HHS Public Access
Author manuscript
Author Manuscript
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Published in final edited form as:
Soc Dev. 2018 August ; 27(3): 586–600. doi:10.1111/sode.12285.
Approach‐related emotion, toddlers' persistence, and negative
reactions to failure
Margaret Wolan Sullivan and Dennis P. Carmody
Rutgers – The State University of New Jersey, School of Nursing
Abstract
Author Manuscript
Author Manuscript
Approach behavior, defined as differences in behavior to an incentive event and anger at its
removal, was assessed during contingency learning in 87 5-month-olds was related to maternal
ratings of mastery behaviors at two years. Mothers reported on infants’ concurrent temperament,
as well as the occurrence of anger and tantrums, and their own anger at 12 months. Approach
behavior was expected to predict persistence with objects and persistent motor behavior, but not
negative reactions to failure. Negative reactions to failure were expected to be mediated by a
distress-prone temperament. The moderating effect of maternal anger on these relations was also
explored using conditional process regression models. Controlling for soothability, early approach
behavior predicted toddlers’ persistence, especially gross motor persistence, moderated by
maternal anger. With more maternal anger, approach behavior and toddler’s persistence were more
strongly related. Distress to limits, infant anger at 12 months, and maternal anger were
significantly correlated, but only infant anger was related to negative reactions to failure. Prior to
six months, goal-directed behavior is related to later behavioral persistence, but maternal responses
to child anger are an important contributor to this relation and by 12 months, infant anger directly
predicts mastery frustration at two years.
Author Manuscript
Evolutionary accounts and functional theories of emotion both agree that anger is
characterized by increased activation directed toward a blocked goal or obstacle and has
negative valence. As such, this adaptive form of anger has long been viewed as part of the
approach emotional system (Carver, Sutton, & Scheier, 2000; Gray, 1991; Greenberg, 2017;
Schneirla, 1959). Approach behaviors are motor actions associated with stimulus attention
(e.g. reaching, pulling to obtain objects), but they also have an emotional component.
Interest, curiosity, and determination are approach emotions, for example. Carver and
Harmon-Jones (Carver & Harmon-Jones, 2009; Carver, et al., 2000; E. Harmon-Jones &.C.
Harmon-Jones, 2013; C. Harmon-Jones, Schmeichel, Mennitt, & E. Harmon-Jones, 2011; E.
Harmon-Jones, Lueck, Fearn, & C. Harmon-Jones, 2006) have called attention to anger’s
unique aspect as a negative approach emotion. Yet, the view of anger as part of approach
motivation has been slow to influence the developmental literature on emotion, most likely
because anger has negative rather than positive emotional valence. Thus, approach-related
aspects of anger have not typically been examined. Our goal was to explore the relations
Correspondence can be addressed to M. W. Sullivan or D. P. Carmody, School of Nursing, 180 University Ave, Rutgers University,
Newark, New Jersey 07102-1803. Phone: 973-353-3854; Fax: 973-353-5810. sullivan@sn.rutgers.edu.
Disclosure Statement: Neither author has any conflict of interest to declare
Sullivan and Carmody
Page 2
Author Manuscript
between young infant’s approach-related emotion and other behaviors with which approachrelated emotion is theoretically or conceptually linked.
Author Manuscript
Infant contingency learning and anger responses to its disruption offer a way of assessing
approach behavior and approach-related anger prior to six months of age. This body of work
consistently has found that greater anger expression during contingency disruption (i.e., an
extinction procedure) and increased instrumental action occur when the contingent goal was
disrupted, although the two measures of increased activation and anger expression were
weakly correlated (Crossman, Sullivan, Hitchcock & Lewis, 2009; Sullivan & Lewis, 2003;
Sullivan, Lewis, & Alessandri, 1992). In other work, individual differences in the infants’
anger expressions to the blocked goal were related to interest expressions when the
contingent goal was restored (Lewis, Sullivan, Ramsay, & Alessandri, 1992). Additionally,
tantrums of 20-month-old toddlers, measured in terms of either onset or severity, were
unrelated to anger expressions at 5 months (Sullivan & Lewis, 2012).
Because these results support the view that individual differences in contingency response
and anger at its disruption (i.e., approach-related anger) may provide a measure of approach
behavior and emotion, we developed a general model to explore relations between infant
approach in this context with theoretically-linked behavioral sequelae such as behavioral
persistence and negative reactions to failed goal-directed activity in toddlers (See Figure 1).
Persistence at play with objects and during motor activities, the primary measures at two
years, were hypothesized to develop from early approach behavior. Negative reactions to
failure at two years, on the other hand, were expected to show different or possibly unique
mediation by early temperamental negativity, because they may be comprised of both poorly
regulated anger to blocked goals and distress.
Author Manuscript
Behavioral persistence at challenging or effortful tasks is a key aspect of mastery motivation
observed in toddlers and preschoolers (Busch-Rossnagel & Morgan, 2013; MacTurk &
Morgan, 1995; Messer, 1993). The behavior and emotion observed in contingency learning
and similar instrumental tasks is a hypothesized precursor of such mastery motivation with
later persistence reflecting the cumulative history of contingency experiences of infants prior
to 9 months (Dichter-Blancher, Busch-Rossnagel & Knauf-Jensen, 1997; MacTurk &
Morgan, 1995; Morgan, Wang, Liao & Xu, 2013). These studies and the linkage of
approach-related anger to subjective feelings of determination in adults (C. Harmon-Jones et
al., 2011) support the idea that individual differences in anger resulting from thwarted
approach may be related to later differences in persistence. We sought to model this relation
taking potential mediators of maternal behavior and child temperament into account.
Author Manuscript
Some empirical evidence on one form of persistence frequently observed in young children
has been related to early approach as well as to maternal behavior. Assertiveness or the
pursuit of one’s agenda despite interruption or competing demands, an early form of selfefficacy or autonomy-striving, was related to approach-related anger, but not to negativity or
protest to the disruption (Lewis, Sullivan & Kim, 2015). Although toddlers’ willfulness has
been linked to positive socialization techniques of mothers, including supportive maternal
behavior, autonomy-granting, and low depressive symptoms, Lewis et al. (2015) found no
consistent association with early maternal contingency. In other work, highly willful toddlers
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 3
Author Manuscript
also initiated more positive interactions with mothers than did low-willful toddlers, leading
the investigators to conclude that in toddlers, active resistance to parental demands reflects
children’s motivation to control events rather than poor parenting (Dix, Stewart, Gershoff, &
Day, 2007). Differences in approach-related anger expression specifically also have been
related to child surgency and attention in preschoolers (He et al., 2013). Such findings,
support the idea that early individual differences in approach behavior and emotion are
related to positive child characteristics and that maternal anger/annoyance would have
limited impact on mastery behaviors except at high levels of maternal anger.
