2

Running Head: Leisure and Children’s Well-Being

The Contribution of Active and Passive Leisure to Children’s Well-Being

Mark D. Holder¹, Benjamin Coleman², and Zoë Sehn¹

¹University of British Columbia, Okanagan

3333 University Way

Kelowna B.C., Canada V1V 1V7

² Okanagan College

Kalamalka Campus

7000 College Way (C 311)

Vernon, B.C.,Canada V1B 2N5

Accepted for Publication in Journal of Health Psychology, November 2008

Word count: 4816 (Abstract, text, and References) plus 2 figures

Email:

Phone: 250 807-8728

Fax: 250 807-8001

The relation between leisure and well-being, including happiness and self-concept, was examined in 375 children aged 8-12 years. Active leisure (e.g., physical activity) was positively correlated with well-being. Passive leisure (e.g., television and video games) was negatively correlated with well-being. Aspects of active leisure (e.g., the importance of sport to the child and how sports made the child feel) as judged by both parents and children accounted for unique variance in children’s well-being; passive leisure did not. Similar to previous research on adolescents and adults, active leisure activities were related to children’s well-being.

Key words: happiness; leisure; children; well-being; exercise
Introduction

Leisure involves voluntary non-work activities engaged in for enjoyment (Hills & Argyle, 1998; Melamed, Meir, & Samson, 1995). Leisure has been associated with many health benefits including reduced stress (Iwasaki, 2001), improved cardio-respiratory fitness (Wong et al., 2003), and enhanced cognition (Singh-Manoux et al., 2003).

Leisure enhances overall psychological well-being (see Caldwell, 2005) and this relation may be stronger than the relation between leisure and physical health (Sacker & Cable, 2006). Subjective well-being refers to individuals’ assessments of their lives in terms of life satisfaction and happiness (Holder & Coleman, in press). The relation between leisure and well-being has been demonstrated for a variety of populations and activities. For example, holiday-taking in Britain modestly increased happiness (Gilbert & Abdullah, 2004), leisure in Chinese university students predicted happiness even after personality traits were controlled (Lu & Hu, 2005), leisure was associated with greater happiness in people with disabilities (Dickie et al., 2002; Lancioni et al., 2005), and hobbies were linked to increased happiness in the elderly in rural Japan (Onishi et al., 2006). The relation between leisure and well-being may be causal and enduring; leisure during adolescence predicted well-being 15 years later (Sacker & Cable, 2006).

Theories attribute the link between leisure and well-being to several factors including that leisure buffers the effects of negative events (Tedeschi & Calhoun, 1995; 2004; Tedeschi et al., 1998) and enhances well-being by acting as a protective factor by providing social support, feelings of competency and meaning, relaxation, and distraction (Caldwell, 2005). Although most research and theories are based on adolescents and adults, the findings and theories may apply to children. For example, leisure may contribute to children’s well-being by enhancing social relations (see Caldwell, 2005) which are associated with children’s happiness (Holder & Coleman, in press).

Leisure involves activities freely chosen (Lu & Hu, 2005) based on individual interest (Hills & Argyle, 1998). This free choice and intrinsic motivation can contribute to increased happiness, autonomy, and self-confidence (Csikszentmihalyi & Hunter, 2003; Frederick-Recascino & Schuster-Smith, 2003). According to Self-Determination Theory (Deci & Ryan, 1985; Ryan & Deci, 2000), leisure activities contribute positively to life because they involve autonomous and self-determined behavior. Typically, there is a negative correlation between activities that an individual has not chosen and their self-esteem (Eccles et al., 2003; Frederick-Recascino & Schuster-Smith, 2003; Vandell et al., 2005). Children’s activities are often chosen by their parents. Therefore, leisure may not be positively correlated with children’s well-being.

The relation between leisure and children’s well-being may depend on the type of activity. Passive activities (e.g., television viewing, reading alone, and computer use) have been negatively correlated with well-being (Argyle, 2001; Csikszentmihalyi & Hunter, 2003; Shaw & Gant, 2002) whereas active pursuits (e.g., exercise) are positively associated with higher levels of well-being (Csikszentmihalyi & Hunter, 2003; Hills & Argyle, 1998).

