Children, Youth and Environments.
Vol 14, No.1 (2004)
ISSN 1546-2250

Perceived Restorative Components: A Scale for Children

 

Kathleen L. Bagot
Department of Psychology Monash University, Australia Department of Psychology Monash University, Australia

Citation: Bagot, Kathleen L. (2004). "Perceived Restorative Components; A Scale for Children." Children, Youth and Environments 14(1): 120-140.

Comment on This Article

This study reports on the development and psychometric validation of a perceived restorative components scale for children (PRCS-C). Children (n=112 boys, n=113 girls) aged 8 to 11 years completed an initial pool of 23 items addressing the components of a restorative environment to assess two familiar, everyday environments- their school playground and their school library. Factor analysis indicated a five-factor model (Being Away– Physical, Being Away- Psychological, Fascination, Compatibility and Extent) of 15 items best fit the data, consistent with prior adult restorativeness measure research and fitting within Kaplan’s Attention Restoration Theory. Satisfactory internal consistency was found for four of the five factors, with a two-item construct of Extent being lowest. School playgrounds had significantly higher restoration potential than school libraries, when compared with school classrooms, indicating divergent validity of the measure. Results were examined by sex and age and differences reported as a broad indicator of the measure’s ability to differentiate between groups of people’s reports of perceived restorativeness and possible developmental differences.

Keywords: children, restorative environments, attention restoration theory, playground, school library

 

Research into environments that are restorative to children’s attention has been hindered by the absence of a measure children can complete to report the perceived restoration potential of their environments. Such a measure would provide researchers with a tool that would improve understanding of restorative components and environments in children’s lives and allow the evaluation and comparison of environments frequented by children from their perspective. In turn, children’s environments could be enhanced or indeed designed to include elements that support and facilitate the restoration of their attention, providing a “user-friendly,” effective, yet economical resource to support children’s development and overall life effectiveness.

Two of the key journals for person-environment researchers have dedicated special issues to environments that facilitate the restoration of attentional capacity (Journal of Environmental Psychology 23(2), 2003 and Environment and Behavior 33(4), 2001). Noticeably absent from both issues, however, was work involving children. Although an argument exists for initially establishing outcomes with one sample only (i.e., adults), this approach results in research with children being delayed, or even disregarded. The aim of this paper is to make a step towards addressing such a delay.

Although research with adults and restorative environments has been dominant (e.g., Hartig, Mang, and Evans 1991; Herzog and Gale 1996; Kaplan 2001a; Korpela, Hartig, Kaiser, and Fuhrer 2001; Kuo and Sullivan 2001), a number of studies have examined the relationship between different types of environments, typically varying in levels of natural elements (e.g., trees and grassy spaces), and children’s attentional capacity (Faber Taylor, Kuo, and Sullivan 2001; Grahn, Martensson, Lindblad, Nilsson, and Elkman 1997; Wells 2000). Some of these studies have drawn on Attention Restoration Theory (ART) (Kaplan 1995; Kaplan and Kaplan 1989) to explain their findings. They reported restoration of attention; however, they did not include an assessment of the level of restorative components within the environments.

Establishing the relationships between the perceived restorative potential in environments, the experience of attention restoration and environmental elements including levels of nature, requires a valid measure of the restorative components for children. To date, this has not been considered in the literature and is the goal of this study. This paper outlines ART, reviews each of the proposed components of restorative environments, touches on measures developed for adults and then reports the development and evaluation of a perceived restorative components scale for children.

Attention Restoration Theory

The capacity to direct attention (CDA) is a crucial component in the performance of specific everyday tasks as well as in overall life effectiveness for both adults and children (Driver 2001; Kuo 2001; Kuo and Sullivan 2001; Piontkowski and Calfee 1979; Rockstroh and Schweizer 2001; Ruff and Rothbart 1996; Wells and Matthews 1994). For example, children with severe attentional deficits (e.g., Attention Deficit Disorder, or ADD) as well as those with less attentional capacity loss demonstrate reduced effectiveness in both academic performance and interpersonal relationships. For all individuals, CDA inhibits distracting, irrelevant stimuli while allowing appropriate information to be extracted from the environment. As CDA requires effort, however, the ability to continue directing attention can be diminished. This reduces the efficacy of the inhibitory mechanism for irrelevant stimuli, in turn resulting in reduced effectiveness. For example, attentionally fatigued individuals (both adults and children) may experience irritability, difficulty with concentrating, problem solving or planning (Kaplan 1995). The restoration of attentional capacity is therefore crucial for both adults and children alike.