Author Manuscript
Negative reactions to failure at two years of age is a measure of frustration at the inability to
accomplish goals, and reflects the negative aspect of mastery attempts. This construct is
related conceptually to tantrums in the second year of life. Tantrums at this age are thought
to reflect developmental characteristics of mastery and autonomy-striving (Birigen, Campos,
Emde, & Applebaum, 2008). Typically peaking between 18 and 24 months (Potegal &
Archer, 2004), tantrums are not considered problematic unless they persist beyond
toddlerhood and are protracted, self-injurious, or aggressive (Belden, Thomson, & Luby,
2008;; Carlson, Potegal, Margulies, Gutkovitch, & Wall, 2009; Lorber, Del Vecchio, & Slep,
2015 ). Severe tantrums after the age of 4–5 years are thought to indicate poor anger
regulation due either to underlying vulnerabilities or to problematic interactive processes
(Carlson, et al., 2009).). In addition to modeing persistence, we explored in a separate model
relations between early approach-related emotion and negative reactions to failure. We
expected negative reactions to failure to be inversely related to approach and to reflect more
negative temperament.
Author Manuscript
By including approach-related anger in our approach construct, we also sought to explore
several issues that complicate study of the development of individual differences in anger. A
key issue addressed by including anger expressions in the operational definition of infant
approach are that negative emotion has been viewed as undifferentiated throughout the first
year of life. Infant emotion research generally finds that anger and sad/distress expressions
co-occur to many emotion elicitors and that there is limited evidence for emotion-context
specificity (Camras & Shutter, 2010).
Author Manuscript
The temperament literature also tends to support the view of undifferentiated anger. Despite
the careful parsing of infant reactivity along several emotion-relevant dimensions, maternal
reports of temperament in early infancy tend to reduce statistically to three broad factors:
positive affectivity, negative affectivity, and orienting/regulatory capacity (Putnam et al.,
2014). At the same time, approach-related anger to contingency disruption has been
unrelated to maternal reports of temperament in any simple way, suggesting the constructs
may differ. Specifically, distress to limits, the dimension with greatest face validity with
anger/frustration, was unrelated to anger expression (Crossman et al., 2009). Likewise,
having a relatively more negative temperament contributed little variance to toddler
persistence, but persistence was predicted by approach-related anger measured during
contingency disruption (Lewis, et al., 2015). Although temperamental irritability during
contingency learning itself has been reported to have a modest relation to learning
performance, this relation was due to the irritability of non-learners included in the sample
(Lemelin, Tarabulsy, & Provost, 2002). Moreover, this relation held whether learning status
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 4
Author Manuscript
was operationally defined as a continuous learning score or as a dichotomous (yes/no)
variable. These investigators suggested that irritability during learning might reflect an
inability to manage incoming information, mediated by greater negativity to environmental
and task novelty, clearly temperament factors. How and when approach-related anger may
be distinguished from irritable anger/distress is a difficult problem. However, findings from
empirical work on emotional components of tantrums and their timing suggested a strategy
for exploring this issue in our model.
Author Manuscript
The constructs of distress to limits and negative reactions to failure may share features of
tantrum behavior which appears to include both anger and distress components, organized as
a bi-phasic process (Potegal & Davidson, 2003). Behaviorally, tantrums are comprised of
rapid-rising, intense anger followed by less intense, slower-fading distress. The anger
component of tantrums has an organized, hierarchical pattern of increased activation and
escalating vocal behavior to blocked play, empirically observable by 15 months of age
(Potegal, Robison, Anderson, Jordan, & Shapiro, 2007). The organization of tantrums as a
dynamic, bi-phasic process suggests that the initial anger of tantrums may be goal-focused.
Additionally, it may be that the secondary distress component is more indicative of
temperamental distress-proneness. Although it is unclear exactly when in development this
biphasic organization of anger and distress arises, its presence by15 months implies that the
anger component of tantrums might be separable from distress, and potentially from
temperamental negativity. Negative temperament seems to reflect either generalized negative
reactivity, or the shared variance between anger and distress processes. The contingency
learning and disruption procedure provided an opportunity to explore this issue at five
months, because the anger expressions observed upon contingency disruption are thought to
result from thrwarted approach rather than heightened distress.
Author Manuscript
Author Manuscript
To test the explanatory value of distress to limits in accounting for any observed relations
between approach and anger expressions and later toddler behavior, we included it as a
mediator in our models. To the extent that anger expression in response to contingency can
be accounted for by temperament, a mediating relation with distress to limits should be
observed (as shown in Figure 1). To the extent that anger/approach behavior is being
measured, distress to limits will not explain an approach-persistence relation. Similarly,
maternal report of infant anger severity at 12 months was included as a second mediator. We
expected 12-month anger to be at least moderately correlated with distress to limits,
reflecting some continuity with early temperament, but we also expected it might also reflect
the more differentiated organization of anger/distress reactions thought to emerge in the
second year. Potentially, different mediating effects might be observed in our models of the
relations between approach and the mastery-related outcomes of persistence and negative
reactions to failure.
We also considered whether any observed relation between approach and toddler behavior
was conditional on maternal responses to toddlers’ anger. We evaluated maternal anger as a
moderator of the hypothesized approach-outcome relations given the large literature on
interactive effects in early socialization. Among clinically-referred dyads, externalizing
problems in preschool-age children were predicted by earlier negative parenting and
mutually angry, hostile interactions, indicating that maternal anger to toddlers’ anger is a
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 5
Author Manuscript
risk factor (Davies & Cicchetti, 2014; Lorber & Egeland, 2009). Maternal emotion,
appraisals of child behavior, and physiological reactivity may lead to over-reactive,
inappropriate extremes of parenting (over-permissiveness, punitiveness or
unresponsiveness), eliciting child negativity as well as less effective behavioral management
by mothers (Lorber, 2012; Lorber, Mitnick, & Slep, 2016). If such processes are present in
non-clinical samples, it is important to know when they first influence child emotionality.
Interactions promoting the exacerbation of anger and tantrums are likely cumulative as
infants become increasingly mobile and independent. It was expected that maternal anger in
response to toddlers’ anger might exacerbate negative reactions at two (Lorber & Egeland,
2011). Influences on persistence were more exploratory, because primarily relations between
positive maternal behavior and persistence have been observed.
Author Manuscript
Finally, soothability, a temperament dimension that captures self-calming or ability to downregulate negative affect, has been related to differences in learning status and cardiacregulation (Crossman, et al., 2009; Sullivan, 2016). We therefore included it as a covariate
(not shown in Figure 1).
