, Bekker, van Strien, & van Heck, dren and that children’s poor emotion regulation ex-
2007). Those parents who engaged in emotional eat- plained the association. Children with low emotion
ing would be more likely to use certain child-feeding regulation skills used food to cope with stress from ma-
practices that taught children to use food for emo- ternal rejection. It is likely that children with poor emo-

Downloaded from https://academic.oup.com/jpepsy/article/42/4/445/2670347 by guest on 08 September 2023

tional purposes rather than for physical hunger (e.g., tion regulation use food to manage other types of
using food to soothe and to reward; Blissett, Haycraft, stressors as well. Overall, the literature linking parent-
& Farrow, 2010; Farrow et al., 2015). Building on ing style and emotional eating suggests that parents
that theory, children who experienced emotional feed- have an important role in children’s eating behavior.
ing may have been more likely to engage in emotional In addition to feeding and parenting styles, parents’
eating as they became autonomous with food choices mental health symptoms may predict emotional eating
in early adolescence, because they have learned that in children. Most research has focused on maternal
food could be used to regulate emotions (Braden internalizing disorders predicting emotional eating in
et al., 2014). Youth with internalizing disorders may young children (Haycraft & Blissett, 2008; Haycraft &
be especially susceptible to emotional eating owing to Blissett, 2012; McPhie, Skouteris, Daniels, Jansen,
underlying emotion dysregulation that predisposes 2014; Ystrom, Barker, & Vollrath, 2012). For ex-
youth to cope with strong emotions through eating ample, mothers of 2-year-olds who experienced greater
(Czaja, Rief, & Hilbert, 2009). depression, anxiety, and stress were more likely to en-
gage in emotional eating, use emotional feeding tech-
niques with their children, and have children who
Parent Influences on Emotional Eating in Youth
engaged in emotional eating (Rodgers et al., 2014).
A substantial body of literature has examined the asso- Other research suggested that mothers with more inter-
ciation between parental feeding styles and emotional nalizing symptoms were more likely to pressure their
eating in children. Parental feeding styles have been children to eat, which undermined children’s natural
associated with increased emotional eating in children. hunger cues and taught emotional eating as a coping
For instance, offering food to soothe negative emo- mechanism (Ystrom et al., 2012). Little research has
tions in children has been shown to predict emotional examined the role of maternal internalizing symptoms
eating in young children (Braden et al., 2014; Farrow, on adolescent emotional eating. However, research by
Haycraft, & Blissett, 2015; Tan & Holub, 2015). Snoek and colleagues (2007) revealed that emotional
Moreover, when parents offered food to their children eating among parents was related to emotional eating
to calm them after a stressful task in an experimental among adolescents. They based their research design in
study, children consumed more cookies than those theories of social modeling to argue that adolescents
whose parents did not use food to soothe (Blissett, learned to engage in emotional eating through observa-
Haycraft, & Farrow, 2010). Similarly, Farrow and tion of parental behavior (Snoek et al., 2007).
colleagues (2015) demonstrated that 5–7-year-old
children were more likely to overeat in response to a
laboratory-based stressor if mothers rewarded chil- Child Influences on Emotional Eating
dren with food and had restricted access to children’s In addition to parent predictors, some research indi-
food 2 years prior. Children may have learned to use cated that children’s psychological health was associ-
food as a way to manage distress when parents either ated with emotional eating. For instance, youth aged
rewarded with food or restricted certain foods. 8–18 years old who were characterized as engaging in
Research on emotional eating in youth has also emotional eating at high levels were more likely to
examined the influence of broader parenting styles. For have internalizing disorders and to consume more
instance, low maternal support and high psychological food during an observed mealtime (Vannucci et al.,
and behavioral control have been implicated in produc- 2013). In another study, van Strien and colleagues
ing more emotional eating in children (Snoek, Engels, (2010) examined the relationships among depression
Janssens, & Van Strien, 2007). Behavioral control of symptoms, genetic risk for low levels of serotonin, and
food may have led to emotional eating in youth be- emotional eating in adolescents. The study found that
cause children stopped relying on their own hunger and adolescent depression symptoms were correlated with
Internalizing Symptoms and Emotional Eating 447

emotional eating, especially for those with a genetic Aims and Hypotheses
predisposition to have low levels of serotonin (van The current study had three aims. The first aim was to
Strien et al., 2010). The authors argued that early ado- examine maternal internalizing symptoms predicting
lescence was the ideal time to examine emotional eat- emotional eating in youth. It was hypothesized that
ing, because the rates of emotional eating are rather maternal internalizing symptoms early in a child’s de-
small in childhood but increase substantially in velopment would have long-term effects on emotional
adolescence. eating in early adolescence. This hypothesis was based
on previous findings of significant associations be-
tween maternal internalizing psychopathology and
emotional eating in children, but extended previous
Study Rationale
research into early adolescence in a longitudinal de-
Because research has primarily been conduced with sign. The second aim was to examine child internaliz-

