Psychology
holiday Jessica
Supervising Children During Parental Distractions
Richard E. Boles,1 PHD and Michael C. Roberts,2 PHD 1Cincinnati Children’s Hospital Medical Center, Cincinnati and 2Clinical Child Psychology Program,
University of Kansas
Objective To examine the effects of parenting distractions on supervising behaviors in relation to child risky
behaviors. Methods Forty preschool-aged children and their parents were randomly assigned to occupy a
simulated home living room for 45 min with the parent involved in either (a) no planned distraction, (b) a
telephone call distraction, (c) a TV show distraction, or (d) a computer assignment distraction. Parent and
child behaviors were recorded and coded. Results Parent supervising behaviors were significantly
intercorrelated but revealed no relation to risky child behavior. Children showed higher risky behavior during
parental distractions and steadily over time when parent distractions occurred. Additionally, younger children
were more likely to engage in risky behavior when compared to older children. Conclusions Parents
showed significant reductions in their ability to supervise children during distractions, limiting the ability to
provide education or to take immediate action necessary to prevent or minimize possible injuries.
Key words childhood injuries; observation; parent distractions; supervision.
Unintentional injuries are the most significant health
challenge for children and adolescents from age 1 to 19
years with home injuries a major proportion of the
problem (Guyer et al., 1999; National Center for Injury
Prevention and Control, 2001, 2002). For preschool
children, home injuries likely result from child, parent,
and environmental variables interacting within family and
cultural contexts. The extent to which these variables
contribute to the occurrence of an injury, however, is still
poorly understood. (National Center for Injury and
Prevention and Control, 2002). One limitation to a
multi-factorial understanding has been the limited empiri-
cal knowledge on child supervision, an often cited parent-
related factor in injuries among young children (Cataldo,
Finney, Richman, & Riley, 1992; Gärling & Gärling, 1993;
Morrongiello & Dawber, 1998). Only recently have
researchers begun to explore supervision in relation to
pediatric injuries, providing conceptual models, assess-
ment instruments, and methodological guidance for
empirical investigations (Morrongiello, Corbett, McCourt,
& Johnston, 2006; Morrongiello & House, 2004; Saluja
et al., 2004).
The Relationship of Supervision to Child Injury
Currently a dearth of knowledge exists regarding the actual
practices of caregivers during interactions with children in
the home. Emerging evidence has shown that closer
supervision provides a protective role and is linked with
fewer child injuries (Morrongiello, Ondejko, & Littlejohn,
2004). Additionally, parents increase their supervision
with children who are reported as risk takers, sensation
seeking, and impulsive (Morrongiello et al., 2006).
However, behavioral observations of parents supervising
children on playgrounds have shown only limited relation
to injury risk (Morrongiello & House, 2004), and much
less is known about the relation of observed supervising
behaviors within a home environment and child risk
behaviors. For example, in a study conducted by
researchers at Safe Kids from 2000–2001, it was shown
that 88% of drowning victims were being supervised by
caregivers (Cody, Quraishi, Dastur, & Mickalide, 2004).
Among parents who supervise children during swimming,
distractions were reported as common occurrences
All correspondence concerning this article should be addressed to: Richard E. Boles, PhD, Cincinnati Children’s Hospital Medical Center, Division of Behavioral Medicine and Clinical Psychology, MLC 3015, 3333 Burnet Avenue, Cincinnati, OH 45229-3039, USA. E-mail: [email protected]
Journal of Pediatric Psychology 33(8) pp. 833–841, 2008 doi:10.1093/jpepsy/jsn021
Advance Access publication March 10, 2008 Journal of Pediatric Psychology vol. 33 no. 8 � The Author 2008. Published by Oxford University Press on behalf of the Society of Pediatric Psychology.
All rights reserved. For permissions, please e-mail: [email protected]
D ow
nloaded from https://academ
ic.oup.com /jpepsy/article-abstract/33/8/833/923482 by A
rizona S tate U
niversity W est user on 24 N
ovem ber 2019
[e.g., talking to someone (38%), reading (18%), eating
(17%), and using the phone (11%)]. Supervision investiga-
tions have not systematically explored how parents modify
their supervision practices during typical home distrac-
tions. This is a key factor when considering most parents
are at home when their children are becoming injured
(Shannon, Brashaw, Lewis, & Feldman, 1992).
Saluja et al. (2004) provided a conceptual model of
parent supervising behaviors in which a hierarchy of
supervision strategies defines supervision in a way that
can be systematically measured for a relation to injury
risk. Three dimensions were described that included
(a) attention, (b) proximity, and (c) continuity. Attention
encompassed the level of engagement between the super-
visor and child in addition to visual and auditory
components, ranging from directly visually focused to no
visual contact with the child. Proximity indicated the
distance between the supervisor and child, ranging from
touching ability to out of reach and beyond reach at large
distances. Continuity described the extent to which the
supervisor demonstrates the first two dimensions in a
continuous, intermittent, or nonexistent (absent)
approach. The present study attempts to test the critical
elements of the Saluja et al. (2004) model of parent
supervision by measuring multiple observable behavioral
components: (a) attention (i.e., visual attention based on
ability to make eye contact with the child), (b) engagement
(i.e., verbal and/or physical interactions between the child
and parent, and (c) proximity (i.e., how close the parent is
to the child).