Author Manuscript
In summary, the hypotheses were: 1) controlling for soothability, approach behavior to
contingency and its disruption would be related to greater persistence, but not to greater
negative reactions to failure, consistent with demonstrated links between persistence and
toddler autonomy-striving (Dichter-Blancher, et al., 1997); 2) controlling for soothability,
distress-proneness and infants’ 12-month anger severity would be correlated and would
mediate the approach-persistence relation to the extent that infant distress-prone reactivity
accounts for any relation between approach-related anger and later persistence; 3)
controlling for soothability, a more distress-prone temperament and/or greater anger severity
at 12 months would be related to negative reactions to failure, mediating any observed
relation between approach and negative reactions to failure, and 4) maternal anger would
moderate the effect of infant approach such that negative reactions to failure would be
greater when maternal anger was greater. Differences in persistence were also expected to be
evident when maternal anger was greater, if moderation was observed.
Method
Participants
Author Manuscript
A convenience sample of 103 healthy newborns (≥37 weeks GA; birth weight > 5 lbs; 54%
female) and their mothers were recruited for study at University-affiliated teaching hospitals
prior to discharge by an experienced research assistant. The diverse sample represented the
surrounding metro/suburban area. Sixteen infants were missing data due to attrition at the
final visit (11) or incomplete or missing 12-month data (5), resulting in a final N of 87.
Infants lost to follow-up did not differ in age, gender, parity, maternal marital status,
ethnicity or other demographic characteristics from those who completed the study. Detailed
demographic information on the sample is provided in the supplementary material.
The attrition rate between each contact point was approximately 10%, within the expected
range for longitudinal visits at our laboratory. Infants with missing data did not differ
significantly in their behavior or their 5-month temperament ratings from those with
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 6
Author Manuscript
complete data. Five additional infants were excluded from the sample initially for health
issues (i.e., maternal expression of a developmental or health concerns, or diagnosis of a
medical or sensory problem including heart murmur; questionable vision or hearing).
Procedures
Author Manuscript
There were two laboratory visits (5 months, 24 months) and one interim phone contact (12
months). At the 5-month visit, each infant was observed in a standard learning procedure. A
detailed description is provided below. Birthday cards were sent when all children reached
12 months of age and mothers were reminded that they would be contacted via phone to
answer questions about their child’s emotions and development at this age. At 24 months,
toddlers received a standardized assessment to confirm that they were developing
appropriately while, mothers completed a questionnaire on tantrums and anger. The
procedures at each contact point were explained to the mothers who had the opportunity to
ask questions and sign IRB-approved consent forms.
Five-month assessment—Infants were scheduled when mothers expected them to be
alert and playful. Mothers completed the Infant Behavior Questionnaire [IBQ] (Rothbart &
Gartstein, 2000) to assess temperament. Infants (M= 23 weeks, SD = 2.0) were trained to
pull a ribbon to activate an audiovisual event using a well-established operant learning
procedure (Crossman et al., 2009; Sullivan, 2016; Sullivan et al., 1992). As in these studies,
rate of pulling (i.e., a motor activity whose goal was to access the contingent incentive) was
assessed during a period of non-incentive. Anger expressions during this period were
simultaneously scored. Increased pulling and anger displays in this context reflect behavioral
persistence in accessing a stimulus and were the measure of approach behavior (Lewis et al.,
2015).
Author Manuscript
Each infant sat in front of a monitor with mothers nearby, but out of view. A ribbon attached
to a soft elastic cuff was placed on the infants’ right wrist. The session consisted of three
phases: baseline, contingency, and disruption/extinction. In the initial 2-m baseline phase,
we recorded the frequency of pulling without incentive. This was followed by a 6-m
contingency learning phase during which each pull activated the 3-s audiovisual event. The
6-m period is sufficient to observe learning at this age, but brief enough to minimize
boredom or crying (Crossman et al., 2009; Sullivan, 2016). Any fussing, defined as
intermittent, voiced, low level negative affect during learning was timed via a stopwatch by
the research assistant. If fussiness escalated into continuous crying that lasted for more than
60 seconds, the session was terminated. Following this learning phase, infants experienced
another 1-m of non-reinforcement; the disruption/extinction phase.
Author Manuscript
The infant’s face and shoulders were recorded throughout the session for subsequent scoring
of facial expressions during the disruption. Two coders were trained to a minimum of 80%
reliability on pre-coded stills and digital videos from a prior study and were blind to infant
performance during the lab. Interim reliability checks were done on 10% of randomly
selected study participants. Kappa coefficients indicated that coders sustained above chance
agreement (κ = .87).
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 7
Author Manuscript
Infants’ anger expressions during the disruption phase were scored second by second using
MAX facial expression codes (Izard, 1995) as infants show little anger during contingency
learning (Crossman et al., 2009; Sullivan, 2016), but approximately 50–60% of infants who
learn tend to display anger during the disruption/extinction phase, ranging in duration from 1
to more than 53 seconds of anger (Crossman et al., 2009; Sullivan & Lewis, 2012).
Measures
Table 1 lists summarizes all the measures and lists the age at which they were collected.
Author Manuscript
Author Manuscript
Infant approach (5 months)—A composite approach score was derived from the
learning behavior (pulling) and the occurrence of anger expression when the contingency
incentive was disrupted. It reflected consistent access of the incentive event while also
accounting for the occurrence of anger expressions when the contingency was disrupted.
Infants’ pulling rate determined whether each individually sustained an increase in pulling
equal to or greater than 1.15 times their baseline average (M contingency/M baseline). This
minimum learning criterion discriminates reliably between those infants who rapidly
respond to the contingency in a single session and those who do not (Crossman et al., 2009;
Sullivan, 2016) and has been used to distinguish learners from nonlearners in other work
(Lemelin et al., 2002). Infants meeting criterion were responsive to the contingency, i.e.,
increased pulling, and therefore displayed approach behavior toward the stimulus. Those
who did not learn were categorized as having shown little or no approach and met the
criterion for excessive fussing during the contingency phase. Anger expression observed
during contingency disruption was divided into three categories: low and high anger by
median split and no anger during disruption. Infants were then grouped into five ordinal
categories reflecting less or more approach behavior. The ranked approach categories were:
1) those showing little or no approach (35% of the sample), 2) approach with increasing
fussiness during the session (13%), 3) approach and anger to disruption less than the median
(15%), 4) approach with anger to disruption above the median (17%), and 5) approach
without anger in response to disruption (20%). The pulling and anger expression data
corresponding to each of these categories is reported in Table 1.