Downloaded from https://academic.oup.com/jpepsy/article/42/4/445/2670347 by guest on 08 September 2023

young children, the research base has mainly focused ing symptoms predicting emotional eating. It was
on maternal feeding behaviors and the emergence of hypothesized that child internalizing psychopathology
emotional eating in children. A smaller subset of stud- would also predict higher emotional eating. Because
ies have also examined the associations between early early adolescents have more independence in their
maternal internalizing symptoms and children’s emo- food choices, it was expected that youth who experi-
tional eating when they were in preschool and early enced greater internalizing symptoms would be more
elementary school. Studies examining associations be- likely to soothe themselves with food. The third aim
tween maternal internalizing psychopathology and was to explore the interplay of maternal and child
children’s emotional eating later in development are internalizing symptoms by testing both mediator and
limited, and we are aware of no longitudinal studies moderator models. Theoretically, it was unclear
spanning early development to adolescence. In add- whether child internalizing psychopathology ex-
ition to examining the influence of maternal internal- plained the relationship between maternal internaliz-
izing psychopathology on emerging emotional eating, ing psychopathology and emotional eating in youth
it is also important to understand how children’s own (mediation) or whether child internalizing psycho-
internalizing problems may predict emotional eating pathology influenced the relationship (moderation).
patterns with the increasing independence of early Both mediator and moderator models were tested in
adolescence (Bassett et al., 2008). This study was built exploratory analyses to determine the relationship be-
on past findings by examining the relationships be- tween maternal and child internalizing psychopath-
tween maternal and child internalizing symptoms and ology on emotional eating.
early adolescent emotional eating in a longitudinal
framework spanning three critical times of develop- Method
ment (preschool, late elementary school, and early Participants and Procedures
adolescence). Participants were 170 children recruited at preschool
age for a longitudinal, lagged cohort-sequential study
that spanned from preschool through early adoles-
Interplay of Parent and Child Internalizing cence. Participants were recruited at a Midwestern
Symptoms Influencing Emotional Eating study site in the United States through flyers distrib-
uted at local preschools, doctors’ offices, the local
Moreover, maternal and child internalizing symptoms health department, and by word of mouth. Before en-
could interact to produce emotional eating in youth. rolling in the study, parents completed a telephone
For instance, child internalizing symptoms could ex- screening, and as the larger project focused on describ-
plain the relationship between maternal internalizing ing normal cognitive development in preschoolers,
symptoms and later emotional eating, such that moth- children with diagnosed developmental, behavioral, or
ers passed on a tendency for depressive and anxious language disorders at the time of initial recruitment
symptoms to their children (Buckholdt, Parra, & were excluded. To be eligible for the study, the pri-
Jobe-Shields, 2014), and children’s internalizing symp- mary language spoken at home had to be English.
toms led to increased emotional eating to cope with At study entry, mothers and children visited the la-
strong emotions (mediation). Alternatively, it is also boratory to complete measures. Mothers provided writ-
plausible that maternal internalizing psychopathology ten informed consent for participation. At the
exerted a unique influence on the risk for emotional preschool time point, children participated within 2
eating in youth, with the combination of maternal and weeks of turning 5.25 years. During this visit, mothers
child mental health symptoms having an interactive ef- completed ratings of their own mental health function-
fect on emotional eating in adolescence (moderation). ing, demographic information, and a background
448 Kidwell, Nelson, Nelson, and Espy

Table I. Descriptive Statistics and Correlations of Study Variables

Study variables N M SD Maternal Maternal Maternal Maternal Child Child

symptom distress somatization depression anxiety anxiety depression

Age at early adolescence 170 12.02 0.90 – – – – – –

Annual household income 170 55,953 35,563 – – – – – –
Maternal years of education 170 14.94 2.16 – – – – – –
Maternal symptom distress 162 50.74 7.49 – – – – – –
Maternal somatization 170 47.18 7.46 .44*** – – – – –
Maternal depression 170 47.91 7.63 .49*** .49*** – – – –
Maternal anxiety 170 46.02 8.72 .42*** .42*** .53*** – – –
Child anxiety 170 46.14 8.60 .069 .10 .14 .09 – –
Child depression 169 45.08 6.74 .24** .13 .32*** .19* .59*** –
Emotional eating 170 1.65 0.63 .18* .16* .10 .21** .16* .15*

Downloaded from https://academic.oup.com/jpepsy/article/42/4/445/2670347 by guest on 08 September 2023