Drawing on James' (1950) distinction between “voluntary attention” (i.e., requires effort) and “involuntary attention” (i.e., effortless), Kaplan’s (1995) ART is based on direct and indirect attention respectively (Kaplan and Kaplan 1989; Kaplan 1995). Direct attention, like voluntary attention, requires effort to maintain. As such, the ability to maintain the effort of directing attention can be weakened, which is experienced as directed attention fatigue (DAF). Conversely, indirect attention is effortless. As the two types of attention are experienced one at one time, ART proposes that involuntary or indirect attention leads to the restoration of directed attention. Different environments vary in their support of the experience of involuntary attention; that is, some environments are experienced as more restorative than others (Kaplan 1995).

ART proposes there are four components required for an environment to be restorative: Being Away, Extent, Fascination and Compatibility (Kaplan and Kaplan 1989). The level of any of these components varies across environments, including those distinguished by the amount of natural or built elements. Although some restoration could occur when only one component is met- for example, physically being away from one’s work environment- ART suggests that greater restoration would be yielded in an environment comprising all four. Further, because these components are considered to be products of person-environment interactions (Kaplan 2001), a determination of whether a specific environment is restorative can differ across people and across time (Hartig, Korpela, Evans, and Garling 1997b; Hartig et al. 1991; Kaplan and Kaplan 1989). It is argued, however, that restorative environments are more likely to be comprised of natural elements such as trees and grassy spaces, than built elements such as brick buildings and concrete surfaces (Hartig, Kaiser, and Bowler 1997a; Hartig et al. 1997b; Hartig et al. 1991; Herzog, Chen, and Primeau 2002; Herzog, Maguire, and Nebel 2003; Kaplan and Kaplan 1989; Kaplan 1995; Korpela, Kytta, and Hartig 2002; Kuo 2001; Kuo, Bacaicoa, and Sullivan 1998; Kuo and Sullivan 2001; Laumann, Garling, and Stormark 2001).

Being Away

One of the key components of a restorative environment is that of Being Away from the attentionally fatiguing features of daily life. This component is further divided into two different types of being away– that of being away physically, i.e., in a different geographical location, and that of being away psychologically, i.e., a conceptual shift (Kaplan and Kaplan 1989; Kaplan 1995, 2001b). The phrase “wanting to get away from it all” in conjunction with a holiday destination is reminiscent of the physical Being Away component of restorative environments (Kaplan and Kaplan 1989). Being away, however, does not require the traversing of any great distance. It could be achieved simply by changing the “direction of one’s gaze” (Kaplan 1995, 173). Psychologically being away can be achieved by changing the usual cognitive content experienced (Kaplan and Kaplan 1989) and by engaging in thoughts or activities that are distinct from the everyday (Kaplan 2001b). There are three ways proposed to elicit “being away” psychologically (Kaplan and Kaplan 1989): removing distractions from the immediate environment, not completing usual work for a period, or ceasing the pursuit of certain purposes. The sense of being away is proposed to be strongest when all three are performed. Yet an environment that meets these three may not be restorative; it may be considered as confining and/or boring, such as a basement without a telephone (Kaplan and Kaplan 1989). Therefore, a restorative environment, while facilitating either Being Away physically or psychologically from the environment of everyday stressors, must be comprised of additional components. One such component is Extent.