Primary Study Goals and Hypotheses
The primary purpose of the present study is to identify
parenting behaviors associated with risky child behaviors
within a simulated home setting, with consideration of the
influence of parent distractions in order to understand
potential changes in parenting behaviors when their
attention shifts away from the child. Our goal was to
address an existing gap in the literature regarding the link
among parent supervision and child behaviors based on
observations within a home setting. Previous studies
utilizing observational methodologies to study injuries
helped shape the present approach (Cataldo et al.,
1992; Morrongiello & Dawber, 1998; Morrongiello &
House, 2004). Due to the potential for participants to
change their behavior after being asked to participate
in an observational study (Haynes & Horn, 1982), we
explored the effects of masking the study aims for half
of the sample. Although the literature is not entirely
consistent, most rigorous studies have shown limited
reactivity effects (Jacob, Tennenbaum, Seilhamer, Bargiel,
& Sharon, 1994).
We hypothesized that parenting behaviors would
change during the distraction periods. Specifically, we
expected engagement and eye contact to decrease between
parents and children during the parent distractions while
proximity would increase during distractions. Additionally,
we hypothesized an inverse association between risky child
behaviors and parental supervision behaviors. Finally, we
expected minimal differences of reactivity between
informed and uninformed groups.
Methods Participants
Participants were 40 parent–child dyads aged 2 to 5 years
(M ¼ 4.4, SD ¼ 1.1). Parents were primarily Caucasian
(93%) and well educated with all having attended some
college and 62.5% had obtained a bachelors degree or
higher. Participating parents were most often the mother
(85%) with an average age of 32.6 years (SD ¼ 5.7).
A power analysis determined that 40 participants were
adequate to detect for possible effect sizes.
Procedure
Parents and their children were recruited from a mid-
western preschool population. Criteria for study eligibility
included that the child: (a) had to be aged 2–5 years,
(b) spoke English as a first language, (c) did not have a
parent reported developmental disability, and (d) had
not participated in an injury related study within the
past 12 months. Parents and children visited a University
clinic and were provided a project description and signed a
consent form (as approved by the Institutional Review
Board). The researcher also collected verbal assent from
each child. A random numbers table was generated by
a computer program to develop four equal groups of 10
(i.e., four distractions) across two conditions (i.e., informed
or misinformed). Half of the sample was randomly assigned
to be told the study was about child supervision and
informed that they would be videotaped while in the room.
The remaining random sample half was informed the study
was about child patience and they were not informed
about being videotaped until after the observation.
The parent and child were then informed that they
were to occupy another clinic room for 45 min. The
parents were read a brief description of what to expect
while in the room, depending on the condition they were
randomly assigned. Specifically, parents in the Phone
Distraction group were told that a cordless phone was
834 Boles and Roberts
D ow
nloaded from https://academ
ic.oup.com /jpepsy/article-abstract/33/8/833/923482 by A
rizona S tate U
niversity W est user on 24 N
ovem ber 2019
placed in the room (out of child reach) and would ring
sometime during the time they are in the room. They were
told to answer the phone to answer a series of questions by
the researcher on child patience (e.g., ‘‘describe your
experiences going to grocery store with your child’’) timed
to last �15 min. Parents in the TV Distraction group were
told an electronic timer would sound after an unspecified
period (i.e., 15 min) of time, which indicated they needed
to turn on the TV and watch a VCR tape of a typical TV
program (i.e., a 15 min segment on meal preparations). In
an effort to increase parent adherence toward watching the
TV program, parents were instructed to pay close attention
to the program in order to answer follow-up questions.
Parents in the Computer Distraction group waited until an
electronic timer alarm sounded after 15 min, which
indicated they needed to turn on the computer that
revealed a program already up on the screen. The computer
program was created by the author using PowerPoint�.
Specifically, an automated presentation was provided to
parents on the subject of child patience. The slide show
was initiated by parents who pushed a marked button. The
slides were timed to advance automatically, creating a
15 min presentation. Parents in the Computer Distraction
group were also told to pay close attention to the computer
program in order to answer follow-up questions by the
researcher. After parents in each distracted condition
completed the distraction, they remained in the room for a
final 15 min. Finally, parents in the condition with no
planned distractions were told to remain in the room for
45 min, in which follow-up questions would be adminis-
tered. At the end, the parents and their children entered
a separate room and completed the structured interview on
room hazards and the demographic questionnaire.
Debriefing
Because this experimental method involved an alteration of
informed consent, a debriefing period occurred directly
after all study protocols had occurred for the entire sample.