***p < .001; **p < .01; *p < .05.

interview including questions about pregnancy, the sex (51.7% female) and oversampling on sociodemo-
home environment, daily routines, and cigarette smok- graphic risk (39.7% at risk). Sociodemographic risk
ing. Because mothers were thought to be the best re- was defined as eligibility for public medical assistance
porters of their own pregnancies, mothers were based on federal poverty guideline of family income to
targeted as the parent reporter. There was a 95.0% par- size. The mean total household annual income was
ticipation rate for mothers who agreed to complete $55,953 (SD ¼ $35,563). Refer to Table I for descrip-
their own measures. At the grade-school time point for tive statistics of continuous variables.
this study, children participated in the spring of 4th Of 180 participants who completed the study meas-
grade. During the 4th grade assessment, after parents ures in early adolescence, 170 youth had complete
provided consent and children provided verbal assent, data on study measures at all three time points
children completed a packet of measures including (94.4%). Nine had missing data on the parent mental
internalizing symptoms in a quiet room of the labora- health measure completed when children were 5.25
tory. In preschool and elementary school, parents were years. One was missing the child internalizing meas-
compensated for study participation with a $75 gift ures in 4th grade.
card per session, and children received small toys.
In early adolescence, parents provided consent and Measures
the youth provided assent to participate in the follow- Maternal Internalizing Symptoms
up study. The youth completed measures of health be- The Brief Symptom Inventory (BSI; Derogatis &
haviors including emotional eating in a quiet room of Melisaratos, 1983) is a 53-item self-report measure in
the laboratory. At the early adolescent time point, the which mothers rated the extent to which they had
youth ranged in age from 10.75 to 15.08 years been bothered (0 ¼ not at all to 4 ¼ extremely) by
(m ¼ 12.02, SD ¼ 0.90). Parents were compensated for various mental health symptoms in the past week. The
study participation with a $75 gift card, and early Positive Symptom Distress Index (hereafter, maternal
adolescents were compensated with a $25 gift card. symptom distress) measures the intensity of distress
The University of Nebraska-Lincoln institutional re- and was included as a broad measure of mental health
view board approved all procedures. problems in the analyses. To assess common specific
In terms of race and ethnicity, 64.9% children re- internalizing problem areas, the following subscales
ported as European American, 18.4% as multiracial, were included in analyses: somatization (e.g., “faint-
14.4% as Hispanic/Latino, 1.7% as African American, ness or dizziness”), depression (e.g., “feeling no inter-
and 1.0% as Asian American. The race and ethnicity of est in things”), and anxiety (e.g., “feeling tense or
the mothers were reported as the following: 85.0% keyed up”). Internal consistency ranged from accept-
mothers as European American, 7.2% as Hispanic/ able to excellent across the scales used in the current
Latino, 2.8% as African American, 2.0% as multira- study (a ¼ .95 for maternal symptom distress, .70 for
cial, 2.0% as Asian American, and 1.0% as American somatization, .81 for depression, and .71 for anxiety).
Indian. The majority of mothers (68.9%) were married.
In terms of maternal education, 2.9% of mothers had Child Anxiety
less than a high school diploma or equivalent, 9.8% The Revised Child Manifest Anxiety Scale–Short Form,
had a high school diploma or equivalent, 36.2% had second edition (RCMAS-2; Reynolds and Richmond,
attended some college, 41.4% had a college degree, 2008) is a measure of children’s anxiety appropriate for
and 9.8% had attended a graduate or professional pro- children aged 6–19 years (for a description see
gram. Recruitment involved stratification by children’s Huberty, 2012). The short form of the RCMAS-2 is a
Internalizing Symptoms and Emotional Eating 449

10-item questionnaire that is a valid measure of child emotional eating. The Preacher and Hayes (2008) SPSS
trait anxiety (Reynolds & Richmond, 2008). During Macro for Multiple Mediation used bootstrapping to es-
the spring of 4th grade, children answered questions timate the direct and indirect effects of maternal internal-
about their worries by circling either Yes or No for izing symptoms on emotional eating with a mediator of
each question (e.g., “I get nervous around people”). child depression or child anxiety. The Hayes and Mattes
The RCMAS-2 provides age-stratified norms based on (2009) MODPROBE Procedure macro was used to esti-
a nationally representative sample of 2,300 youth. The mate the moderation models with a maternal internaliz-
anxiety t-score was calculated using standard scoring ing psychopathology variable as the focal predictor,
procedures for the RCMAS and included in analyses. child depression or anxiety as the moderator, and emo-
Internal consistency of the RCMAS-2 was acceptable in tional eating as the dependent variable. A series of linear
the current study (a ¼ .72). regressions were estimated following the MODPROBE
Procedure to obtain beta coefficients for the moderator

Downloaded from https://academic.oup.com/jpepsy/article/42/4/445/2670347 by guest on 08 September 2023