Extent

For an environment to have Extent, it must be “rich enough and coherent enough so that it constitutes a whole other world” (Kaplan 1995, 173). The sense of another world is congruent with the “Being Away” component just described. Extent, or the sense of a whole other world, is defined by the presence of / experience of two components: “connectedness” and “scope” (Kaplan and Kaplan 1989). Extent, via connectedness and scope, can pertain to the physical environment or to psychological processes. Connectedness is defined by a series of relationships between environmental features. That is, a connected environment is comprised of features that relate to each other and which, when considered as a set of features, are related to a larger environment (Hartig et al. 1997b; Kaplan and Kaplan 1989). Scope refers to breadth and indicates there is more than the immediate environment available, either physically just out of sight or even within the imagination (Hartig et al. 1997b; Kaplan and Kaplan 1989). Psychological examples include “debugging a computer program or repairing a car or producing a showpiece cake” (Kaplan and Kaplan 1989, 184); activities which can be absorbing to the point of an individual no longer being aware of their surroundings. Physical examples include both non-natural environments such as shopping malls, casinos and zoos, all designed to give a sense of a whole other world, as well as natural environments such as distant wilderness or trails and paths designed to give the illusion of small areas being larger (Kaplan 1995). Although not in relation to ART, perhaps this component is better explained by Cobb (1969): “The child’s sense of wonder, displayed as surprise and joy, is aroused as a response to the mystery of [the] stimulus [of nature] that promises more to come or, better still, more to do” (28). Such promises may draw children into environments, but for children to remain, the environment must be interesting, the third component.

Fascination

Elements of Extent elicit Fascination, another component of restorative environments. Fascinating elements can be comprised of physical stimuli within the environment such as “sex and violence, competition and cooperation” (Kaplan and Kaplan 1989, 184) or “animals, people, water, fire” (Herzog, Black, Fountaine, and Knotts 1997, 165). Fascination can also occur via cognitive involvement- i.e., processes such as learning (Kaplan and Kaplan 1989), story telling, or problem solving (Herzog et al. 1997).

The contribution of Fascination is crucial for two reasons. First, fascinating physical stimuli or cognitive processes are by definition interesting and therefore “attract people and keep them from getting bored” (Kaplan and Kaplan 1989, 184). Secondly, fascinating stimuli or processes do not require the use of conscious, effortful directed attention (Kaplan and Kaplan 1989). Fascination alone however, does not complete a restorative experience. “People can be fascinated by events that may negatively affect their mental and emotional states, such as violence” (Hartig et al. 1997b, 178). Therefore, researchers formed the concept that fascination ranges along a continuum from hard fascination to soft fascination based on the elements of the stimuli (Kaplan and Kaplan 1989). Hard fascination involves dramatic attention grabbing stimuli such as watching auto racing (Kaplan 1995) and staring at horrific sights, or giant trees and vast waterfalls (Herzog et al. 1997; Williams 1998). Soft fascination is based on stimuli that “hold the attention in an undramatic fashion” (Williams 1998, 60) and includes activities such as learning (Kaplan and Kaplan 1989) and elements of nature such as sunsets and waterfalls, caves and fires (Kaplan and Kaplan 1989), leaves moving in a breeze and cloud formations (Kaplan 1995; Williams 1998). Elements closer to the soft fascination end of the spectrum are considered to be restorative (Kaplan 1995; Williams 1998) as they provide opportunities for effortless attention (Kaplan 2001b). As such, attention can be maintained over an extended period of time. Remaining in an environment, however, is dependent upon its compatibility with the purpose of entering the environment- the final component considered to comprise a restorative environment.

Compatibility

The fourth and final component of a restorative environment is that of Compatibility; that is, the environment is believed to be supportive of an individual’s purpose while in the environment (Kaplan 1983; 1995; 2001b). There are a number of ways an environment could be incompatible with an individual’s desired behavior (Kaplan 1983). For example, environmental conditions may dictate certain behaviors at odds with individual goals, or stimuli within the environment may distract from goals. Alternatively, some predetermined activities may elicit spontaneous behaviors. For example, children may decide to play catch, a relatively stationary game, and end up playing brandy, where the ball is thrown at a running person. For an environment to be considered restorative, it must not only support the specific activities, but also any inclinations of the individual. Therefore compatibility is “a function of environmental dictates and personal intentions” (Hartig et al. 1997b, 178). If the environment or elements of the environment are not, or are perceived to be not supportive of specific activities proposed, then the environment is not compatible (Kaplan 1983). If the component of Compatibility is met, environmental elements are supportive of and congruent with required behavior, and functions can be performed without difficulty or distraction, an important component of a restorative environment.

