Associations of Parental Influences with Physical Activity and Screen Time among Young Children: A Systematic Review

Parents play a critical role in developing and shaping their children's physical activity (PA) and sedentary behaviours, particularly in the early years of life. The aim of this systematic review is to identify current literature investigating associations of parental influences with both PA and screen time in young children. This systematic review was conducted in November 2013 using 6 electronic databases covering research literature from January 1998 to November 2013. Thirty articles that met inclusion criteria were identified. These studies covered five important aspects of parenting: (1) parenting practices; (2) parents' role modelling; (3) parental perceptions of children's PA and screen viewing behaviours; (4) parental self-efficacy; and (5) general parenting style. Findings suggest that parents' encouragement and support can increase children's PA, and reducing parents' own screen time can lead to decreased child screen time. Improving parenting practices, parental self-efficacy or changing parenting style may also be promising approaches to increasing PA time and decreasing screen time of young children.


Introduction
Physical activity (PA) and sedentary behaviours (predominantly screen time) impact on the weight status of children [1][2][3]. Research evidence suggests that helping young children establish an active lifestyle can prevent them from overweight and obesity [4][5][6][7]. In the early years of life, parents play a critical role in developing and shaping their children's PA and sedentary behaviours through role modelling and creating a healthy home environment that increases PA and reduces screen time [8].
There have been several systematic reviews investigating correlates of PA and sedentary behaviours in young children [9][10][11][12][13]. These reviews used a social-ecologic framework [14] to summarize the correlates of PA and screen time of young children in five domains: (1) demographic and biological; (2) psychological, cognitive, and emotional; (3) behavioural attributes and skills; (4) social and cultural; and (5) physical environmental. Although parents' role modelling, parenting practices and parental perceptions of children's PA, and sedentary behaviours are part of the social and cultural domain, they have not been explicitly discussed in these reviews. General parenting style and parental self-efficacy that are also part of the social and cultural domain have not been investigated in these reviews.
In addition to parents' role modelling and parenting practices, general parenting style and parental self-efficacy also influence the development of PA and sedentary behaviours of young children, particularly in the early years of life [8]. Parenting style refers to a general pattern of parenting that provides the emotional background in which parents' behaviours are expressed and interpreted by a child [8]. It can be conceptualized as a context that moderates the influence of specific parenting practices on a child. Closely related to parenting practices and parenting style, parental self-efficacy is regarded as a parent's belief that he or she is capable of organizing and executing tasks related to parenting a child [15].
Findings from a recent systematic review [16] investigating associations between parental factors (parents' role 2 Journal of Obesity modelling, parenting practices, and parental perceptions of importance and value of PA) and young children's PA were inconsistent, reporting that parental and family dynamics associated with children's PA are undeveloped. In particular, the review did not examine the parental influences on children's screen time (the time spent on watching TV, DVDs, or videos, using a computer and playing with an electronic game system) despite increasing research interest in investigating children's screen time and its independent association with childhood obesity. Further investigation was called for to clarify and understand specific parental influences that are associated with PA in children using comprehensive reviews of well-defined parental influences and their effects on both PA and screen time.
To fill in the knowledge gaps in this area, we aimed to update the current literature investigating parental influences and their associations with both PA and screen time in young children.

Search Strategy.
In November 2013, a systematic search of the literature was conducted. Literature included in this review was retrieved from six electronic databases, including Medline, the Cochrane Central Register of Controlled Trials (CENTRAL), Cochrane Database for Systematic Review (CDSR), PsycINFO, EMBASE, and Web of knowledge. Research papers were limited to those written in English and published or included in databases from 1998 to November 2013. There was no restriction of study designs. Study participants in searched papers were limited to parents, father, and/or mother with young children aged ≤6 years. Parental influences included (1) parenting practices, (2) parents' role modelling (parents' own PA and screen time), (3) parental perceptions of children's PA and screen viewing behaviours, (4) parental self-efficacy, and (5) general parenting style. Children's PA included parent-reported "outdoor play" or "active play" and objectively measured PA level (e.g., accelerometer). The search strategy used for the Medline database is displayed in Table 1. A similar search strategy was used for other databases. Additional manual searches of the references of selected articles were also conducted for other relevant articles. Grey literature, such as unpublished studies and dissertations, was also included.