The actual purpose of obtaining information on super-
vision behaviors was explained to the uninformed sample
half with clarifications of why it was necessary to refrain
from revealing the study purpose until after the observa-
tions and information were collected. Participants in the
previously uninformed group were also informed that they
were videotaped and provided an explanation that it was
necessary to reliably code each of the parent and child
behaviors using a recording device. They were reminded
of their right to withdraw from the study and remove
their data from the project at this time without penalty
before completing a second informed consent form.
No participants in the uninformed group withdrew from
the study after debriefing. Parents in the informed group
were also debriefed on the study aims immediately after
participation. Parents were given $25 for their time
participating.
Materials
Simulated Home Environment
Observations took place in a university clinic room
designed to replicate a typical home environment. The
room contained furniture (i.e., a couch, end tables, a floor
lamp, plants, padded chairs, a desktop computer,
telephone, TV, and VCR) in addition to objects classified
into two groups: (a) low injury risk (e.g., a pillow) and
(b) high injury risk (e.g., a knife, medicine, a step ladder,
5-gallon bucket, a lighter, and a spray cleaner).1 The
furniture was arranged on the perimeter of the room walls,
with the risky objects were placed in varied locations (e.g.,
a pill container was on an end table, the knife was on a
table, and the lighter was in the seat portion of a padded
chair at the side edge, similar to where a lighter might be if
one had fallen out of a pocket. Objects were classified into
group 1 or 2 based on a sorting task by two experts on
child injury with 100% agreement. The selected risky
objects were based on developmentally related injuries to
this particular age group (National Center for Injury and
Prevention and Control, 2001, 2002).
Measures
Demographic Form
Demographic information was collected from parents via a
self-report questionnaire. Information collected included:
(a) parent education, (b) occupation, (c) housing condi-
tion, (d) child health and past injury history, and
(e) amount of time spent on a computer, watching TV,
and talking on the phone.
Structured Interview on Room Hazards
A brief structured interview on parent recall of conditions
was conducted in order to assess the parent’s level of
awareness of risky and nonrisky items in the simulated
1Each risky object had been modified to be relatively safe and
many had been utilized in our prior studies. The knife was a folding
type with a dulled blade, the medicine was candy placed in a 7-day
medicine organizer without safety locking lids, the bucket did not
contain any water and had a large warning label attached to the
side, the lighter (which did not have a safety feature) was drained of
the fluid and the flint was removed to prevent spark, the spray
cleaner was water mixed with food coloring, and a step stool
utilized one step and multiple safety labels were attached to its
surface.
Supervising Children During Parental Distractions 835
D ow
nloaded from https://academ
ic.oup.com /jpepsy/article-abstract/33/8/833/923482 by A
rizona S tate U
niversity W est user on 24 N
ovem ber 2019
home room, the interaction their child had with such
items, and the potential level of risk for injury for their
child if they interacted with a risky item. Specifically,
parents were asked to consider how likely they thought an
injury could occur when their child interacted with an
identified home hazard by providing a rating from 1 to 5,
where 1 indicated that a parent did not think any injury
would likely happen and 5 meant an injury would
absolutely happen.
Design
The study utilized comparisons of multi-phase conditions,
in which the 45 min period of time in the room is divided
into three phases: (a) no planned distractions, (b) a plan-
ned distraction period, and (c) no planned distractions.
Each phase lasted 15 min that was monitored by the
researcher or parent using digital stopwatches. Ten parent–
child dyads were randomly assigned to one of four
conditions: (a) Phone Distraction, (b) TV Distraction,
(c) Computer Distraction, or (d) No Distraction.
Data Analysis
Observational Data Processing
Observational data were obtained by recording 45 min
segments of parent–child interaction on to VHS tapes,
using a video camera permanently mounted in the upper
corner of the observation room. A VCR was affixed to a
shelf beneath the camera and hidden by a plant. The VHS
tape was later converted to DVD format in order to be
analyzed by observational software. The observational
software used to code and analyze behavior was The
Observer� and The Observer Video Pro� (Noldus, version
5.0.25, The Netherlands).
Coding Scheme
In order to reliably code behavioral data, a behavioral
coding scheme was developed based on the model
proposed by Saluja et al. (2004), although not all
components of the model could be examined given that
the parent and child dyads remained in the room together
and in relatively close proximity (i.e., auditory attention
and far proximity beyond a single room were not coded).
Three parent behaviors (proximity, visual attention, and
engagement) and one child behavior (risky behavior) were
coded as the primary behaviors for analyses using the
following operational definitions. Close Proximity was
coded when the parent was touching or within an arm’s
reach of their child; Close Visual Attention was coded when
a parent positioned her/his head in a direct line of vision or
within peripheral view of their child; Engagement was
coded during activity in which the parent and child are
both simultaneously involved. Involvement may have been
verbal (e.g., communicating with each other), nonverbal
(e.g., actively listening to story telling or to a direction with
parental eye contact), and active (e.g., following a request
to retrieve an item). Examples include: a parent reading a
book to their child who has occasional eye contact with the
book or parent, the parent and child throwing a ball back
and forth, or sitting together talking. Risk Taking behavior
was coded when the child physically touched a predeter-
mined risky object.