Restorative Component Measures

Although the four components of restorative environments were initially proposed two decades ago (Kaplan 1983), the development of measures for these constructs has only recently been addressed and has focused on young adult samples. In 1997, Hartig and colleagues, continuing work initially published in 1991, conducted a series of studies in order to develop the Perceived Restorativeness Scale (PRS) with varied success (Hartig et al. 1997b). Overall, their data yielded two factors: a General Restorativeness Factor and a Coherence Factor. They suggested further improvements were warranted to strengthen the measure, a point reiterated by Korpela et al. (2001) after finding a similar factor structure with favorite and unpleasant places. Hartig et al. (1997b) did, however, caution against a simplified model being pursued.

Development of a measure of the five restorative components was subsequently conducted by Laumann et al. (2001), with 22 items loading onto five factors. This measure will be referred to as the Restorative Components Scale (RCS) in keeping with Herzog et al. (2003). Interestingly, the a priori Being Away items split into two factors: a physical being away, novelty (e.g., I am in a different setting than usual) and a psychological being away, escape (e.g., I am away from my obligations). This empirical finding is in keeping with Kaplan’s original definitions of Being Away referring to either physical or psychological aspects. The other factors were as proposed by ART: Extent (e.g., the elements here go together), Fascination (e.g., this setting has many things that I wonder about) and Compatibility (e.g., I am capable of meeting the challenge of this setting).

Both these measures, however, have been developed for use with adult participants. The goal of this study was to develop and evaluate a measure of restorative components appropriate for the use with children. Consistent with Hartig et al.'s (1997b) goals, the measure for children should capture the components of ART, as well as differentiate environments that differ theoretically in their level of restorativeness. Therefore, two familiar, everyday environments were evaluated by each child in this study; that is, their school playground and their school library. These environments were compared to their school classroom. A school library is clearly more similar to the classroom (i.e., indoors, chairs and tables, used during class time, few natural elements) than the school playground (i.e., outdoors, play equipment, used during breaks, more natural elements) and therefore would theoretically be perceived as less restorative.

The decision to focus on children’s school environments is based on theoretical and social considerations. First, for an environment to be restorative it needs to be accessible (Kaplan 1995). Given the wide reaching changes in recent decades and, in turn, increased parental reluctance to permit children to play unsupervised, opportunities for children to be able to wander from home to local fields or parks have decreased (Kellert 2002). Second, a number of studies have indicated that children aged six through 12 more often spend time in nearby or familiar places rather than unusual or spectacular places (Moore 1986; Nabhan and Trimble 1994; Sobel 1990). Schoolyards may be one of the few open areas that children have available to them on an everyday basis. Finally, as school attendance is compulsory, children not only attend school environments regularly but for lengthy periods; that is, up to seven years at the same campus (i.e., Prep to grade 6 in Australia). As proposed previously (Faber Taylor et al. 2001), the combination of accessibility and the academic goals of education systems suggests the importance of considering the perceived and actual restorativeness of school environments.

Sex differences across restorative environment components have yet to be examined with adult samples. Prior research with children, however, has reported sex differences in environments close to home. For example, studies have demonstrated associations between views of nature from home and attentional benefits for girls (Faber Taylor et al. 2001; Faber Taylor, Kuo, and Sullivan 2002), but not boys, possibly because boys may not play as close to home (Faber Taylor et al. 2001). Bagot and Gullone (2003) report that girls indicated higher positive affect from natural stimuli (e.g., being with animals, activities in natural settings) than boys did. Although natural elements and stimuli are not specifically captured in this study, and the environments of interest are away from the home setting, such research suggests sex differences could be expected in school settings as well.