Author Manuscript
Temperament (5 months)—This well-constructed, widely-used, reliable maternal report
of infant temperament allows mothers to rate six dimensions on a 1–7 scale (Rothbart &
Gartstein, 2000). For this study, the dimensions selected for analysis were Soothability and
Distress to Limits. Soothability is often included in composite positive temperament scores,
but it served in this study as a measure of infants’ ability to scale down-regulation of
negative affect, rather than as a measure of positive temperament. Soothability has been
associated with differences in cardiac regulation during learning contingency (Sullivan,
2016). Distress to Limits or negativity in response to restriction is a component of negative
temperament and has considerable face validity with tantrums and negative responses to
frustration. It also predicts later externalizing behavior in clinically-referred samples by 8
months of age (Lorber et al., 2014).
Twelve-month anger—At 12 months, mothers reported on tantrums and anger using an
instrument developed for this study (Sullivan, 2012). The 17 scenarios were selected from a
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 8
Author Manuscript
set of 25 that were piloted with an independent sample of 50 mothers of toddlers who rated
the frequency of occurrence of each scenario. The scenarios retained in the current study
were reported to occur at least weekly by 75% of mothers. To further insure that maternal
reports of their own anger were based on actual, rather than hypothetical, events in their
children’s lives, mothers in the study were also asked to report the frequency of occurrence
of each scenario. Only scenarios with non-zero occurrence were used to calculate maternal
anger severity in this study.
Author Manuscript
Mothers rated 1) the frequency of tantrum occurrence on a 7-point scale ranging from never
to more than once per day, and 2) the frequency of everyday contexts in which toddler’s
anger expressions might occur a 5-point scale, ranging from “Never or less than once per
week” to “Very often, every day or more than once per day”, as well as the severity of anger
in those contexts, also using a 5-point scale ranging from “No anger” to “Full tantrum”. The
twelve-month anger severity served as the measure of negativity at this age in addition to the
earlier rating of distress to limits at 5 months. Tantrum frequency was assessed for
descriptive purposes and for cross-validation of the 12-month anger measures. Tantrums
were defined for mothers as a sudden outburst of negative emotion in relation to a restriction
or the child’s inability to do something. Mothers were also provided a checklist of behaviors
to indicate those they typically observed to confirm that they were characterized by at
minimum, screams, shouts or intense crying.
Author Manuscript
The contexts rated for anger severity were 17 common, pretested scenarios in which toddlers
might become angry or frustrated (See supplementary material). Cronbach’s alpha,
calculated on 80 questionnaires with complete data on all items was .76. Deletion of any
single item from the analysis did not suppress alphas below .74. In addition, mothers also
rated their own anger/annoyance in each scenario from (1) not at all angry/annoyed to (5)
very angry. The ratings were averaged across reported scenarios to yield the average levels
of 12-month anger for the child and anger severity of the mother. Split-half reliability for
each scale was .80 for context frequency; .84 for child anger severity and .84 for maternal
anger. Reliability for tantrum frequency cannot be reported because it was a single item.
Only 12-month anger severity and maternal anger were included in the models.
Author Manuscript
Toddler Behavior—Persistent behavior and negative reactions to failure at two years of
age were assessed on the Dimensions of Mastery Questionnaire [DMQ] (Busch-Rossnagel
& Morgan, 2013). This 45-item questionnaire obtains maternal report of a toddler’s
motivation to persist at challenging tasks without undue frustration. Mothers rate several
behavioral domains of persistence (35 items) and Negative Reactions to Failure (5 items).
The scales have reported internal consistencies of .68 –.83, consistent with the range of
alphas observed in this study (.65 – .81). Parent and teacher ratings, as well as older
children’s self-ratings on DMQ correlate significantly with observed behavioral persistence
(Morgan, et al., 2013). The analyses focused on two scales that were conceptually similar to
the infant measure of persistent pulling and goal-directed anger in response to removal of the
contingency goal. Object-Related Persistence (Object P) and Gross Motor Persistence
(Gross Motor P), as well as the Negative Reactions to Failure scale. Thus, both persistence
(2 scales) and negative reactions to failed goal-directed behaviors were assessed as outcomes
in separate models.
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 9
Author Manuscript
Results
Data Analysis Plan
Author Manuscript
Table 1 summarizes all the measures used in modeling, their source, and the variables
derived from them. The distributions and descriptive statistics of all data were inspected
initially, and preliminary analyses were run to determine whether any demographic or family
variables (e.g., gender, birth order, number of hours of nonmaternal care, etc.) should be
included as covariates. None were significantly related to the study variables (all ps > .10).
Then, the bivariate correlations between these variables were inspected. Thereafter, three
conditional process regression models were assessed using techniques developed by Hayes
(2013) to evaluate the conceptual model (Figure 1) for each of the three outcomes and to
obtain population-based estimates of any observed mediation or moderation. A fourth model
predicting tantrum frequency was not significant (p > .05), so only the three models of the
DMQ persistence and negative reactivity scales are reported. Using G-POWER (Faul,
Erdfelder, Buchner, & Lang, 2009), the sufficient sample size for an alpha of 0.05, power of
0.80, and a medium effect size (0.15) was an N of 85.
Author Manuscript
Descriptive data at 5 months—The learning data were inspected to assess whether the
approach scores reflected the raw learning and emotion data. The means of the learning ratio
score (contingency response/baseline response) and anger expressions during disruption are
presented in Table 1 adjacent to the category descriptors. The ranked categories differed
significantly on each measure by Kruskal-Wallis test (p < .05 for both), but did not differ on
distress to limits or soothability ratings, supporting the approach categorization. The
correlations between the learning ratio score and distress to limits and soothability were
nonsignificant (r87 =.−.16 and .02, respectively). Correlations between infant anger
expressions and these temperament dimensions were.03 for soothability, and. −.04 for
distress to limits.
Author Manuscript
Descriptive data at 12-months—The onset and the frequency of tantrums reported by
mothers for this sample are summarized for comparison with other samples. The mean age
of tantrum onset was 12.26 months (SD = 0.37). Although 94% of the children reportedly
expressed anger at 12 months, tantrums were less frequent. The modal response (53%) was a
tantrum occurrence of “less than once per week”, with 2% having more than one per day.
Tantrum frequency at this age was related to Negative Reactions to Failure at two years (r88
= .25, p < .05). Tantrum frequency was modestly related to both 12-month anger severity
(r101 = .37, p < .01) and to infant distress to limits (r103 = .22, p < .05), but not to
soothability (−.17, ns), anger expression at 5 months (.21), the learning score (−.02, ns).
Maternal anger was also unrelated to tantrum frequency at 12 months (.02, ns). Means of 12month anger-severity and maternal anger, as well as their correlations with 5-month
variables, are reported in Table 2. Maternal anger was related to 5-month distress to limits
ratings as well as to 12-month anger.