Study Selection.
Study selection was based on predefined inclusion and exclusion criteria. The inclusion criteria were (1) individual quantitative studies that examined relationships of parental influences (covering at least one of the five aspects) with PA or screen time of young children; (2) studies with children aged ≤6 years old, or studies with a wide age range but describing the results specifically for children aged ≤6 years old; and (3) full text articles or dissertations, written in English. Studies were excluded from this review, if they were (1) pilot studies, (2) validation studies, (3) qualitative studies, (4) review papers, (5) studies that examined correlates of children's PA or screen time but did   (5) and (21) and (47) (553) (49) limit (48) to (English language and humans and yr = "1998-current") (483) * The asterisk sign stands for any character(s). not include any aspects of defined parental influences for this review, and (6) studies involving children aged >6 years.
A total of 1414 articles were identified through database searching. Duplicate articles ( = 307) were removed, resulting in 1107 individual articles for consideration. By screening the titles, 1062 articles were considered to be irrelevant and thus excluded. Forty-five papers including grey literature remained as a result of the initial search. The references of these remaining 45 articles were further screened manually to identify other relevant articles. Five additional articles were included. A total of 50 full texts were further assessed. After excluding 20 articles according to the criteria, 30 articles were included in the present review. The process of study selection is reported in Figure 1.

Assessment of Included Articles.
Two reviewers (Huilan Xu and Li Ming Wen) independently screened the study titles and abstracts and then critically appraised the selected articles. Due to heterogeneity of these studies (i.e., differences in study design, study quality, and statistical analysis method), it was not possible to conduct a meta-analysis that uses statistical methods to summarize the results. Therefore, the results from this review are presented descriptively. We critically evaluated the papers using a previously established quality checklist [17] with some modifications. This quality assessment tool was originally adapted from the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement [18]. For this review, the quality checklist consists of eight query items as follows: (1) was the study longitudinal or randomised controlled trail (RCT)? (2) Did the study describe the participant eligibility criteria? (3) Were the study participants randomly selected (or representative of the study population)? (4) Did the study report the sources and details of assessment for parental influences and did the instruments have acceptable reliability? (5) Did the study report the sources and details of assessments for PA and screen time and did all the methods have acceptable reliability for specific age group? (6) Did the study report a power calculation and was the study adequately powered to detect hypothesized relationships? (7) Did the study report the numbers of individuals who completed each of the different measures and did participants complete at least 80% of measures? (8) Did statistical analysis take into account the confounding? A score of "1" was assigned to "yes" to the query  item, or a score of "0" was assigned. The range of score was from "0" to "8" for each paper. A paper with a score above 5 was regarded as a good quality paper.
The quality scores of the papers ranged from 2 to 7 with an average score of 4.9. Most studies ( = 24) used cross-sectional design, with only 6 longitudinal studies having follow-up duration from 1 to 5 years. Sample sizes of these studies were reasonable based on their effect size and significance level except two studies had less than 100 participants.
(3) The amount of time parents report engaging in PA together with their child decreased significantly over time. 4 Oliver et al. (2) Parental enjoyment of PA and perceived importance of child PA were not significantly associated with parental support for PA.  (2) Parental PA and family support for PA were positively associated with children's PA (AOR 1.62 and 2.18, respectively, < 0.10).

Home-based children's PA
(3) Importance of child's PA ability was not associated with children's PA.     (1) Family support was directly and significantly associated with child MVPA ( = 0.28).
(3) Parenting style was not associated with child active play.
(5) Parenting style did not moderate the relationship between parental support and child active play behaviour. 6 Taylor et al. Cross-sectional analysis: (1) Parental activity was weakly associated with child's PA measured by AAC at   (2) Coviewing TV with child was positively associated with more child TV time ( = 0.16, < 0.05) (3) Parent TV time restriction was associated with lower child TV time.