Four conceptually similar behaviors were also coded
during each interval. For example, Far Proximity was coded
when a child was beyond the reach of parent; No Visual
Attention was coded when a parent was turned away from a
child; No Engagement was coded during any activity
performed by the parent, in which the child has no
involvement or interaction. This behavior was defined as
when a parent provides no verbal and physical attention
during an activity (e.g., visual focal or peripheral atten-
tion is withheld and no interaction is made with the
child). Examples include the parent watching the TV but
the child is NOT watching the TV and the parent is not
talking to the child (even though the child may still be in
the view of the parent). Finally, No Risk Taking was coded
when no interaction with hazards occurred during the
interval. Preliminary data analyses of the four primary
study variables revealed significant negatively correlated
findings with their opposite behavior (e.g., Close Proximity
and Far Proximity, r ¼ �.81, p ¼ .005). Therefore, only
three supervision behavioral components representing
potentially safer parenting behaviors and one child
behavior were analyzed for this study (e.g., Close
Proximity, Close Visual Attention, Engagement, and
Risk Taking).
Behaviors were coded as events during 15 s intervals.
That is, regardless of duration, a behavior was counted
once and only once if it occurred during each 15 s interval.
The available software enabled coders to rewind, pause,
and fast forward observational data during coding,
maximizing adequate assessment of all behaviors.
Reliability Training
All reported data were coded by the primary coder after
training and an independent coder established interrater
reliability. A Cohen’s Kappa coefficient was calculated in
order to evaluate the level of performance using randomly
selected data. A minimum of .75 kappa values, suggesting
strong agreement above chance, was used to indicate
reliable coding has occurred (Fleiss, 1981). After being
trained to reliability four randomly selected 45-min
observations (representing 10% of the total data) were
836 Boles and Roberts
D ow
nloaded from https://academ
ic.oup.com /jpepsy/article-abstract/33/8/833/923482 by A
rizona S tate U
niversity W est user on 24 N
ovem ber 2019
selected for coding by an individual who had limited
knowledge of research objectives. Twelve Kappa coeffi-
cients (three parenting behaviors across four participants)
for supervising behaviors showed an average value of .88
(SD ¼ 0.09), ranging from .72 to 1.00. A separate analysis
of agreement for Risk Taking behavior for all 12
observations indicated a 100% agreement for all observa-
tions, suggesting satisfactory interrater reliability.
Preliminary Data Screening
Examinations of standardized values for all variables
revealed five values greater than �3.29 (p < .001, two-
tailed test). Affected variables included Risk Taking (one
score from each phase for a total of 3, but not from the
same individual), one score on Engagement during the first
phase, and one score on Close Visual Attention from the
postdistraction phase. These extreme values were across
participants, indicating a univariate approach could be
utilized to reduce the impact of such scores on other data.
Specifically, these raw scores were assigned a value of 1
unit larger or smaller, depending on the direction of the
outlier, than the next most extreme score in the
distribution.
Results
Means and standard deviations for parent supervising
behaviors and child risk taking behaviors across time
periods and between distraction groups are displayed in
Table I. The results for parent and child behaviors are
presented separately in which four one-way repeated-
measures analysis of variance (ANOVA) were conducted
for each observed parent and child behavior (see Fig. 1, in
which the y-axis represents the average number of times
the target behavior occurred during intervals across
participants).
Risky Child Behaviors
Overall, a marginally significant omnibus interaction effect
for child Risk Taking across distraction groups was
observed, F(6, 72) ¼ 1.9, p ¼ .09 (partial eta squared ¼ .14;
see Fig. 1A). As a result, no further follow-up tests were
conducted.
Parent Supervising Behaviors
When examining parent behaviors (shown in Fig. 1B),
Close Visual Attention observations showed a significant
interaction effect by Group, F(6, 72) ¼ 6.43, p < .001,
(partial eta squared ¼ .35). Specifically, during the
Computer Distraction phase, a significant reduction
(p < .001) in Close Visual Attention was revealed when
compared with all other distraction groups. That is,
parents showed a significant reduction in the number of
times they made Close Visual Attention with their child
while using the computer.
A significant interaction effect was also detected
between Close Proximity X Group, F(6, 72) ¼ 2.56,
p < .05 (partial eta squared ¼ .18). Specifically, a significant
quadratic effect (p < .001) can be seen in Fig. 1C, in which
parents in the Phone Distraction group significantly
reduced their proximity to their child during the distraction
phase and then significantly increased in proximity during
the postphase when compared with both distraction and
baseline phases. Additionally, significant differences were
shown between distraction groups during only the post-
phase in which parents in the Computer Distraction group
showed significantly less proximity to their children when
compared with all other groups (p < .05).