As research into restorative environments has focused on adult samples, there has not been an examination of developmental differences in perceived restoration potential or attention restoration to date. Research has however revealed differences across developmental stages in environmental perception and preferences (e.g., Balling and Falk 1982; Bernaldez, Gallardo, and Abello 1995; Herzog, Herbert, Kaplan, and Crooks 2000; Malinowski and Thurber 1996; Zube, Pitt, and Evans 1995), suggesting differences in perceived restoration may be revealed with children of different ages.

As the RCS will be used as a base for the children’s adapted measure, it is expected a five-factor measure (due to Being Away being split across two factors) will similarly form. As the level of restorative components can vary across environments, the evaluation of two environments allows a test of discriminant validity. It is proposed that school playgrounds will score higher than the school library on all five restorative components, as well as on a combined potential restorativeness score. In addition, as mentioned earlier, restorative environments differ across people and time. Although time is not captured in this study, differences across people may be considered as broadly reviewed with an examination of response by sex and age. It is expected girls will report the playground (i.e., the more natural environment) as more restorative than boys. As this is a preliminary study involving children with a limited age range, no specific hypotheses are generated regarding developmental differences; however, differences between younger and older children will be examined and reported.

Method

Participants

A total of 230 primary school children participated in the study. As five were deleted due to missing data and outliers (detailed in Results), the final sample consisted of 112 boys (age: M=9.9, SD=.97 years) and 113 girls (age: M=9.96, SD=1.01 years) with all children being aged between 8 and 11 years (age: M=9.93, SD=.99 years). This age group was selected as they are able to read and complete questionnaires alone and in Australia, children in Grades 3 to 6 (approximately 8 to 11 years) share the same school (i.e., primary school). There were 74 children aged 8 and 9 forming a younger sub-sample (age: M=8.7, SD=.46 years) and 151 children aged 10 and 11 for the older sub-sample (age: M=10.53, SD=.50 years). Participants were recruited from nine schools across north-east and south-west suburbs of Melbourne (a large south-eastern Australian city) covering a broad range of socio-economic and cultural groups.

Via an active consent procedure, there was a response rate of approximately 20 percent. It should be noted, however, that only three of approximately 1,200 parents approached actively denied permission for their child to participate.

Measures

The initial item pool that formed the basis of the Perceived Restorative Components Scale for Children (PRCS-C) was drawn from previously published adult measures. With the permission of the first authors, items were adapted for children, from both the RCS by Laumann et al. (2001) and the PRS by Hartig et al. (1997b).

Items were selected on the basis of the following criteria: they should be representative of Kaplan’s restorative components, applicable to children’s environments (i.e., for this study they would make sense in a school setting), comprehended by children or readily adapted for use with children, and they should have loaded onto appropriate factors in prior research involving the development of adult perceived restorative environment measures. Due to adult research published previously, a smaller initial item pool than otherwise was used, minimizing the disruption to participating schools and the required time and effort of children. A primary school teacher and a primary school environmental educator reviewed the final wording of items for suitability with children.

The adapted items comprised 23 questions representing the factors proposed by ART and more specifically, the five factors as reported by Laumann et al. (2001): seven items for Being Away (three for physical and four for psychological), five items for Fascination, six items for Extent and five items for Compatibility.

To ensure children’s responses were directly related to the environments of interest, the survey included more specific phrasing than has been used with adult samples. For example, prior research with adults (e.g., the RCS) asked participants to complete the following item - “I do something different than I usually do when…” while imagining themselves in an environment presented by video. For the PRCS-C, the name of the specific environments (in italics here for identification only) were included in the item, i.e., “When I am in the playground, I do different things than in the classroom” or “The things I like to do can be done in the library.” Although these items contain direct language, it was deemed to be appropriate to facilitate accurate completion by children as young as eight years.

Children were asked to think about how true each statement was for them and to circle the answer that suited them best. The response scale was a five-point Likert scale (0-Not at all, 1-A little, 2-Sometimes, 3-A lot, 4-Completely) with only the words printed across the page under each question for children to circle; that is, there was not a numbered scale on the page. Children readily answered questions by circling the answer appropriate for them. Items for each factor were distributed throughout the questionnaire, so no items for the same factor followed each other.