Child Depression analyses presented in Table II. The predictor and moder-
The Children’s Depression Inventory 2 (CDI-2) is a 28- ators were mean centered before creating the interaction
item measure of child depression symptoms (Kovacs, term. Interactions were probed at the mean and 61 SD
2015). It is appropriate for children aged 7–17 years of the mean.
old, and the CDI’s psychometric properties are excel-
lent (for a review see Tobin, 2016). Children completed
Results
the CDI during the spring of the 4th grade by answer-
ing questions about symptoms experienced in the last 2 Preliminary Analyses
weeks on a 3-point Likert scale (e.g., “I feel sad”). The Before the regression analyses, preliminary analyses
depression t-score was included in the analyses. T- were conducted to determine the association of demo-
scores above 65 were clinically elevated and scores graphic factors with emotional eating. Gender, age,
above 55 were considered above average compared ethnicity, race, and maternal education did not signifi-
with children of the same age. Internal consistency of cantly correlate with emotional eating (p > .05). Given
the CDI-2 was acceptable in the current study (a ¼.79). the lack of association between demographic factors
and the outcome variable, demographics were not
Emotional Eating included in the subsequent regression models for parsi-
Emotional eating was measured using the Dutch mony. A post hoc power analysis for multiple regres-
Eating Behaviors Questionnaire, a commonly used 33- sion revealed that a sample size of 170 provided ample
item questionnaire, in adolescence (van Strien, 2002; power to find a medium effect. Refer to Table I for de-
van Strien, Frijters, Bergers, & Defares, 1986). Early scriptive statistics and correlations of study variables.
adolescents answered questions about emotional eat-
ing patterns on a 5-point Likert-type scale (e.g., “Do Main Analyses
you get the desire to eat when you are anxious, wor- Maternal Internalizing Symptoms and Emotional
ried, or tense?”). For the analyses, emotional eating in Eating
response to all emotions was used as the outcome Linear regression analyses were examined using the fol-
measure. Internal consistency in the current sample lowing predictors in separate models: maternal symp-
was excellent (a ¼ .92). tom distress, maternal somatization, maternal anxiety,
and maternal depression. First, maternal symptom dis-
Analysis Plan tress significantly predicted adolescent emotional eat-
Preliminary Analyses ing, ß ¼ .18, R2 ¼ .03,
Pearson’s correlations were calculated to determine F(1, 160) ¼ 5.58, p ¼ .019, with greater maternal
whether there were significant associations of demo- symptoms predicting greater adolescent emotional eat-
graphic factors with study variables. Post hoc power ana- ing. Second, the maternal somatization scale ac-
lyses for multiple regression were conducted to ensure counted for significant variance in adolescent
that there was enough power to find significant effects. emotional eating, ß ¼ .17, R2 ¼ .03, F(1, 169) ¼ 4.60,
p ¼ .001. Third, maternal anxiety significantly pre-
Main Analyses dicted adolescent emotional eating, ß ¼ .21, R2 ¼ .04,
A series of linear regression models were used to estimate F(1, 169) ¼ 7.41, p ¼ .007. Fourth, maternal depres-
the relationships among maternal internalizing symp- sion did not significantly predict adolescent emotional
toms, child internalizing symptoms, and youth emotional eating, ß ¼ .01, R2 ¼ .01, F(1, 169) ¼ 1.72, p ¼ .191.
eating. Moderator and mediator analyses were used to
determine whether child internalizing psychopathology Child Psychopathology Predicting Emotional Eating
was a moderator or mediator of the relationship between The next set of regression analyses investigated child
maternal internalizing psychopathology and youth internalizing symptoms and emotional eating.
450 Kidwell, Nelson, Nelson, and Espy

Table II. Summary of Moderator Regression Results With Emotional Eating as the Outcome

Predictor variables B SE ß DR2

Model 1
Step 1 .03*
Maternal somatization .01* .01 .17*
Step 2 .02
Maternal somatization .01* .01 .16*
Child anxiety .01* .01 .14*
Step 3 .02*
Maternal somatization .02* .01 .18*
Child anxiety .01* .01 .10*
Maternal somatization x child anxiety .001* .01 .16*
Model 2

Downloaded from https://academic.oup.com/jpepsy/article/42/4/445/2670347 by guest on 08 September 2023

Step 1 .03*
Maternal somatization .02* .01 .18*
Step 2 .02
Maternal somatization .01* .01 .16*
Child depression .01* .01 .13*
Step 3 .02*
Maternal somatization .02* .01 .18*
Child depression .01* .01 .15*
Maternal somatization x child depression .001* .001 .15*
Model 3
Step 1 .04*
Maternal anxiety .01* .01 .21*
Step 2 .02
Maternal anxiety .01* .01 .18*
Child anxiety .01 .01 .14
Step 3 .01
Maternal anxiety .01* .01 .19*
Child anxiety .01* .01 .15*
Maternal anxiety x child anxiety .001 .001 .10
Model 4
Step 1 .04**
Maternal anxiety .01** .01 .21**
Step 2 .01
Maternal anxiety .01* .01 .18*
Child depression .01 .01 .12
Step 3 .02
Maternal anxiety .01* .01 .19*
Child depression .02* .01 .18*
Maternal anxiety x child depression .001 .001 .14
Model 5
Step 1 .01
Maternal depression .01 .01 .10
Step 2 .02
Maternal depression .01 .01 .10
Child anxiety .01 .01 .15
Step 3 .03*
Maternal depression .01 .01 .10
Child anxiety .01* .01 .17*
Maternal depression x child anxiety .002* .001 .19*
Model 6
Step 1 .01
Maternal depression .01 .01 .10
Step 2 .02
Maternal depression .01 .01 .06
Child depression .01 .01 .14
Step 3 .05**
Maternal depression .01 .01 .08
Child depression .03** .01 .33**
Maternal depression x child depression .002** .001 .30**

Note.
**p < .01; *p < .05.
Internalizing Symptoms and Emotional Eating 451

Specifically, child anxiety symptoms (ß ¼ .16, R2 ¼ .03, variance in emotional eating, R2D ¼ .03, b ¼ .002,
F(1, 169) ¼ 4.32, p ¼ .039) and child depression symp- p ¼ .014. Next, child depression was included as the
toms (ß ¼ .15, R2 ¼ .02, F(1, 167) ¼ 4.05, p ¼ .040) moderator. Similarly, child depression significantly
were both significant predictors of emotional eating in predicted emotional eating, with the interaction term
early adolescence. accounting for a unique amount of variance in emo-
tional eating, R2D ¼ .05, b ¼ .002, p ¼ .003. The
Exploratory Analyses negative interaction terms indicated that when child
Mediation internalizing symptoms were low, children engaged
Mediation analyses were examined using the follow- most in the most emotional eating when their mothers’
ing predictors in separate models: maternal somatiza- symptoms were high. When maternal depression was
tion, maternal anxiety, and maternal depression. low, children with high anxiety/depression engaged in
Child anxiety or child depression were included as me- more emotional eating than those children with low