Parent and child Engagement also showed a significant
interaction effect, F(6, 72) ¼ 5.80, p < .001 (partial eta
squared ¼ .33). Specifically, with the exception of parents
in the No Distraction Group, all parent and child dyads
significantly reduced their Engagement with each other
during the distraction phase (all p < .05). Additionally, as
shown in Fig. 1D, parents and their children who were in
Table I. Means (SD) for Parent and Child Behaviors During all Phases
and Across Groups
Behavior Baseline Distraction Post
Phone Distraction Group
Parent
Visual attention 59.9 (0.3)a1 59.8 (0.4)a1 59.6 (1.3)a1
Close proximity 41.7 (6.9)a1 26.3 (16.5)a2 48.8 (12.1)a3
Engagement 59.3 (1.1)a1 23.5 (15.9)a2 59.5 (1.3)a1
Child risk taking 3.4 (5.4)a1 2.2 (3.2)a1 3.3 (7.7)a1
TV Distraction Group
Parent
Visual attention 59.9 (0.3)a1 59.5 (1.3)a1 59.6 (0.8)a1
Close proximity 29.7 (13.5)a1 35.9 (20.6)a1 41.7 (15.4)a,b1
Engagement 59.9 (0.3)a1 42.0 (12.6)b2 58.5 (1.4)a1
Child risk taking 3.7 (5.2)a1 7.3 (10.2)a1 1.7 (1.8)a1
Computer Distraction Group
Parent
Visual attention 59.9 (0.3)a1 41.5 (11.5)b2 59.5 (1.0)a1
Close proximity 40.6 (13.8)a1 22.2 (16.1)a1 28.5 (19.6)b1
Engagement 60.0 (0.0)a1 38.2 (14.4)b2 59.1 (1.6)a1
Child risk taking 4.1 (5.4)a1 6.1 (7.4)a1 3.4 (5.7)a1
No Distraction Group
Parent
Visual attention 59.9 (0.3)a1 59.8 (0.4)a1 59.8 (0.4)a1
Close proximity 42.3 (14.1)a1 30.6 (20.0)a1 36.7 (15.7)a,b1
Engagement 59.9 (0.3)a1 59.8 (0.4)c1 60.0 (0.0)a1
Child risk taking 1.2 (1.4)a1 3.9 (9.0)a1 6.8 (9.7)a1
Subscripts that differ by letter are significant across groups (p < .05); different
numerical subscripts significantly differ within groups (p < .05).
Supervising Children During Parental Distractions 837
D ow
nloaded from https://academ
ic.oup.com /jpepsy/article-abstract/33/8/833/923482 by A
rizona S tate U
niversity W est user on 24 N
ovem ber 2019
the Phone Distraction group showed significantly less
Engagement compared with the TV and Computer Distrac-
tion groups who were all significantly lower on engagement
behaviors compared with the No Distraction group.
In order to test the effects of whether knowing the
study was about child supervision, multiple one-way
MANOVAs across all three time periods were conducted
which revealed no group differences between informed and
misinformed parent–child dyads on all parent and child
behaviors (p > .05). Parents who were not aware of being
in a study on child supervision while being videotaped
showed no significant differences in their parental
behaviors related to supervision when compared with
parents informed of the study aims.
Correlation Analyses
Intercorrelations were calculated among Parent age, Child
age, Risk Taking, and three parenting behaviors: Close
Visual Attention, Close Proximity, and Engagement.
During the baseline phase a significant negative relation
between Child Age and Risk Taking occurred in which
risky behavior was less likely to occur as children neared
age 5 (r ¼ �.53, p < .01). Additionally, a significant
positive relation was found between Parent Age and
Close Proximity in which older parents showed greater
Proximity (r ¼ .44, p < .01). Finally, parents who demon-
strated more Close Visual Attention were significantly more
likely to also show Engagement during the baseline phase
(r ¼ .34, p < .05).
During the distraction phase, a significant negative
relation between Child Age and Risk Taking remained from
baseline phase (r ¼ �.49, p < .01). Additionally, a sig-
nificant positive relation between Close Visual Attention
and Close Proximity occurred (r ¼ .39, p < .05) as well as a
significant positive relation between Close Proximity and
Engagement (r ¼ .35, p < .05). During the postphase, a
significant negative relation between Child Age and Risk
Taking remained from the previous phases (r ¼ �.44,
p < .01). Interestingly, no significant correlations were
found among child gender, major injury history, and Risk
Taking. Furthermore, analyses of parent report of room
hazards after the observation revealed no significant rela-
tionships on child or parent behaviors or family demo-
graphic variables.