Questionnaire packages comprised a demographics page (items included grade level, age and sex) and the restorative component items to be completed for both the school playground and the school library. Approximately ten minutes separated the completion of the two sets of restorative component items (i.e., for the two environments) during which timed performance tasks were completed. Children did not exhibit any signs of fatigue and indicated they enjoyed participating in the research. The data presented comes from a larger study designed to evaluate a number of measures for children (i.e., restorative components, attention performance). Only the psychometric evaluation of the PRCS-C is reported here. The items were presented in the same order, but were completed for playground and library in counterbalanced order across schools.

Procedure

After the study was approved by the Monash University Ethics Committee and Victorian governing bodies of schools, principals of schools were approached to participate in the study. Children in grades three to six of participating schools received an explanatory statement and consent form for their parents’ review. In addition to parental approval, children’s assent was also obtained prior to data collection (note: two children decided not to participate).

Data collection was conducted during school hours in groups of up to approximately 30 children, based on children’s grade levels (Grades 3 and 4 together, Grades 5 and 6 together). The majority of participants could see a small section of their playground (from the classroom window), but not the library during completion. Items were read aloud to all groups of children, taking approximately 30 minutes, with older children working at their own pace. Children were encouraged to ask questions if they were unsure of the meaning of any words and advised there were no right or wrong answers. Once complete, children were advised of the purpose of the study and thanked for their time.

Results

Data was analysed using SPSS (v. 11.0) and included factor analysis to determine the factor structure of the restorative components measure. Resulting factors were then examined for internal consistency with Cronbach’s alpha, with factor scores being compared across environments and by sex and age via dependent and independent t-tests. Factor analysis was chosen over Principal Components Analysis (PCA), as a theoretical solution was of primary interest.

First, missing value analysis was conducted. Three participants had more than half of their data missing and were therefore deleted. The majority of variables had less than 1 percent of data missing and no variable had more than 1.8 percent missing (i.e., four participants did not answer). Missing values were replaced by estimated values via Expectation Maximisation (EM) method and were deemed to reflect actual values. The data formed part of a larger study where univariate outliers were retracted to ±3.29 standard deviations from the mean (no univariate outliers for the restorative components items) and two multivariate outliers were reviewed and deleted. No major violations of factor analysis assumptions were detected. Sample size was not optimal (i.e., N<300), an issue discussed later, and as expected, some items were not normally distributed. They were however, not transformed due to the sample size (i.e., N>30) and difficulty in interpreting transformed variables (Tabachnick and Fidell 1996). Initial analysis indicated the playground restorativeness data was suitable for factor analysis. Features noted were: presence of correlations over .3, Kaiser-Meyer-Olkin Measure of Sampling Adequacy was .82 and Bartlett’s Test of Sphericity was significant.

A preliminary factor analysis was conducted using Maximum Likelihood extraction with direct oblimin (Delta = 0) rotation, as it was expected the factors would be correlated. There were seven initial eigen values over 1.0 explaining 61 percent of the variance. The scree plot indicated five factors could be extracted, with an initial break between the first two and then the next three. The presence of five factors is also consistent with prior research developing the adults’ RCS, on which this study is based.

A second factor analysis was then conducted requesting five factors to be extracted. Eigen values ranged from 3.38 to .53, explaining 47 percent of the variance. Although three eigen values were less than 1.0 after extraction, examination of the results continued as it is argued a larger sample size would yield five factors with this number of items. Four items were deleted as they did not load onto any factor, one item was deleted as it loaded at .3 on two factors and two items were deleted as they did not load onto theoretically-driven a priori factors. A third factor analysis was then run resulting in an additional item to be deleted as it did not load onto any factor. This resulted in 15 items remaining (see Table 1 for items and factor loadings) comprising five factors, revealing Thurstone’s Simple Structure. (Note: The negative item loadings on the Compatibility factor indicate this factor correlates negatively with others, within the Factor Correlation Matrix.) The two Being Away factors were sub-titled Physical (instead of Escape) and Psychological (instead of Novelty), to simplify interpretation.