Modeling toddler persistence and negative reactions to failure—Table 2 presents
the bivariate correlations between the variables used in the conceptual models and the means
and standard deviations for the continuous measures. All significance levels are two-tailed.
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 10
Author Manuscript
Correlations among the DMQ scale scores showed that Object P and Gross Motor P were
related to one another (p’s < .01), but were inversely related to Negative Reactions to
Failure, the relation with Gross Motor P being significant (p< .05). Object P and Gross
Motor P were retained as separate measures of persistence as they reflect different
behavioral domains and because pulling, a gross motor behavior, was the index of learning.
Because the bivariate correlations indicated that soothability was related to DMQ persistence
scores, but not to Negative Reactions to Failure, distress to limits or to 12-month anger
severity, it was retained as a covariate in the model. Distress to limits was entered as a
mediator along with its correlate, 12-month anger severity, in each of the models. Infant
approach behavior was the focal predictor of each of the DMQ outcomes. The models also
examined whether the observed relation between Approach and each DMQ outcome was
moderated when mothers expressed anger in response to their toddlers’ anger.
Author Manuscript
To locate effects, tests of indirect effects were conducted with each model. Interactions were
assessed using percentiles and bias-corrected, confidence intervals estimated using
bootstrapping (Hayes, 2013). The number of bootstrap samples used to generate these biascorrected intervals was 1,000. The final, summary model is described for each outcome.
Graphical presentation of interaction terms employs recommended techniques for locating
significant effects by plotting the conditional direct effect of Approach on each outcome at
10th, 25th, 50th, 75th, and 90th percentiles of Maternal Anger in the models (Hayes, 2013).
Infant Approach Emotion and Toddler’s Object Persistence (Object P)
Author Manuscript
Author Manuscript
Main hypothesis—We expected that persistence with objects would be related to infants’
early approach behavior, potentially moderated by maternal anger. The overall model
predicted persistence with objects (R2 = .17, F6,80 = 2.78, p = .02). Soothability, the
covariate, was related to Object P (Beta = .11, SE. = .05, t = 2.39, p < .02 [CI, .02:.20]). A
conditional effect of maternal anger was observed (Beta = −.38, SE. = 04, t = 2.38 p < .02
[CI, .02:.19]). The interaction between approach and maternal anger severity approached
statistical significance (Beta = .17, SE = .09, t = 1.92, p < .06, [CI, .01: 0.35]). Figure 2
shows these effect sizes with 95% confidence intervals plotted against the size of the
interaction. The effect estimate appears on the vertical axis and the moderator of parent
anger severity appears on the horizontal axis. A horizontal reference line indicates the effect
estimate of zero. The plotted lines show the mean value at selected points on the range of
parent anger severity in the model (specifically, the 10th, 25th, 50th, 75th and 90th percentiles)
as well as the 95% confidence intervals over the 10th through 90th percentiles. Where the
intervals do not include the effect estimate of zero, a significant moderation effect is found.
Maternal anger had little impact on the relation between approach style and persistence
except at the highest percentiles. Approach was related to Object P when maternal anger
level was higher. At this level of maternal anger, approach behavior enhanced Object P as
the confidence intervals exclude zero.
Mediation—We hypothesized that the direct relation between approach and persistence
might be accounted for in part by distress to limits and/or 12-month infant anger severity.
Tests of indirect effects found no evidence of mediation. That is, controlling for soothability,
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 11
Author Manuscript
neither infants’ distress to limits at 5 months nor 12-month anger severity accounted for the
observed relations between infant approach behavior and later persistence with objects.
Infant Approach Emotion and Toddler’s Gross Motor Persistence (Gross Motor P)
Main hypothesis—We expected that persistence in gross motor activities would be related
to infants’ early approach behavior, moderated by maternal anger. This regression model
also predicted Gross Motor P (R2 = .15, F6,80 = 2.28, p = .04). Soothability again directly
predicted greater persistence (Beta = .15, t = 2.61, p < .01 [CI, 0.14: 1.05]). A conditional
effect of maternal anger was again observed (Beta = .11, t = 2.16, p < .03). Figure 3 shows
these effect sizes with their respective 95% confidence intervals, plotted as described above.
When maternal anger was highest, its effect on the Approach/Gross Motor P relation was
strongest. At all other levels, the effect was nonsignificant as the confidence intervals
include zero.
Author Manuscript
Mediation—No mediation effects were observed. That is, controlling for soothability,
neither infants’ distress to limits at 5 months nor 12-month anger severity accounted for the
observed relations between infant approach behavior and later gross motor persistence.
Infant Approach and Negative Reactions to Failure
Main hypothesis—We expected that negative reactions to failure would be related to
infants’ early approach behavior, moderated by maternal anger. This model also significantly
predicted Negative Reactions to Failure (R2 = .20, F 6,80 = 3.30 p < .006). However, there
were no direct effects of soothability or approach, and no evidence of moderation by
maternal anger severity.
Author Manuscript
Mediation—The mediation hypothesis was unsupported for distress to limits or its
downstream correlate, 12-month anger severity. There was a direct effect of 12-month anger
severity on the outcome without significant mediation, however (Beta = .48 SE = .17, t =
2.84, p >.006, [CI, .14:.82]). Infants reported to express more anger to in common everyday
contexts at 12-months reacted more negatively when they were unsuccessful at two years of
age. Those in the highest quartile of 12-month anger were most likely to display more failure
reactions at 24 months; F3,91 = 7.55, p < 001, MQ1= 2.65, [CI’s 2.5:2.76] vs. MQ4. = 3.44
[CI’s 3.38: 3.50]. Quartiles 1–3 differed statistically from the highest quartile but not from
each other (Dunnett’s C: all p’s < .03, mean diff range: 0.43 – 0.70).
Discussion
Author Manuscript
Persistence and negative reactions to failure assess positive and negative emotional processes
when toddlers are unable to attain goals. We found differential predictability of these
constructs from infant measures of temperament and emotion. In this community sample,
appropriate approach responses to contingency at 5 months, indirectly predicted gross motor,
and to a lesser extent object-related persistence, but not negative reactions to failure. A
soothable temperament, a covariate, was also related to both forms of persistence, but not to
negative reactions to failure. Only 12-month infant anger directly predicted negative
reactions to failure.
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 12
Persistence
Author Manuscript
Soothability’s relation with both forms of persistence was expected as mastery motivation
has been associated with positive affect in that literature. Readily soothable infants may be
more able to focus their attention on external events and thus spend more time mastering
their own gross motor skills and interacting with objects. The ability to self-distract by
engaging with objects and motor play may also help dampen negativity.