Downloaded from https://academic.oup.com/jpepsy/article/42/4/445/2670347 by guest on 08 September 2023

diators. Emotional eating was always included as the anxiety/depression. Refer to Figure 2 for the relation-
outcome. There was no evidence of mediation when ship between maternal depression and childhood de-
using nonparametric bootstrapping analyses that were pression and anxiety predicting emotional eating.
repeated 1,000 times.
Discussion
Moderation
The results of the moderation analyses are displayed The current study examined the relationships among
in Table II. Moderation was first examined with ma- maternal and child internalizing symptoms and emo-
ternal somatization as the focal predictor and child tional eating, as well as mediation/moderation models
anxiety as the moderator of emotional eating. When of how maternal and child internalizing symptoms
including maternal somatization as the predictor on interacted to produce youth emotional eating.
Step 1, child anxiety as the predictor on Step 2, and Consistent with the first hypothesis that early mater-
the centered interaction term on Step 3, there was sig- nal internalizing symptoms would have a long-term
nificant moderation. Each variable was significantly influence on youth emotional eating, the results dem-
predictive of emotional eating, with the interaction onstrated that both global and specific aspects of ma-
term accounting for a unique amount of variance in ternal internalizing problems predicted early
emotional eating, R2D ¼ .02, b ¼ .001, p ¼ .040. A adolescent emotional eating (with the exception of
similar pattern was observed for child depression as maternal depression). In support of the second hy-
the moderator (R2D ¼ .02, b ¼ .001, p ¼ .046), such pothesis that child internalizing symptoms would pre-
that when child internalizing symptoms were low, dict early adolescent emotional eating, the results
children engaged in the most emotional eating when indicated that both child depression symptoms and
their mothers’ symptoms were high. When maternal anxiety symptoms predicted later emotional eating.
somatization was low, children with high anxiety/de- Finally, the results indicated that child psychopath-
pression engaged in more emotional eating than those ology influenced the relationship (moderation) be-
children with low anxiety/depression. See Figure 1 for tween maternal psychopathology (except for maternal
the relationship between somatization and childhood anxiety) and early adolescent emotional eating. There
depression and anxiety predicting emotional eating. was no evidence of mediation.
Moderation was then examined with maternal anx- Most of the research on maternal internalizing psy-
iety as the focal predictor and child anxiety or child chopathology and child emotional eating has focused
depression as the moderator variable. When including on maternal feeding practices with young children
maternal anxiety as the predictor on Step 1, either (Haycraft & Blissett, 2008; Haycraft & Blissett, 2012;
child anxiety or child depression as the predictor on Ystrom et al., 2012). This is one of the first studies to
Step 2, and the centered interaction term on Step 3, examine the relationship between maternal internaliz-
there was not significant moderation, although mater- ing symptoms and early adolescent emotional eating.
nal anxiety, child depression, and child anxiety contin- Consistent with research that found that early mater-
ued to be significant predictors. Refer to Table II for a nal internalizing symptoms were associated with
summary of moderator regression results. young child emotional eating (McPhie et al., 2014),
Finally, moderation was examined with maternal the current study extended the findings to demonstrate
depression as the focal variable and child anxiety as a that maternal internalizing psychopathology early in
moderator. In the final model, there was significant development was a risk factor for later emotional eat-
moderation, but maternal depression was not signifi- ing in early adolescence. All examined predictors of
cantly related to emotional eating. Child anxiety sig- emotional eating in adolescence significantly predicted
nificantly predicted emotional eating, with the emotional eating except for maternal depression. It is
interaction term accounting for a unique amount of unclear why maternal depression did not predict
452 Kidwell, Nelson, Nelson, and Espy

Downloaded from https://academic.oup.com/jpepsy/article/42/4/445/2670347 by guest on 08 September 2023

Figure 1. Graph of the moderating effect of child anxiety (top figure) and depression (bottom figure) on the relationship
between maternal somatization and emotional eating.
Note. Low, average, and high child internalizing symptoms reflect 1 SD below the mean, the mean, and 1 SD above the mean, respectively.
Internalizing Symptoms and Emotional Eating 453

Downloaded from https://academic.oup.com/jpepsy/article/42/4/445/2670347 by guest on 08 September 2023

Figure 2. Graph of the moderating effect of child anxiety (top figure) and depression (bottom figure) on the relationship be-
tween maternal depression and emotional eating.
454 Kidwell, Nelson, Nelson, and Espy

emotional eating, but consistent with expectations, children whose mothers had minimal depression/
maternal anxiety and somatization predicted youth somatization, children who were highly depressed or
emotional eating. Rodgers and colleagues’ (2014) con- anxious engaged in more emotional eating than chil-
tingent reinforcement learning model theorized that dren who were low in depression and anxiety.
maternal internalizing disorders led to emotional feed- Overall, the results indicated that both child and ma-
ing of toddlers, because food was used to temper emo- ternal internalizing symptoms were unique risk factors
tions of both parents and their children. This study for later emotional eating, but they were not necessar-
extended that theory by demonstrating that adoles- ily cumulative.
cents whose mothers had internalizing disorders were The findings of the mediation analyses did not sup-
also more likely to emotionally eat. Perhaps children port the theory that maternal internalizing problems
who were emotionally fed are less in tune with in- would lead to child internalizing problems, which in
ternal physical hunger cues (Ystrom et al., 2012), turn would lead to adolescent emotional eating.