It should be noted that results are reported here (see Table 1) for Factor Analysis via Maximum Likelihood as the focus is on the theoretical structure of the data/measure. This provides a more stringent test than PCA as both error and communality variance have been removed from results.


Analyses were repeated for the Library data with remarkably similar results. Results indicated data was suitable for factor analysis: presence of correlations over .3, Kaiser-Meyer-Olkin Measure of Sampling Adequacy was .83 and Bartlett’s Test of Sphericity was significant. Factor analysis with Maximum Likelihood extraction and direct oblimin rotation (Delta=0) was conducted. Again, seven eigen values over 1.0 were reported, explaining 60 percent of the variance. Scree plot examination revealed five factors to be extracted. Factor analysis was re-run with five factors requested. Eight items were deleted; four did not load onto any factor, two were single-item factors, and another two items loaded onto factors that did not make theoretical or empirical sense. Six items were the same as those deleted from the playground data. A third factor analysis was then run, comprising five factors and revealing Thurstone’s Simple Structure with the 15 remaining items.

Library data yielded three factors comprising the same items revealed in the playground data– Being Away- Physical, Being Away- Psychological and Fascination. Compatibility, however, was split over two factors and an Extent factor was not revealed (a priori Extent items were either not loading at all, or loading onto the Compatibility or Being Away– Physical factors).

Table 1.Factor Analysis Pattern Matrix for 15 Items of ERM-C (Playground Data)

I determined that factors for both environments would be based on the 15 items retained from the Playground data as they met both theoretical and statistical criteria. Although the library data yielded a similar factor structure, it did not exactly replicate the structure revealed within the playground data; that is, no Extent factor was revealed.

Cronbach’s alpha was calculated for the five factors for both the playground and the library data to determine internal consistency. As shown in Table 2, highest values were yielded with the four-item factor Fascination and, as would be expected, the lowest values were yielded with the two items of Extent.

Table 2. Cronbach’s Alpha for Factors of School Playground and School Library

Correlations between the five factors were conducted and are reported in Table 3. All associations across both environments were positive and significant except for Fascination and Being Away– Physical for the playground. Generally the relationships between the factors were stronger for the library environment.

Table 3. Pearson’s Correlations between the Five Factors for the Two Environments

In order to evaluate whether the PRCS-C would distinguish between environments, results for the playground and the library environments were compared. A mean item score for each of the five factors was computed for both environments. As shown in Table 4, means for the five factors and the total perceived restorativeness score (computed by adding values of the five factors) were higher for playground than library, except for Fascination. Six (five factors, one total score) dependent samples t-tests were conducted yielding significant differences (i.e., p<.001) for all values (t values are reported in Table 4). The playground environment scored significantly higher than the library on Being Away- Psychological, Being Away- Physical, Compatibility and Extent. The library however, scored significantly higher than playground on Fascination. The total perceived restorativeness score of the playground was significantly higher than the library. These results (i.e., factor scores across environments) were then examined by sex and age group (refer to Table 4). Girls and older children reported similar results to the total sample; that is, the playground scores were higher across all factors and the total perceived restorativeness score, except that Fascination scored significantly lower. For boys and younger children however, the level of Fascination was equal across both environments.

Table 4. Means, Standard Deviations and Dependent Sample t- tests for the Five Factors and Total Restorativeness Score between Two Environments for the Total Sample and by Sex and Age

To investigate whether the PRCS-C yielded differences between groups of people, mean factor scores were calculated for each factor across both environments and are reported in Table 4. Ten (five factors x two environments) independent samples (sex) t-tests were conducted (Bonferroni adjusted significance alpha .05/10 = .005). Results revealed that boys scored significantly higher than girls on the playground Being Away- Psychological factor [t(223)=3.29, p=.001]. Results also revealed boys scored higher than girls on the library Being Away- Physical factor [t(223)=2.54, p=.01]. Independent sample t-tests were repeated by age group, with significance alpha again at .005. No significant differences were revealed between younger and older children. The total perceived restorativeness score was also examined by sex and age for each environment. Two (environment) independent samples (sex) t-tests and two (environment) independent samples (age group) t-tests were conducted (Bonferroni Adjusted significance alpha .05/2 = .025 for each) revealed boys scored the playground significantly higher on total perceived restorativeness restorative than girls [t(223)=3.26, p=.001]. No significant differences were reported by age group.