A B
C
0
10
20
30
40
50
60
Baseline Distraction Post
M e a n
I n
te rv
a l F
re q
u e n
c y
Phone TV Computer No Distraction
Baseline Distraction Post
Phone TV Computer No Distraction
Baseline Distraction Post
Phone TV Computer No Distraction
Baseline Distraction Post
Phone TV Computer No Distraction
0
10
20
30
40
50
60
M e a n
I n
te rv
a l F
re q
u e n
c y
0
10
20
30
40
50
60
M e a n
I n
te rv
a l F
re q
u e n
c y
D
0
10
20
30
40
50
60
M e a n
I n
te rv
a l F
re q
u e n
c y
Figure 1. The average number of times the target behavior occurred during intervals across participants. (A) Mean number of Child Risk Taking behaviors by children during each experiment phase, by distraction group; (B) Mean number of parents’ Close Visual Attention with children during each experiment phase, by distraction group; (C) Mean number of parents’ Close Proximity with children during each experiment phase, by distraction group; and (D) Mean number of parents’ Engagement with children during each experiment phase, by distraction group.
838 Boles and Roberts
D ow
nloaded from https://academ
ic.oup.com /jpepsy/article-abstract/33/8/833/923482 by A
rizona S tate U
niversity W est user on 24 N
ovem ber 2019
Discussion
Our findings are that parents show a shift in parental
supervision during typical home distractions. Compared
with parents who were not experimentally distracted, all
distractions significantly reduced parental engagement and
the parents’ visual attention was reduced while the parents
were engaged with the computer. Additionally, parents
were significantly more likely to be farther away from their
child during TV and phone distractions. These findings
largely support our hypotheses that parents would reduce
their supervision behaviors during common home distrac-
tions. The distractions in the current study were believed
to have generalizibility to the real home world of families,
given the common experiences of parents talking on the
phone, watching TV, or using a computer while super-
vising their child. This study provides a number of
important advancements in explicating the role of parental
supervision of preschool children during typical home
distractions. In addition, this study empirically addresses
methodological concerns often raised during observational
investigations, which can be useful when designing future
supervision investigations.
Although not statistically reliable, children increased
their risk taking behaviors during distractions, except when
parents were on the phone, and showed a steady increase
in risk taking over time when parents were not distracted.
Additionally, there was a negative relation between the
variables measuring child age and risky behavior during all
three time periods. Even though not systematically
evaluated for this study, children who interacted with
risky objects most often attempted to bring the object to
a parent. Such behavior may be the result of prior parental
warnings regarding the dangerousness of similar objects. It
remains unclear, how many children become injured by
handling objects they have been told are dangerous and
attempt to move it to a safer location or to an adult. The
fact that children in the No Distraction group gradually
increased their risky behavior over time may indicate
boredom with the simulated environment and a tendency
to explore items, including risky objects initially perceived
as ‘‘off limits.’’
When combined across distraction groups, child risk
taking behaviors during the baseline (M ¼ 3.1; SD ¼ 4.6),
distraction phase (M ¼ 4.9; SD ¼ 7.9), and postphase
(M ¼ 3.8; SD ¼ 6.9) indicates that children showed an
average risky behavior about once every 5 min in an
unfamiliar environment. Our prior observational investiga-
tions without a parent present showed a lower rate of
contact with dangerous items (1.78 contacts every 15 min;
Boles, Roberts, Brown, & Mayes, 2005), which may reflect
the child’s belief that the environment is safer with a parent
present. Additional investigations may benefit from
including more postobservational interviews with children
to address these cognitions.
Contrary to our hypothesis, parents’ eye contact,
distance away from, and involvement with their children
were not statistically related to child risk taking behaviors.
This finding may be due, in part, to the belief that parents
may not have considered their child’s risky behaviors as
actually being dangerous but rather typical child behavior,
resulting in minimal systematic changes in parent
behaviors. In previous research, for example, parents
have reported that childhood injuries are expected during
childhood and not likely to be preventable by parents
(Morrongiello & Dayler, 1996). Additionally, parents’
report of the room’s risky objects and probability of their
child becoming injured showed no significant association
with actual supervision behaviors. This finding may show
how noninjurious events, despite a danger being present,
reinforces the belief that injury is unlikely and that
parenting behaviors need to be modified for the environ-
mental conditions.
After a distraction, parents made greater eye contact,
were more often close to their child, and were more
engaged with their child. Nonetheless, supervision beha-
viors were significantly intercorrelated during only the first
30 min, suggesting parents may change their supervision
practices over time. Thus, the present study provides only
limited support of the model of supervision behaviors from
Saluja et al. (2004) for understanding risky behaviors. In
particular, parents showed no significant changes in their
supervising behaviors related to their child’s risky behavior.
This finding is particularly striking given the tendency for
children to increase risky behaviors during most parent
distractions as well as across time. Moreover, during
distractions, parents showed significant reductions in their
ability to supervise their children, limiting the ability to
provide education or to take immediate action necessary to
prevent or minimize a possible injury after risky behavior.
This finding also provides evidence that the manipulation
was effective (and simulated real life).