12
Journal of Obesity  (4) Parent believes the amount of TV child watches would not affect child health was not significantly associated with girls meeting Australian/Canadian screen time recommendations.
(5) Parent limits child TV viewing was not significantly associated with girls meeting Australian/Canadian screen time recommendations. 6 14 Journal of Obesity    (2) Parent monitoring activity was not associated with child PA time and screen time.  (2) Parental TV program rules did not predict less TV viewing by children (children whose parents have program rules watched 32.5 more minutes/day TV, < 0.001). The relationship between program rules and child TV viewing was mediated by parental presence during viewing.
(3) Children of parents with program rules tended to spend more time playing outdoors (mean difference of minutes playing outdoors = 17.7, < 0.1).
(4) Parental negative attitude towards TV was directly associated with child TV time ( = −0.09, <

0.05) and
was mediated by parental TV time rules.
(6) Children of parents with time rules and program rules tended to watch TV more frequently (mean difference of frequency watching TV were 0.28, < 0.0001, and 1.07, < 0.001, respectively). 4 No (1) Parental participation in PA (mother's frequency of walking >30 mins/day and frequency of organised sport) was associated with greater outdoor play in preschool children ( = 0.04-0.008).
(2) Sedentary activity rules (frequency that TV is on, rules about TV viewing, and parents limit exposure to TV advertising) were associated with greater outdoor play in preschool children ( < 0.001).
(3) Fewer rules about TV watching were associated with increased sedentary indoor behaviour 3 AOR: adjusted odds ratio.

18
Journal of Obesity or let their child know that PA is good for health); (2) setting rules (e.g., restricting indoor games and having TV viewing rules); (3) parents preference (e.g., preferring child to do the same activities with their older siblings). For children's screen time, parenting practices included (1) setting TV rules (e.g., time or program rules); (2) watching TV with their child; (3) monitoring screen time; (4) having meals without watching TV; and (5) stimulating children to be active.

Parenting Self-Efficacy.
Parental self-efficacy for limiting screen time was assessed by a single item: the level of confidence that parent could say "No" to their child's request for screen time (TV/computer/video games) [36][37][38].
One study assessed parental self-efficacy for influencing their child's PA in eight challenging situations such that the parent cannot think of activities to suggest; the parent is not able to participate in the activity [47].

Parenting Style.
Only one study investigated the association of general parenting style and children's PA [31]. Parenting style was classified by using Maccoby and Martin's [49] classifications of parenting style with two dimensionsresponsiveness (parental warmth/hostility) and demandingness (parental control).

Classification and Measurement of Children's PA and
Screen Time. Children's PA was measured by parent selfadministered questionnaire in 10 studies [19,23,24,26,27,31,[45][46][47][48], followed by accelerometer in 7 studies [20-22, 28-30, 32], or heart rate monitoring and direct observation in one study [25]. Screen time was assessed by totaling the time spent on TV, DVDs, electronic game, and computer ( = 16 studies), which was always measured by parent self-administered questionnaire. Table 5 summarized the associations between parental influences and children's PA. Through examining three aspects of parenting practices on children's PA, there was moderate to strong evidence of linkage between parental encouragement/support and children's PA. For example, of 11 studies examining the relationship between parents' encouragement/support and children's PA, eight studies [21, 23, 26-29, 31, 45] with a mean quality score of 5.6 found that children whose parents encouraged or supported them to do PA were more likely to have higher levels of PA, yet three studies (with a mean quality score of 4.7) did not find such association [20,22,25]. The associations of setting rules and parental preference with children's PA were weak due to a small number of studies and inconsistent findings. One study (a quality score of 7) examining both setting PA rules and parent preference found that parental rules were positively associated with boy's PA, but parent preference was negatively associated with boy's PA [29]. Among three studies examining the association between setting TV viewing rules and children's PA, two studies (a mean quality score of 3.5) found that children with TV viewing rules spent more time playing outdoors [46,48], which was not supported by Gubbels et al. 's study [45]. There was also moderate to strong evidence of positive association between parental PA level and young children's level of activity. Eight studies with a mean quality score of 5.1 found that parents' own PA level was positively associated with young children's PA [19, 20, 23, 28-30, 32, 48] with only two studies revealing no such association [22,24].