Author Manuscript
Author Manuscript
Although approach-related emotion did not directly predict either form of persistence,
moderation by maternal anger was observed. The uppermost range of maternal anger
reported, which in this study was only mild, moderated effects of infants’ approach on both
object-related and gross motor persistence. An impact of mild levels of maternal anger on
persistence is consistent with considerable literature suggesting that maternal negative
emotion has an impact on children’s emotional behavior. However, it begs the question of
why a relation between approach behavior and persistence with objects and gross motor
behavior is enhanced. Two possible mechanisms may account for enhanced approach/
persistence relations. Approach behaviors reflect determination and autonomy-striving in
response to thwarted action (C. Harmon-Jones et al., 2011; Lewis et al., 2015; Sullivan,
2014). Low levels of maternal anger in response to toddlers’ behavior may elicit greater
child persistence through increased arousal or agitation. We do not know how mothers
express or display their annoyance, but this account assumes that they do in a way that is
perceived by the child. The literature also supports that infants are sensitive to other’s
emotion changes (Geangu, Hauf, Bhardwaj, & Bentz, 2011). A second possibility is that
mild maternal anger elicits similar levels of anger and determination in the child, an emotion
contagion process described as “coercive cycles” of negative interaction (Lorber & Egeland,
2011) . Child anger to unresponsive parenting is a demonstrated predictor of subsequent
externalizing behavior, but has not typically been studied in community samples (Davies &
Cicchetti, 2014).
Negative reactions to failure
Negative reactions to failure at two years are frustration at thwarted mastery attempts. They
were modestly associated in the bivariate correlations to tantrum frequency and to 5-month
distress to limits, but were related only to infants’ 12-month anger in the final model. They
were unrelated to early approach behavior. This pattern confirmed the conceptual coherence
of this construct with the 12-month rating with loss of efficacy whereas the correlation of
distress to limits with both these measures reflects shared variance with distress negativity.
Lack of mediation and alternative models
Author Manuscript
Inclusion of anger to contingency disruption as a component of approach behavior allowed
us to evaluate whether a specific temperament dimension, distress to limits, explained any
observed relation between early approach behavior and toddler’s behavior and negative
reactions to failure. The absence of any effect of distress to limits in any of the models was
unexpected because a distress-prone temperament has been shown to contribute to both
anger and/or tantrum behavior in clinically-referred samples (Lorber et al., 2014). Two
factors are suggested to account for the lack of mediation. Distress to limits, infant anger
severity, and maternal anger were all significantly correlated. If the mediating effect of either
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 13
Author Manuscript
infant variable is small, the moderating effect of maternal anger, or controlling for
soothability, may have outweighed them. Alternatively, the mediating effect was
hypothesized to be at the higher levels of distress to limits, so it may be that an insufficient
number of infants at the highest levels (11 infants in this community sample were above the
90th percentile) precluded our observing this effect.
Author Manuscript
Negative temperament cannot be ruled out in explaining variance in the observed relation
between 12-month anger severity and negative reactions to failure, however. We were unable
to test mediation of this relation due to the structure of our measurement plan. Our
assessment of distress to limits at 5 months preceded the 12-month anger assessment which
violates the required temporal sequence of mediators to be intervening variables or
contemporaneous with the predictors. In our model, 12-month anger was modeled only as a
downstream correlate of distress to limits, which was confirmed, suggesting some reliability
in maternal report across this age. Distress to limits was also related to maternal reports of
their own anger, suggesting that having a distress-prone infant may feed maternal anger over
time.
Limitations and implications
Author Manuscript
Overall this study illustrates that the transition between the first and second year is an
important time to study anger, tantrum onset, and maternal reactions to anger as well as the
contribution of temperament and maternal behavior to anger in relation to later mastery
behavior. The results imply that children who are high in approach, as assessed by their
behavioral and emotional responses to continency learning, may be differentially susceptible
to even mild levels of parental anger. Direct observation will need to confirm this
mechanism. All other measures were based on maternal report. The range of maternal anger
in this study was also limited. It is striking, that effects were observed at these low levels, but
under-reporting may have occurred. Further, whether and how this level of maternal
annoyance is communicated to the child cannot be addressed without behavioral measures.
On the other hand, the IBQ and the DMQ scales are reliable and well-validated. The toddler
anger questionnaire developed for the study also may have potential usefulness as a research
tool as the results provide some support for its validity. Observation of toddlers and mothers,
and the inclusion of reports from fathers or alternative caregivers, would allow replication of
the model, reduce self-report bias, and allow better identification of mechanisms.
Author Manuscript
Finally, for those interested in emotional development, the results imply that the field should
not be satisfied with the current consensus that early negative affect is undifferentiated.
Approach-related anger may be parsed from other forms of negative reactivity by focusing
on behavioral activation and anger expressions surrounding access to and loss of an
incentive event. Although this study did not use sophisticated statistical methods for deriving
an approach score (e.g., cluster analysis) due to limited sample size, the behaviors on which
our categorization was based were reliably observed and operationally defined. As such,
they are amenable to replication. The benefits of deriving such a measure, if replicable, is
that it would allow better understanding of 1) how undifferentiated negative emotion in the
first months of life becomes organized over the first two years of life, and 2) whether
individual differences in this organization are related to temperament or socialization factors.
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 14
Author Manuscript
The answers to these and related questions are of broad interest to those seeking to
understand the trajectories of emotional development in normal as well as clinical
populations.
Supplementary Material
Refer to Web version on PubMed Central for supplementary material.
Acknowledgments
This work was supported by NIMH grant #MH61778. We thank Michael Lewis for his insight and support during
the conduct of this study, Stacey Napoli, April Van Bergen, for their work in data collection and preparation data
and Angela Crossman, Ph. D. for performing the Bayley assessments.
This work was supported in part by NIMH Grant# xxxxx to the first author.
Author Manuscript
References
Author Manuscript
Author Manuscript
Belden AC, Thomson NR, Luby J. Temper tantrums in healthy vs. depressed and distruptive
preschoolers: Defining tantrum behaviors associated with clinical problems. Journal of Pediatrics.
2008; 152:117–122. DOI: 10.1016/j.jpeds.2007.06.030 [PubMed: 18154912]
Birigen Z, Campos JJ, Emde RN, Applebaum M. The development of autonomy: The role of walking
onset and its timing. Perceptual and Motor Skills. 2008; 106:395–414. DOI: 10.2466/pms.
106.2.395-414 [PubMed: 18556897]
Busch-Rossnagel NA, Morgan GA. Introduction to Section Three: Overview and analysis. In: Barrett
KA, Fox NA, Morgan GA, Fidler DJ, Daunhauer LA, editorsHandbook of self-regulatory processes
in development: New directions and international perspectives. New York, NY: Psychology Press;
2013. 247–264.