Studies report that physical activity is positively correlated with well-being in children (Parfitt & Eston, 2005) and sedentary behavior is negatively correlated with well-being in adolescents (Ussher et al., 2007). However, these studies primarily focused on negative well-being and did not assess dimensions (e.g., happiness) of positive well-being. For example, Parfitt and Eston (2005) assessed anxiety and depression. Though they also assessed self-worth/esteem, they did so by including only six of 36 items from The Child and Youth Physical Self-perception Profile and some of these items assessed negative self-esteem. Similarly, Ussher et al (2007) relied on the Strengths and Difficulties Questionnaire which assesses mental disorders including emotional symptoms, conduct problems, hyperactivity/inattention, and peer relationship problems. Though traditionally science has focused on pathologies (see Joseph et al., 2004), positive strengths (including happiness) are related to health (e.g., Mahon, Yarcheski, & Yarcheski, 2005). Evaluating positive well-being is critical because positive and negative psychological states may be independent, and positive states may exert more influence on health (Lai et al., 2005).

The present study investigated the relation between active and passive leisure and positive well-being in children. Both happiness and positive self-concept were assessed and both self-report and non self-report measures were used.

Methods

Participants

Letters of information and informed consent, and questionnaires, were distributed to 1,633 children in Grades 4-6 to deliver to their parents. Although 531 parents consented, only 514 (31.5%) students participated (e.g., some students were absent during testing). The children (49% boys & 51% girls), aged 8-12 years, were from 13 private and public schools in both rural and urban areas in Western Canada.

Materials

Questionnaires.

When appropriate, Likert-type scales were used because children may not comprehend visual analogue scales (Shields et al., 2003). Because, children prefer filling in circles and having more response options (Rebok et al., 2001), the scales used circles and multiple response options.

Piers-Harris 2. Children completed the Piers-Harris Children’s Self Concept Scale 2 (Piers-Harris 2), a standardized self-report questionnaire that assesses self-concept (Piers & Herzberg, 2002). It is a modification of the 1984 Piers-Harris Children’s Concept Scale which provides a multidimensional widely-adopted assessment with high test-retest reliability and internal consistency, is the most frequently used and recommended instrument for children, and can be administered to groups (Marsh & Holmes, 1990; Piers & Harris, 1984; Piers & Herzberg, 2002). The Piers-Harris 2 has 60 True-False items that express how children may feel about themselves. These items form an overall score of positive self concept (Self-Concept), and standardized scores from the Physical Appearance and Attributes (Physical-Appearance), and Happiness and Satisfaction (Happiness/Satisfaction) subscales. Similar to other studies (e.g., Holder & Coleman, 2008; Young & Bradley, 1998), Happiness/Satisfaction was used as one estimate of children’s happiness. Physical-Appearance and Self-Concept were used as estimates of self-concept.

Faces Scale. Children completed the Faces Scale which assessed their happiness. It uses a Likert-type scale with seven drawings of faces ranging from very sad to very happy. Children rated how happy they were “most of the time”. The Faces Scale is appropriate because children perform most accurately when emotions are represented as drawings, not photographs, and they are best at labeling happiness, followed by sadness (MacDonald, Kirkpatrick, & Sullivan, 1996). This measure has been used successfully to assess children’s happiness (Holder & Coleman, 2008, in press).

Children’s Questionnaire. The Children’s Questionnaire included six questions related to active leisure: “Last week, how many hours did you do athletic activities?”; “How good are you at sports?”; “How important are athletic activities to you?”; “How involved are your parents in your athletic activities?”; “How important are your athletic activities to your parents?”; “Overall, how does your involvement in athletic activities make you feel?”. In Canada, “athletics” refers broadly to exercise and sports and not just to track and field. Four questions related to passive leisure: “How many hours last week did you spend watching television?”; “How many hours did you spend on a computer?”; “How many hours last week did you spend playing computer and video games?”; and “How many hours last week did you talk on the phone?”. Children filled in a circle representing one of seven response options for each item.

Parents’ Questionnaire. The Parents’ Questionnaire included items similar to those on the Children’s Questionnaire. Additionally, parents rated their child’s happiness using the Faces Scale (i.e., “Overall, how does your child feel most of the time?”). Ratings by others of happiness and life satisfaction are valid and show good agreement with self-reports (Lepper, 1998) including those of children (Holder & Coleman, 2008, in press).

Procedure

Following approval from UBC’s ethics board, informed consent was obtained from the school boards, teachers, and parents. Parents completed the Parent’s Questionnaire without consulting their child. If their parents consented, children were asked for their assent and assessed 10 days after their parents.

Researchers explained the study to the children and written standardized instructions asked the children to read the questions carefully, think about their answers, examine all response options, and choose the best option for them. Researchers answered children’s questions. Teachers did not assist. Participation took approximately 30 minutes.