Parental Influences and Young Children's PA.
Weak evidence was found in examining associations of parental perception, parental self-efficacy, and parenting style with children's PA due to limited studies and mixed findings. For example, contradictive findings were reported on parental perception of the importance of PA and children's PA in two studies [23,28]. But another two studies found that parent perception of children's physical competence was positively associated with children's PA (MVPA) [21,22]. Only one study found that parents who have a high sense of self-efficacy are more likely to have their children meeting the PA guidelines [47]. General parenting style was not found to be associated with child PA. However, parental warmth, one of the parenting style dimensions, was positively associated with child PA [31]. Table 6 summarized the associations between parental influences and children's screen time. Weak and mixed evidence was found in examining association between parenting practices and screen time. Of 7 studies examining TV time rules, four studies (a mean quality score of 5) found that setting TV time rules resulted in less screen time [33][34][35]46] and two studies (a mean quality score of 4.5) found no such association [38,43] with another study suggesting a negative effect of TV time rules on screen time [45]. Mixed evidence was found from two studies in examining setting TV program rules and screen time [43,46]. The study conducted in 2005 found children whose parents had TV program rules watched more TV [46], but the finding was not supported by another study in 2007 [43]. In addition, one study revealed that more family rules about TV viewing were associated with less screen time [48]. Coviewing TV with a child [35,46] and having meals when TV is on [34,43] were associated with increased screen time. In addition, one study [45] found that monitoring child screen time was not associated with screen time, but stimulating a child to be active was associated to less screen time.

Parental Influences and Young Children's Screen Time.
In contrast to parenting practices, the review found that there was moderate evidence suggesting parental self-efficacy and parents' own TV time were associated with children's screen time. Nine studies with a mean quality score of 4.8 consistently revealed that parents' own TV time was positively associated with their child's screen time [33-35, 37-40, 42, 44]. The evidence was also consistent in four studies (a mean quality score of 5.5), which concluded that high parental self-efficacy in reducing children's screen time was associated with less screen time in children [36][37][38]47].
In terms of associations of parental perceptions and parenting style with screen time, the evidence was weak and inconsistent from only four studies reviewed. For example, two studies revealed that parental perception of their child spending too much time on playing video games or watching TV was associated with increased screen time [33,38]. Parental perception of "TV helps" was found to be associated with increased screen time [37,43]. Parental perception of "TV hurts" was associated with decreased screen time [46]. But this association was not supported by another study conducted by the same author [43].
The authoritative parenting style was found to be associated with decreased children's screen time by only one study [41].

Discussion
By critically assessing and synthesizing evidence from individual studies, the present systematic review updates the current literature and fills knowledge gaps in relation to associations of parental influences with young children's PA and screen viewing behaviours. Moderate to strong evidence was found in relation to the associations of parental encouragement/support for PA and parents' own PA level with children's PA. Moderate evidence was also found regarding associations of parental self-efficacy and parents' own screen time with children's screen time. Associations of other aspects of parenting practices, parental perceptions, and parenting style with children's PA and screen time were indeterminate due to limited studies and contradictory results from studies included in this review.  [10,12]. Another systematic review that included several qualitative studies suggested that parent involvement, encouragement, and modelling of PA may be important influences impacting young children's PA and are worthy of further systematic study [50].

What Is Already
Of four previous systematic reviews of association between parents' TV time and children's screen time, two concluded a positive association [9,11], while the other two found that there was a positive but indeterminate association [10,13]. Mixed findings about the association between parental TV rules and children's screen time were reported by three reviews [9,10,13]. Cillero and Jago and Hinkley et al. found that children with TV rules had less screen time [9,13], while de Craemer et al. found that this association was indeterminate [10].
Mixed evidence in relation to associations of parental influences with children's PA and screen time was further explored by some qualitative studies. For example, Dwyer et al. found that that parental modelling and/or encouragement of activity was a key influence and predictor of PA and sedentary behaviour in children, which was especially acknowledged by parents [51]. Another qualitative study found parents' own screen viewing habits was one of the two most frequently mentioned factors that influence children's screen time (the other one was weather conditions) [52]. Parents also liked the idea of implementing parental rules for TV viewing (e.g., time rules, no TV viewing during meals) [52].