Camras LA, Shutter JM. Emotional facial expressions in infancy. Emotion Review. 2010; 2:120–129.
DOI: 10.1177/1754073909352529
Carlson GA, Potegal M, Margulies D, Gutkovich Z, Basile J. Rages--what are they and who has them?
Journal Of Child And Adolescent Psychopharmacology. 2009; 19:281–288. DOI: 10.1089/cap.
2008.0108 [PubMed: 19519263]
Carver CS, Harmon-Jones E. Anger is an approach-relate affect. Psychological Bulletin. 2009;
135:183–204. DOI: 10.1037/a0013965 [PubMed: 19254075]
Carver CS, Sutton SK, Scheier MS. Action, emotion, and personality: Emerging conceptual
integration. Personality & Social Psychology Bulletin. 2000; 26:741–751. DOI:
10.1177/0146167200268008
Crossman AM, Sullivan MW, Hitchcock DM, Lewis M. When frustration is repeated: behavioral and
emotion responses during extinction over time. Emotion. 2009; 9:92–100. DOI: 10.1037/a0014614
[PubMed: 19186920]
Davies PT, Cicchetti D. How and why does the 5-HTTLPR gene moderate associations between
maternal unresponsiveness and children’s disruptive problems? Child Development. 2014; 85:484–
500. DOI: 10.1111/cdev.12148 [PubMed: 24033129]
Dichter-Blancher TB, Busch-Rossnagel NA, Knauf-Jensen DE. Mastery motivation: Appropriate tasks
for toddlers. Infant Behavior and Development. 1997; 20:545–548.
Dix T, Stewart AD, Gershoff ET, Day WH. Autonomy and children’s reactions to being controlled:
Evidence that both compliance and defiance may be postive markers in development. Child
Development. 2007; 78:1204–1222. DOI: 10.1111/j.1467-8624.2007.01061.x [PubMed:
17650134]
Faul F, Erdfelder E, Buchner A, Lang AG. Statistical power analyses using G*Power 3.1: tests for
correlation and regression analyses. Behavior Research Methods. 2009; 41:1149–1160. DOI:
10.3758/BRM.41.4.1149 [PubMed: 19897823]
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 15
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Geangu E, Hauf P, Bhardwaj R, Bentz W. Infant pupil diameter changes in response to others’ positive
and negative emotion. PLoS ONE. 2011; 6:1–10. doi:www.plosone.org.
Gray JA. Neural systems, emotion and personality. Neurobiology of Learning, Emotion, and Affect.
1991; 4:273–396.
Greenberg G. Approach/Withdrawal Theory. In: Vonk J, Shachelford TK, editorsEncyclopedia of
Animal Cognition and Behavior. Springer International Publishers AG; New York: 2017. doi:
10:10071978-3-319-47829_1074
Harmon-Jones C, Schmeichel BJ, Mennitt E, Harmon-Jones E. The expression of determination:
Similarities between anger and approach-related positive affect. Journal of Personality and Social
Psychology. 2011 Mar.100:172–181. DOI: 10.1037/a0020966 [PubMed: 20853981]
Harmon-Jones E, Harmon-Jones C. What is approach emotion? Emotion Review. 2013; 5:291–295.
DOI: 10.1177/1754073913477509
Harmon-Jones E, Lueck L, Fearn M, Harmon-Jones C. The effect of personal relevance and approachrelated action expectation on relative left frontal cortical activity. Psychological Science. 2006;
17:434–440. DOI: 10.1111/j.1467-9280.2006.01724.x [PubMed: 16683932]
Hayes AF. Introduction to mediation, moderation, and conditional process analysis. New York: The
Guilford Press; 2013.
He J, Qiu P, Park K, Young, Xu Q, Potegal M. Young Chinese children’s anger and distress: Emotion
category and intensity identified by the time course of behaviors. International Journal of
Behavioral Development. 2013; 37:349–356.
Izard CE. The Maximally Disrciminative Facial Coding System (Revised). Newark, DE: University of
Delaware; 1995.
Lemelin J, Tarabulsy G, Provost MA. Relations between measures of irritability and contingency
detection at 6 months. Infancy. 2002; 3:543–554.
Lewis M, Sullivan MW, Ramsay D, Alessandri SM. Individual differences in anger and sad
expressions during extinction: Antecedents and consequences. Infant Behavior and Development.
1992; 15:443–452.
Lewis M, Sullivan MW, Kim HM. Infant approach and withdrawal in response to a goal blockage: Its
antecedent causes and its effect on toddler persistence. Developmental Psychology. 2015;
51:1553–1563. DOI: 10.1037/dev0000043 [PubMed: 26389608]
Lorber MF. The role of maternal emotion regulation in over-reactive and lax discipline. Journal of
Family Psychology. 2012; 26:642–647. DOI: 10.1037/a0029109 [PubMed: 22888786]
Lorber MF, Del Vecchio T, Slep AM. Infant externalizing behavior as a self-organizing construct.
Developmental Psychology. 2014; 50:1854–1861. DOI: 10.1037/a0036985 [PubMed: 24866289]
Lorber MF, Del Vecchio T, Slep AM. The emergence and evolution of infant externalizing behavior.
Development and Psychopathology. 2015; 27:663–680. DOI: 10.1017/S0954579414000923
[PubMed: 25212988]
Lorber MF, Egeland B. Infancy parenting and externalizing psychopathology from childhood through
adulthood: Developmental trends. Developmental Psychology. 2009; 45:909–912. DOI: 10.1037/
a0015675 [PubMed: 19586169]
Lorber MF, Egeland B. Parenting and infant difficulty: Testing a mutual exacerbation hypothesis to
predict early onset conduct problems. Child Development. 2011; 82:2006–2020. DOI: 10.1111/j.
1467-8624.2011.01652.x [PubMed: 22026438]
Lorber MF, Mitnick DM, Slep AM. Parents’ Experience of Flooding in Discipline Encounters:
Associations With Discipline and Interplay With Related Factors. Journal of Family Psychology.
2016; 30:470–479. DOI: 10.1037/fam0000176 [PubMed: 26690330]
MacTurk RH, Morgan GA. Mastery motivation: Origins, conceptualizations, and applications.
Westport, CT: Ablex; 1995.
Messer DJ. Mastery motivation in early childhood: Development, measurement and social processes.
New York: Routledge; 1993.