Downloaded from https://academic.oup.com/jpepsy/article/42/4/445/2670347 by guest on 08 September 2023

which promoted increased emotional eating across Maternal internalizing symptoms remained a signifi-
development. cant unique predictor of emotional eating even after
The finding that child internalizing problems in late controlling for child internalizing symptoms.
elementary school were predictive of emotional eating
in early adolescence is consistent with the Affect Implications
Regulation Model. The Affect Regulation Model de- Vast literature indicated that maternal depression and
scribes the occurrence of using food as a coping mech- anxiety were associated with poor child outcomes
anism to distract and reduce negative emotions (Burton, more broadly (Brennan et al., 2000). This study adds
Stice, Bearman, & Rohde, 2007; for a meta-analysis see to that literature while extending the findings to emo-
Haedt-Matt & Keel, 2011). Eating can be a way to tem- tional eating. Pediatric psychologists are encouraged
porarily numb uncomfortable emotions, including to assess for emotional eating and internalizing dis-
anger, fear, sadness, anxiety, loneliness, resentment, orders and to provide appropriate emotion regulation
and shame. For example, children experiencing bore- interventions. Early screening and treatment of mater-
dom and sadness may use food as entertainment or as a nal internalizing disorders are critical for many child
way to stay busy. This study demonstrated that emo- health outcomes (Glover, 2014), including emotional
tional eating continued into adolescence. eating. Practitioners can intervene with maternal emo-
This study is one of the first to examine the rela- tional functioning by treating internalizing disorders
tionship between maternal and child internalizing and providing education on modeling healthy coping
symptoms in predicting emotional eating. The ex- skills. Family and individual therapy may be useful for
ploratory models of moderation and mediation addressing psychopathology and emotional eating in
favored moderation, as there was an interaction of both mothers and their children. Because child anxiety
maternal and child internalizing symptoms in produc- and depression were also associated with emotional
ing early adolescent emotional eating. Child internal- eating, youth presenting for treatment of internalizing
izing symptoms moderated all associations between disorders could be taught healthy coping skills to re-
maternal internalizing and adolescent emotional eat- place emotional eating as a way to manage emotions
ing, except for models with maternal anxiety. It was (Hemmingsson, 2014).
surprising that maternal anxiety did not interact with In terms of research, much of the existing literature
child internalizing symptoms when child symptoms has examined family influences on the onset of emo-
influenced the relationships with maternal somatiza- tional eating (Blissett et al., 2010; Farrow et al., 2015;
tion and depression. Maternal internalizing symptoms Rodgers et al., 2014). Because research has primarily
influenced child emotional eating at differing levels of been conduced with young children, the research base
child internalizing symptoms; that is, the effects were has mainly focused on maternal feeding behaviors and
not additive, but multiplicative, suggesting that some the emergence of child emotional eating (Braden et al.,
children may be especially vulnerable to the effects of 2014; Tan & Holub, 2015). Less research has been
their mothers’ psychopathology. conducted with adolescents (Snoek et al., 2007).
The results of the moderation models indicated that Especially because emotional eating increases with age
maternal symptoms were most predictive of emotional (van Strien et al., 2010), longitudinal research is
eating in children who had lower levels of depression needed with older adolescents. Future researchers are
and anxiety. Specifically, mothers’ internalizing symp- encouraged to replicate this study and extend it into
toms had the most influence on emotional eating in samples of older adolescents.
children who were not depressed or anxious. Children
who exhibited high depression or anxiety were rela- Limitations and Future Directions
tively consistent in their emotional eating, with mater- Several limitations of the current study are noted.
nal depression not having a large influence. For First, the study had small effect sizes, which is
Internalizing Symptoms and Emotional Eating 455

common in research examining eating behaviors a novel addition to the literature. Third, parents and
owing to the many contributing factors to emotional children completed distinct, validated measures to re-
eating. Second, the research design examined the pres- duce common method variance that occurs when the
ence of internalizing symptoms with relatively brief, same person reports on each of the constructs.
but reliable, measures (BSI, CDI, RCMAS-2) rather
than conducting a thorough assessment of whether Conclusions
mothers and children had diagnosable, clinical inter-
In conclusion, this study found that early maternal
nalizing disorders. Future researchers may choose to
and child internalizing symptoms predicted later emo-
examine emotional eating in subsamples of youth pre-
tional eating in early adolescence. Moreover, child
senting with clinical depression and anxiety to build
internalizing symptoms moderated the association be-
on the current study that examined the constructs in a
tween maternal internalizing symptoms and youth
community sample. Third, the study design followed a
emotional eating. Pediatric psychologists are encour-