Additional investigations are needed to assess the
impact of environmental modifications on not only risky
child behavior but also parental supervision. That is,
parents may be likely to modify their supervision behaviors
in terms of how close they are, how often they make direct
eye contact, and how often they engage in verbal or
physical interactions based on how they perceive the
environmental risk as well as beliefs about typical child
behavior. Clearly, this is a complex model of reciprocal
Supervising Children During Parental Distractions 839
D ow
nloaded from https://academ
ic.oup.com /jpepsy/article-abstract/33/8/833/923482 by A
rizona S tate U
niversity W est user on 24 N
ovem ber 2019
interactions rather than a simple relationship of super-
vision ! risky behavior ! injuries.
Assessments regarding parents’ perceptions of child
behaviors during observations may also help identify
possible reasons parents’ supervision behaviors were not
related to risky child behaviors. Further, additional
variables such as verbal behavior and parental environ-
mental modifications require analyses to explicate the
relationship among supervising behaviors and risky child
behaviors in the home environment. For instance,
Morrongiello and Dawber (2000) found that when parents
observed children playing on a playground, girls received
more verbalizations of concern about their behavior, while
boys were given encouragement for risky behavior. Our
findings also revealed parents increased their level of
engagement following a distraction which may also be a
factor of general parenting practices. For instance, parents
who show authoritative parenting styles (e.g., high levels of
structure and warmth/involvement) may also have an
association with positive supervision behaviors. Future
supervision investigations may benefit from the inclusion
of parental assessments of parenting behaviors that impact
other child–parent interactions (e.g., compliance with
directions), beyond risk taking behaviors.
The present study provided initial evidence that
participant knowledge in fact does not change participant
behaviors in regard to socially desirable behaviors. That is,
parents who were aware they were being videotaped
showed no greater positive supervision behaviors (e.g.,
using more direct/peripheral eye contact) than parents who
were informed the study was about child patience. This
finding may make future observational studies, which are
potentially an important methodological procedure toward
validating supervision as a construct and subsequent
supervision questionnaires, more likely to be conducted.
Deceptive practices do not appear necessary in order to
minimize socially desirable behaviors.
The limitations of the present study should be
considered when evaluating the present findings. Most
importantly, the observations of parents and children took
place in a simulated home environment. Although precau-
tions were taken to help minimize this potential problem,
the simulated environment might have altered some
behaviors. For example, parents might have thought the
room had been safety proofed before entering with their
child. This potential limitation is countered, however, by
the fact that 100% of parents identified dangerous items in
the room, with an average of 4.75 items listed (ranging
from 3 to 9). Parents in the current study were well
educated and nearly all Caucasian. However, unintentional
injuries remain the leading cause of death in America for
individuals between the ages of 1 and 44 years, across
gender, race, and economic status (Centers for Disease
Control and Prevention, 2005). Developing the present
methodology to obtain these initial findings provided a
significant foundation to build on for future studies to
determine generalizibility to different types of parents. In
particular, there may be differential rates of supervision
behaviors related to diverse backgrounds, including ethnic
diversity and socioeconomic diversity. Understanding these
differences is important to designing tailored interventions
for at-risk populations and their associated risky behaviors.
In addition, our sample size may have limited our
ability to detect some relationships, despite having ade-
quate power for others. Although using a behavioral
observation methodology can be time and labor intensive,
which can limit the number of participants for feasibility
reasons, additional observation-based research may benefit
from using larger samples in order to test multiple num-
bers of hypotheses that deal with relationships that have
varied effect sizes.
As supervision of children continues to be explored as
a necessary construct toward identifying active-based
injury prevention programs, investigations are still
needed to explicate the role of not only nonverbal parent
behaviors, but also the way parents interact with their
environment to make supervising their children produc-
tive, manageable, and within their particular belief systems.
In particular, parents potentially use a combination of
parenting skills during supervision, including modifying
the environment, providing verbal and physical redirec-
tions, and making continuous estimates of risk for their
children for various contexts. However, given the current
number of home injuries still sustained by preschool
children each year, parents are likely not implementing
such skills each day or recognizing the role they have in
reducing the risk of injury for their child. In fact, parents
still often report not being able to prevent injuries and
‘‘accidents’’ are merely the result of bad luck (Morrongiello
& Dayler, 1996; Morrongiello & House, 2004). Therefore,
supervision investigations must also consider how best to
address such erroneous cognitive beliefs about environ-
mental risk as well as limited knowledge on typical child
development that likely impede behavioral interventions
that focus only on changing behaviors. Although it is
encouraging that supervision is increasingly being empiri-
cally investigated as a component of understanding
unintentional injuries, much greater attention is needed
on the development and assessment of comprehensive
models that capture the complex nature of unintentional
840 Boles and Roberts
D ow
nloaded from https://academ
ic.oup.com /jpepsy/article-abstract/33/8/833/923482 by A
rizona S tate U
niversity W est user on 24 N
ovem ber 2019
injuries. Only until such investigations are conducted will
the most effective interventions and prevention programs
be realized.