Morgan GA, Wang J, Liao H, Xu Q. Using the Dimensions of Mastery Questionnaire (DMQ) to assess
mastery motivation of English- and Chinese-speaking children: Psychometrics and implications
for self-regulation. In: Barrett KA, Fox NA, Morgan GA, Fidler DJ, Daunhauer LA,
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 16
Author Manuscript
Author Manuscript
Author Manuscript
editorsHandbook of self-regulatory processes in development: New directions and international
perspectives. New York, NY: Psychology Press; 2013. 305–335.
Potegal M, Archer J. Sex differences in childhood anger and aggression. Child and Adolescent
Psychiatric Clinics of North America. 2004; 13:513–528. DOI: 10.1016/j.chc.2004.02.004
[PubMed: 15183371]
Potegal M, Davidson RJ. Temper tantrums in young children: 1. Behavioral composition. Journal of
Developmental and Behavioral Pediatrics. 2003; 24:140–147. [PubMed: 12806225]
Potegal M, Robison S, Anderson F, Jordan C, Shapiro E. Sequence and priming in 15 month-olds’
reactions to brief arm restraint: evidence for a hierarchy of anger responses. Aggressive Behavior.
2007; 33:508–518. DOI: 10.1002/ab.20207 [PubMed: 17918278]
Putnam SP, Helig AL, Gartstein MA, Rothbart MK, Leerkes E. Development of short and very short
forms of the Infant Behavior Questionnaire-revised. Journal of Personality Assessment. 2014;
96:445–458. DOI: 10.1080/00223891.2013.841171 [PubMed: 24206185]
Rothbart MK, Gartstein MA. The Infant Behavior Questionnaire. 2000. Retrievable via https://
research.bowdoin.edu/rothbart-temperament-questionnaires/instrument-descriptions/the-infantbehavior-questionnaire/
Schneirla TC. An evolutionary and developmental theory of biphasic processes underlying approach
and withdrawal. In: Jones MR, editorNebraska Symposium on Motivation. Vol. 7. Lincoln, NB:
University of Nebraska Press; 1959. 1–42.
Sullivan MW. The Toddler Behavior Questionnaire. 2012. Copyright Rutgers University and M. W.
Sullivan. Unpublished instrument. Available from Rutgers Scholarly Open Access Repository.
http://dx.doi.org/doi:10.7282/T3028VNP
Sullivan MW. Infant expressions in an approach/withdrawal framework. J Genetic Psychology. 2014;
175:472–493. DOI: 10.1080/00221325.2014.964169
Sullivan MW. Vagal tone during infant contingency learning and its disruption. Developmental
Psychobiology. 2016; 58:366–381. DOI: 10.1002/dev.21376 [PubMed: 26517573]
Sullivan MW, Lewis M. Relations of Early Goal Blockage Response and Gender to Subsequent
Tantrum Behavior. Infancy. 2012; 17:159–178. DOI: 10.1111/j.1532-7078.2011.0007 [PubMed:
22408573]
Sullivan MW, Lewis M, Alessandri SM. Cross-age stability in emotional expressions during learning
and extinction. Developmental Psychology. 1992; 28:58–63.
Author Manuscript
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 17
Author Manuscript
Author Manuscript
Figure 1.
Author Manuscript
Author Manuscript
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 18
Author Manuscript
Author Manuscript
Figure 2.
Author Manuscript
Author Manuscript
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 19
Author Manuscript
Author Manuscript
Figure 3.
Author Manuscript
Author Manuscript
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Sullivan and Carmody
Page 20
Table 1
Author Manuscript
Overview of Measures and Variables by Contact Point
Measure
Source
Variable
Five Months (M age = 23 w SD 2.0)
Approach Score:
a. Online record of pulling during learning
phase
b. Coded seconds of MAX anger during
contingency disruption
Infant
Rank order (1–5) based on behavioral responses: 1. Little or no
approach with distress
Did not meet criterion; Pulling
rate is below baseline. No
disruption phase
M Learning ratio = 0.70 SE = 0.04;
> 60 s vocal distress during
contingency
2. Low approach
Learning criterion met, but
cried to criterion prior to
disruption. No disruption phase
M Learning ratio = 1.22 SE = 0.38;
> 60 s vocal distress during
contingency
3. Approach with low anger to disruption
Author Manuscript
Met criterion. Below median
anger (11.0) displayed to
contingency disruption.
M Learning ratio = 1.68; SE = 0.13
M Anger expression = 4.17 SE =
0.69
4. Approach with more anger to disruption
Met criterion. Above median
anger (11.0) displayed to
contingency disruption.
M Learning ratio = 2.25; SE = .03
M Anger expression = 27.30 SE =
0.69
5. High approach
Met criterion. Approach
behavior during disruption with
no observable anger expression.
Temperament:
a. Distress to Limits
b. Soothability
Maternal Report
M learning ratio = 12.57; SE =
5.81
M anger expression: 6.86 SE =
2.84
Rating on a 7-point scale of low to high on each dimension of the IBQ.
Twelve Months (M age = 13.03 m SD 2.05)
Author Manuscript
Child Anger Severity in everyday contexts
Maternal Report
Rating on a 5-point scale, ranging from little or no anger/annoyance to
extreme anger.
Maternal Anger in response to child’s anger in
the same contexts.
Self-Report
Rating on a 5-point scale, ranging from little or no anger/annoyance to
extreme anger.
Frequency of tantrums
Maternal Report
Rating on a 7-point scale ranging from none to > 1 per day.
Two Years (M age = 22.56 m SD 2.02)
Dimensions of Mastery Persistence subscales:
a. Object-related
b. Gross Motor
Maternal Report
Rating on a 5-point scale ranging from low to high.
Negative Reactions to failure on the DMQ
Maternal Report
Rating on a 5-point scale ranging from low to high
Author Manuscript
Soc Dev. Author manuscript; available in PMC 2019 August 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Table 2
Variable Name
2
3
4
5
6
Soc Dev. Author manuscript; available in PMC 2019 August 01.
1
Approach Rank
2
Soothability
3
Distress to Limits
−.04
.14
4
Anger Severity
−.08
.06
.52**
5
Maternal Anger
−.02
.04
.28*
.58**
6
Object Persistence
.18
.22*
−.22*
−.22*
−.19
7
Gross Motor
.13
.26*
−.03
−.16
−.14
.74**
8
Negative Reaction to Failure
−.09
−.09
.25*
.42**
.17
−.16
**
*
1
7
Mean (SD)
1.80 (1.50)
−.02
4.89 (1.03)
3.57 (0.84)
Sullivan and Carmody
Bivariate Correlations of Variables Used in the Models (N = 87)
3.52 (0.52)
1.49 (0.38)
3.53 (0.46)
3.89 (0.55)
−.27*
2.91 (0.60)
Correlation is significant at the 0.01 level (2-tailed).
Correlation is significant at the 0.05 level (2-tailed).
Page 21