What This Study
Adds. The present review examines the evidence of associations of parental influences with both PA and screen time of young children. Findings from the review generally support and reinforce the evidence found in some of the previous reviews. This review also examined the role of parental rules for PA and screen viewing, parent perceptions of children's PA, and screen viewing and found their effects on children's PA and screen time remained indeterminate because of contradictive findings or a small number of studies.
Unlike previous reviews that only focused on parenting practices, the present review also explores parental selfefficacy and general parenting style as part of the parental influences. Moderate evidence was found that increased parental self-efficacy was associated with reduced children's screen time. But, the effect of parental self-efficacy on increasing children's PA cannot be concluded with only one study found in this review. The association of parenting style with children's PA and screen time remains unclear due to the limited number of studies.

Evidence Gaps and Future
Research. Despite substantial evidence suggesting that an authoritative parenting style was associated with older children's PA and sedentary behaviours [53,54], the present systematic review was not able to make such conclusion for young children with only one study included in the review [31,41]. Therefore, the associations of parenting style and young children's PA and sedentary behaviours need more attention in future research. In this review, only one longitudinal study found that the relationship between parent's PA and preschool children's PA was stronger than that in older children (at follow-up) [32]. It seems likely that the associations of various parental influences and children's PA and screen viewing behaviours would change with advancing age of children. Hence, further longitudinal research on parental influences is needed. In addition, there is a clear evidence gap regarding the effect of parental self-efficacy on children's PA.
As discussed above, associations of some parenting practices (e.g., setting PA and TV rules) and parental perceptions with children's PA and screen time are still indeterminate and further investigation is needed to inform the development of health promotion programs. Inconsistent findings of the association between setting TV rules and children's screen time could be a result of different levels of obedience of rules. One qualitative study involving six European countries revealed that, in general, parents of preschoolschool children only had informal rules about TV viewing [52]. In this review, two individual studies were conducted by Vandewater et al. in United States in 2005 and2007 [43, 46]. The first study found that parental perceptions of "TV hurts" and setting TV time rules were associated with decreased screen time [46]. However, such associations were not found in a later study [43]. It may be that the association between parental perception of "TV hurts" and children's screen time was related to setting TV time rules. Thus, more research is needed to investigate whether the parental perception of TV "hurts" influences establishing and enforcing TV rules. In addition, whether parenting practices are moderated by general parenting style also needs to be investigated.

Conclusions
In the early years of life, some parental influences were significantly associated with young children's PA and screen time with moderate to strong evidence. Results from the present review suggest that parents' encouragement and support can increase their children's PA and reducing parents own screen time can lead to decreased child screen time. Improving parenting practices, parental self-efficacy and parenting style may also be promising approaches to increasing PA time and decreasing screen time of young children.

Strengths and Limitations of This Review
One of the strengths of this review is that we used comprehensive and systematic inclusion and exclusion criteria and a modified quality assessment tool for the critical appraisals of papers. With this quality assessment tool, the strength of evidence from individual study could be assessed based on a quality assessment score. In the present review, the concept of parental influences included not only parents' role modelling and parenting practices that were commonly included in previous reviews, but also general parenting style, parental self-efficacy, and parental perception of PA and screen time, which are important aspects of parenting. In addition, this review searched six databases which enable a wider range of studies to be found.
However, due to the heterogeneity of studies included in the review we were not able to conduct a meta-analysis. Evidence derived from this review was limited by various measurements used regarding parents' and children's PA as well as aspects of parental influences. The definition of children's screen time was inconsistent in the papers reviewed. Study sample sizes in these papers reviewed noticeably varied from 69 to 10,694 where small sample size undermines the study findings. In addition, most studies were crosssectional which means it is difficult to make causal inference. Therefore, consideration must be given when interpreting the results.