Acknowledgment
This study was supported in part by the 2004–2005
Student Fellowship Award from The Society for Public
Health Education Fellowship provided by the Centers for
Disease Control and Prevention and was based on R.E.B.
doctoral dissertation submitted to The University of
Kansas.
Conflicts of interest: None declared.
Received August 12, 2007; revisions received February 18,
2008; accepted February 20, 2008
References
Boles, R. E., Roberts, M. C., Brown, K. J., & Mayes, S.
(2005). Children’s risk taking behaviors: The role
of child-based perceptions of vulnerability and
temperament. Journal of Pediatric Psychology, 30,
562–570.
Cataldo, M. F., Finney, J. W., Richman, G. S., & Riley, A.
W. (1992). Behavior of injured and uninjured
children and their parents in a simulated hazardous
setting. Journal of Pediatric Psychology, 17, 73–80.
Centers for Disease Control and Prevention, National
Centers for Injury Prevention and Control. Web-
based Injury Statistics Query and Reporting System
(WISQARS) [online] (2005). Retrieved February 14,
2008, from: www.cdc.gov/ncipc/wisqars
Cody, B. E., Quraishi, A. Y., Dastur, M. C.,
& Mickalide, A. D. (2004). Clear danger: A national
study of childhood drowning and related attitudes and
behaviors. Washington (DC): National SAFE KIDS
Campaign.
Fleiss, J. L. (1981). Statistical methods for rates and
proportions (2nd ed.). New York: Wiley.
Gärling, A., & Gärling, T. (1993). Mothers’ supervision
and perception of young children’s risk of uninten-
tional injury in the home. Journal of Pediatric
Psychology, 18, 105–114.
Guyer, B., Hoyert, D. L., Martin, J. A., Ventura, M. A.,
MacDorman, M. F., & Stobino, D. M. (1999).
Annual summary of vital statistics 1998. Pediatrics,
104, 1229–1246.
Haynes, S. N., & Horn, W. F. (1982). Reactivity in
behavioral observation: A review. Behavioral
Assessment, 4, 369–385.
Jacob, T., Tennenbaum, D., Seilhamer, R. A., Bargiel, K.,
& Sharon, T. (1994). Reactivity effects during
naturalistic observation of distressed and nondis-
tressed families. Journal of Family Psychology, 8,
354–363.
Morrongiello, B. A., Corbett, M., McCourt, M.,
& Johnston, N. (2006). Understanding unintentional
injury-risk in young children I. The nature and scope
of caregiver supervision of children at home.
Journal of Pediatric Psychology, 31, 529–539.
Morrongiello, B. A., & Dawber, T. (1998). Toddlers’ and
mothers’ behaviors in an injury risk situation:
Implications for sex differences in childhood injuries.
Journal of Applied Developmental Psychology, 19,
625–639.
Morrongiello, B. A., & Dawber, T. (2000). Mothers’
responses to sons and daughters engaging in
injury-risk behaviors on a playground: Implications
for sex differences in injury rates. Journal of
Experimental Child Psychology, 76, 89–103.
Morrongiello, B. A., & Dayler, L. (1996). A community-
based study of parents’ knowledge, attitudes and
beliefs related to childhood injuries. Canadian Journal
of Public Health, 87, 383–388.
Morrongiello, B. A., & House, K. (2004). Measuring
parent attributes and supervision behaviors relevant
to child injury risk: Examining the usefulness of
questionnaire measures. Injury Prevention, 10,
114–118.
Morrongiello, B. A., Ondejko, L., & Littlejohn, A. (2004).
Understanding toddlers’ in-home injuries. II.
Examining parental strategies, and their efficacy, for
managing child injury risk. Journal of Pediatric
Psychology, 29, 433–446.
National Center for Injury and Prevention and Control.
(2001). Injury fact book 2001–2002. Atlanta, GA:
Centers for Disease Control and Prevention.
National Center for Injury and Prevention and Control.
(2002). Injury research agenda. Atlanta, GA: Centers
for Disease Control and Prevention.
Saluja, G., Brenner, R., Morrongiello, B. A., Haynie, D.,
Rivera, M., & Cheng, T. L. (2004). The role of
supervision in child injury risk: Definition,
conceptual, and measurement issues. Injury Control
and Safety Promotion, 11, 17–22.
Shannon, A., Brashaw, B., Lewis, J., & Feldman, W.
(1992). Nonfatal childhood injuries: A survey at the
Children’s Hospital of eastern Ontario. Canadian
Medical Association Journal, 146, 361–365.
Supervising Children During Parental Distractions 841
D ow
nloaded from https://academ
ic.oup.com /jpepsy/article-abstract/33/8/833/923482 by A
rizona S tate U
niversity W est user on 24 N
ovem ber 2019