Downloaded from https://academic.oup.com/jpepsy/article/42/4/445/2670347 by guest on 08 September 2023

developmental trajectory in which the constructs were
aged to provide early screening of and interventions
measured at different occasions to preserve the tem-
for maternal and child internalizing symptoms to pre-
poral sequence of the theory (maternal internalizing
vent children’s emotional eating.
symptoms at 5.25 years ! child internalizing symp-
toms in 4th grade ! emotional eating in early adoles-
cence). Perhaps, if we had included internalizing Acknowledgments
disorders from the early adolescent occasion in the We thank the participating families and acknowledge the in-
analyses rather than the 4th grade time point, more valuable assistance with data collection and coding by re-
internalizing symptoms would have been endorsed, search technicians and undergraduate and graduate students
because internalizing disorders increase in prevalence of the Developmental Cognitive Neuroscience Laboratory at
the University of Nebraska-Lincoln.
as children age. Further, the ability to test alternative
models, such as co-emergence of traits/symptoms or
reverse causal sequences, was limited by having the Funding
constructs measured at different time points. Future This work was supported by National Institutes of Health
research can examine the association between adoles- (grant number MH065668) and an award from the Office of
cent internalizing symptoms and emotional eating at Research, College of Arts and Sciences and Department of
multiple time points to augment the results of the cur- Psychology at the University of Nebraska-Lincoln.
rent study. A fourth limitation is that emotional eating
was not assessed in childhood and then controlled for in Conflicts of interest: None declared.
the analyses. This limits the conclusions that can be
drawn in terms of longitudinal relationships and causal- References
ity. A fifth limitation is that this study included only Bassett, R., Chapman, G. E., & Beagan, B. L. (2008).
mothers, and future work examining the role of fathers’ Autonomy and control: The co-construction of adolescent
psychopathology would be valuable. Finally, future re- food choice. Appetite, 50, 325–332.
searchers are encouraged to explore gender differences Blissett, J., Haycraft, E., & Farrow, C. (2010). Inducing pre-
in the relationships. In the current sample, there were school children’s emotional eating: Relations with parental
not significant gender differences in emotional eating, feeding practices. The American Journal of Clinical
but studies with larger samples may have more power Nutrition, 92, 359–365.
Braden, A., Rhee, K., Peterson, C. B., Rydell, S. A., Zucker,
to explore gender differences. Sixth, the theory by
N., & Boutelle, K. (2014). Associations between child
Rodgers and colleagues (2014) posited that parents emotional eating and general parenting style, feeding prac-
with greater internalizing problems would use emo- tices, and parent psychopathology. Appetite, 80, 35–40.
tional eating to modulate their own emotions and use Brennan, P. A., Hammen, C., Andersen, M. J., Bor, W.,
feeding practices that promoted emotional eating in Najman, J. M., & Williams, G. M. (2000). Chronicity, se-
children. Maternal emotional eating was not assessed in verity, and timing of maternal depressive symptoms:
this study, and therefore this aspect of the theory was Relationships with child outcomes at age 5.
not tested and should be examined in future research. Developmental Psychology, 36, 759–766.
Despite the limitations of the current study, there Buckholdt, K. E., Parra, G. R., & Jobe-Shields, L. (2014).
are many strengths that expand on previous research to Intergenerational transmission of emotion dysregulation
through parental invalidation of emotions: Implications
produce novel findings. First, the extended longitudinal
for adolescent internalizing and externalizing behaviors.
design is a contribution to this research that, to our Journal of Child and Family Studies, 23, 324–332.
knowledge, is new. Second, to preserve the temporal se- Burton, E., Stice, E., Bearman, S. K., & Rohde, P. (2007).
quence of the developmental trajectory, the constructs Experimental test of the affect-regulation theory of bulimic
of interest were examined at three separate time points symptoms and substance use: A randomized trial.
(preschool, 4th grade, and early adolescence), which is International Journal of Eating Disorders, 40, 27–36.
456 Kidwell, Nelson, Nelson, and Espy

Croker, H., Cooke, L., & Wardle, J. (2011). Appetitive be- Schuetzmann, M., Richter-Appelt, H., Schulte-Markwort, M.,
haviors of children attending obesity treatment. Appetite, & Schimmelmann, B. G. (2008). Associations among the per-
57, 525–529. ceived parent–child relationship, eating behavior, and body
Czaja, J., Rief, W., & Hilbert, A. (2009). Emotion regulation weight in preadolescents: Results from a community-based
and binge eating in children. International Journal of sample. Journal of Pediatric Psychology, 33, 772–782.
Eating Disorders, 42, 356–362. Snoek, H. M., Engels, R. C., Janssens, J. M., & van Strien, T.
Derogatis, L. R., & Melisaratos, N. (1983). The brief symp- (2007). Parental behaviour and adolescents’ emotional eat-
tom inventory: An introductory report. Psychological ing. Appetite, 49, 223–230.
Medicine, 13, 595–605. Spoor, S. T., Bekker, M. H., Van Strien, T., & van Heck, G.
Evers, C., Stok, F. M., & de Ridder, D. T. (2010). Feeding your L. (2007). Relations between negative affect, coping, and
feelings: Emotion regulation strategies and emotional eating. emotional eating. Appetite, 48, 368–376.
Personality and Social Psychology Bulletin, 36, 792–804. Tan, C. C., & Holub, S. C. (2015). Emotion regulation feed-
Farrow, C. V., Haycraft, E., & Blissett, J. M. (2015). ing practices link parents’ emotional eating to children’s

Downloaded from https://academic.oup.com/jpepsy/article/42/4/445/2670347 by guest on 08 September 2023