Data Analyses

Five measures of well-being were used. Three measures came from the Piers-Harris 2: Happiness/Satisfaction, Physical-Appearance, and Self-Concept. The remaining two measures used the Faces Scale: children’s ratings of their happiness (ChildOwnFace), and parents’ ratings of their children’s happiness (ParentChildFace). Participants who failed to report a score on at least one measure were not included leaving 375 children. The distribution of scores for Happiness/Satisfaction (Skewness = -2.12, SE = .12 & Kurtosis = 5.49, SE = .23), and ParentChildFace (Skewness = -0.65, SE = .12 & Kurtosis = 1.00, SE = .23), violated the assumption of normality. Therefore, these distributions were transformed by reflecting the scores (see Tabachnick & Fidell, 2001), and natural logarithms were computed which improved the Happiness/Satisfaction (Skewness = .82, SE = .12 & Kurtosis = -.22, SE = .23), and ParentChildFace (Skewness = -.51, SE = .12 & Kurtosis = .08, SE = .23) distributions.

Multivariate regression used items from the Children’s and Parent’s Questionnaires to predict variance on three measures: ChildOwnFace, ParentChildFace, and Self-Concept. Significant multivariate predictors were then used in standard multiple regression analyses.

Results

Ratings of Happiness and Self-concept

Over 93% of parents’ and children’s responses on the Faces Scales were in the three happiest categories. Children’s responses on the Happiness/Satisfaction, Physical-Appearance, and Self-Concept scales were within 8%, 8%, and 12% of the standardized norms for the Piers-Harris 2 (Piers & Herzberg, 2002), respectively.

Ratings from all five measures of well-being were positively correlated, .15 r .76, p < 0.01., with some indications of multicolinearity, but not singularity (e.g., r > .90, Tabachnick & Fidell, 2001). Multicolinearity (~ r > .7, Tabachnick & Fidell, 2001) existed between Self-Concept and Happiness/Satisfaction (r = .66) and Physical-Appearance (r = .76). Therefore, only Self-Concept was chosen as a criterion variable for the regression analyses. All measures were included in the correlation analyses.

Children’s Report on Leisure Activities

All active and passive leisure predictors were correlated with ChildOwnFace, except for “Hours spent on the phone” (see Table 1). All correlations between active leisure and well-being were positive whereas all correlations between passive activities and well-being were negative. Similarly, all leisure predictors, other than talking on the phone and hours spent in sports, significantly correlated with ParentChildFace. None of the passive activities significantly correlated with Happiness/Satisfaction, Physical-Appearance, or Self-Concept, whereas all active leisure measures correlated with these scales.

Insert Table 1

Regression of Children’s Leisure. The inter-item correlations between all 10 predictors ranged between .003 and .640. The largest correlation was between “How important are athletics to you?” and “Overall, how does your involvement in athletic activities make you feel?” and approached multicolinearity levels (Tabachnick & Fidell, 2001). Therefore, “How important are athletics to you?” was considered redundant and not included in the following regression analyses. ChildOwnFace, ParentChildFace, and Self-Concept were used as the dependent measures in the following analyses (Happiness/Satisfaction and Physical-Appearance were subsets of Self-Concept and therefore were not included).

The multivariate results, semipartials (sr²) for each predictor, R², and the standardized regression coefficients (β) for the three measures are shown in Table 2. The multivariate regression test (Wilk’s lambda) indicated three significant predictors related to happiness and positive self-concept, Fs(3,355) > 2.70, ps < .05. The remaining six predictors were not significant, Fs(3,355) < 2.60, ps > .05, and were not included in the following multiple regression analysis.

All three predictors accounted for approximately 18% of variance in Self-Concept. Two significant predictors emerged showing that active leisure is associated with children’s self-concept: “Your sport ability” (r2 = .03) and “Sports make you feel” (r2 = .04). The three predictors also accounted for 16% of the variance in ChildOwnFace, with one predictor (“Sports make you feel” r2 = .06) accounting for a significant proportion of the variance. These three predictors accounted for about 7% of the variance in ParentChildFace; children who reported that their involvement in sports made them feel good (r2 = .02), and that their involvement in sports was important to their parents (r2 = .08), also received higher happiness ratings from their parents.

Insert Table 2

Parent’s Report on their Child’s Leisure Activities

All leisure items were significantly correlated with ChildOwnFace, except for “Parental involvement in sport”, “Hours child watched TV”, and “Hours child was on phone” (see Table 1). The same items were significantly correlated with ParentChildFace and Physical Appearance, with the addition of “Parental involvement in sport” as a significant item. All items where significantly correlated with Happiness/Satisfaction, except for “Hours child watched TV”, “Hours child played video games”, and “Hours child was on phone”. All of the athletic activities, and one passive activity (“Hours child played video games”), were significantly correlated with Self-Concept. Similar to the children’s responses, parents’ responses indicated that active leisure was positively correlated with the children’s well-being and passive leisure was negatively correlated or only weakly correlated with well-being.