Teaching our children when to eat: How parental feeding emotional eating: A moderated mediation study. Journal
practices inform the development of emotional eating—a of Pediatric Psychology, 40, 657–663.
longitudinal experimental design. The American Journal Tobin, R. (2016). Test review of the Children’s Depression
of Clinical Nutrition, 101, 908–913. Inventory (2nd ed.). In J. F. Carlson, K. F. Geisinger, & R.
Glover, V. (2014). Maternal depression, anxiety and stress A. Spies (Eds.), The nineteenth mental measurements year-
during pregnancy and child outcome: What needs to be book. Retrieved from the Buros Institute’s Test Reviews
done. Best Practice & Research Clinical Obstetrics & Online website: http://www.buros.org/
Gynaecology, 28, 25–35. Topham, G. L., Hubbs-Tait, L., Rutledge, J. M., Page, M. C.,
Haedt-Matt, A. A., & Keel, P. K. (2011). Revisiting the affect Kennedy, T. S., Shriver, L. H., & Harrist, A. W. (2011).
regulation model of binge eating: A meta-analysis of stud- Parenting styles, parental response to child emotion, and
ies using ecological momentary assessment. Psychological family emotional responsiveness are related to child emo-
Bulletin, 137, 660–681. tional eating. Appetite, 56, 261–264.
Haycraft, E., & Blissett, J. (2008). Controlling feeding prac- Vandewalle, J., Moens, E., Beyers, W., & Braet, C. (2016).
tices and psychopathology in a non-clinical sample of Can we link emotional eating with the emotion regulation
mothers and fathers. Eating Behaviors, 9, 484–492. skills of adolescents? Psychology & Health, 31:857–872.
Haycraft, E., & Blissett, J. (2012). Predictors of paternal and Vandewalle, J., Moens, E., & Braet, C. (2014).
maternal controlling feeding practices with 2-to 5-year-old Comprehending emotional eating in obese youngsters: The
children. Journal of Nutrition Education and Behavior, role of parental rejection and emotion regulation.
44, 390–397. International Journal of Obesity, 38, 525–530.
Hayes, A. F., & Mattes, J. (2009). Computational proced- Vannucci, A., Tanofsky-Kraff, M., Crosby, R. D.,
ures for probing interactions in OLS and logistic regres- Ranzenhofer, L. M., Shomaker, L. B., Field, S. E., . . .
sion: SPSS and SAS implementations. Behavior Research Yanovski, J. A. (2013). Latent profile analysis to determine
Methods, 41, 924–936. the typology of disinhibited eating behaviors in children
Hemmingsson, E. (2014). A new model of the role of psycho- and adolescents. Journal of Consulting and Clinical
logical and emotional distress in promoting obesity: Psychology, 81, 494–507.
Conceptual review with implications for treatment and van Strien, T. V. (2002). Dutch Eating Behaviour
prevention. Obesity Reviews, 15, 769–779. Questionnaire: Manual. Bury St. Edmunds: Thames Valley
Huberty, T. J. (2012). Assessment of anxiety. Anxiety and Test Co.
Depression in Children and Adolescents, 193–218. van Strien, T., Engels, R., Van Leeuwe, J., & Snoek, H.
Kovacs, M. (2015). Children’s Depression Inventory (CDI (2005). The Stice model of overeating. Tests in clinical and
and CDI 2). The Encyclopedia of Clinical Psychology. non-clinical samples. Appetite, 45, 205–214.
Published Online: 23 JAN 2015. DOI: 10.1002/ van Strien, T., Frijters, J. E. R., Bergers, G. P. A., & Defares,
9781118625392.wbecp419 P. B. (1986). The Dutch Eating Behavior Questionnaire
McPhie, S., Skouteris, H., Daniels, L., & Jansen, E. (2014). (DEBQ) for assessment of restrained, emotional, and ex-
Maternal correlates of maternal child feeding practices: A ternal eating behavior. International Journal of Eating
systematic review. Maternal & Child Nutrition, 10, Disorders, 5, 295–315.
18–43. van Strien, T., & Ouwens, M. A. (2007). Effects of distress,
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and alexithymia and impulsivity on eating. Eating Behaviors,
resampling methods for estimating and comparing indirect 8, 251–257.
effects. Behavior Research Methods, 40, 879–891. van Strien, T., van der Zwaluw, C. S., & Engels, R. C.
Reynolds, C. R., & Richmond, B. O. (2008) Revised Children’s (2010). Emotional eating in adolescents: A gene (SLC6A4/
Manifest Anxiety Scale, Second Edition (RCMAS-2): 5-HTT)–Depressive feelings interaction analysis. Journal
Manual. Los Angeles, CA: Western Psychological Services. of Psychiatric Research, 44, 1035–1042.
Rodgers, R. F., Paxton, S. J., McLean, S. A., Campbell, K. J., Ystrom, E., Barker, M., & Vollrath, M. E. (2012). Impact of
Wertheim, E. H., Skouteris, H., & Gibbons, K. (2014). mothers’ negative affectivity, parenting locus of control
Maternal negative affect is associated with emotional feed- and child-feeding practices on dietary patterns of 3-year
ing practices and emotional eating in young children. old children: The MoBa Cohort Study. Maternal and Child
Appetite, 80, 242–247. Nutrition, 8, 103–114.