Discussion

Overall, the goals of the current study were met. The development of a measure suitable for children, comprising the proposed factors of restorative environments and able to differentiate between environments and groups was accomplished– the PRCS-C. The specific items and administration order are included in the first two columns of Table 1.

Playground Data

As predicted, five factors could be extracted from the data, reflecting a priori scales reported by Laumann et al. (2001). Items for each factor are presented in Table 1. Factors included: Being Away- Physical (“… I feel as though I am in different surroundings …”), Being Away- Psychological (“… I am away from things I must do”), Fascination (“There are many things … that I find fascinating”), Compatibility (“The things I want to do can be done …”), and Extent (“I can do different things in different areas …”). This 15-item five-factor structure was found with the playground data.

Library Data

A very similar although not identical factor structure was also yielded with the library data. Of eight items removed from playground data, six were also removed from the library data indicating similar underlying structures in the data. The two items retained by the library data, but excluded from the playground data were a Fascination item and a Compatibility item. Overall, differences lay in the Compatibility and Extent items. Compatibility items were split into two factors and no Extent items remained within the library data.

PRCS-C

In short, both sets of data yielded five factors with 15 items with three factors consistent: Being Away– Physical, Being Away– Psychological and Fascination. Compatibility was similar in both data sets, albeit split in the library data, and Extent was not replicated in the library data. Extent may have been particularly difficult to capture with the library environment due to its typically small size and the similarities in use to the classroom.

As the factor analysis results using the playground data yielded both theoretically and statistically appropriate outcomes, these items were used to comprise the restorative components for both environments for further analysis. For both the playground and library data, internal consistency for the factors was considered acceptable for four of the five factors. Although alpha values of around .7 are considered optimal, lower values are expected and acceptable with scales containing few items (Anastasi and Urbina 1997). Results for the Extent factor, however, are particularly low. This is consistent with prior research, as the two Extent items that were retained were adapted from the PRS developed by Hartig et al. (1997b) who also reported similar issues (i.e., number of items, internal consistency) with the Extent factor. While additional Extent items were included in the initial pool of items, they did not perform well.

All the factors were positively associated. Consistent with prior research (Laumann et al. 2001), the factors correlated significantly across both environments except for Being Away- Physical and Fascination for the playground, the more natural environment. This result is, however, somewhat inconsistent with Laumann et al.’s finding, which occurred within the less natural (i.e., city) environment.

Environment Differences

When comparing the two environments, the more natural environment of the playground yielded higher scores than the library across all factors as expected (Hartig et al. 1997a; Laumann et al. 2001), excluding Fascination, which scored lower. These relationships were all significant, indicating the PRCS-C was distinguishing between the two school environments on the basis of each of the five factors. The unexpected finding of Fascination scoring lower for the playground may be due to the environments in question. Perhaps school libraries, with computers and Internet access contain more opportunities for interesting elements to be considered than the different areas and play equipment of the school playground.

When the playground and library factor scores were compared by sex and age group, the overall results were replicated for girls and the older age group; that is, playground scored higher across restorative components excluding Fascination, which scored lower. Unlike girls and the older age group, however, results indicated that boys and the younger age group, reported that both environments contained interesting elements. It would appear girls and older children found the libraries of their school environments more interesting than their playgrounds and that this accounts for the lower playground Fascination score reported with the total sample. Although not highlighted as a potential sex nor age difference, it was proposed that Fascination could comprise physical stimuli or cognitive involvement (Kaplan, 1989). It can be argued the playground perhaps provides more physically focused elements whereas the library provides more cognitively focused elements of interest. While boys and younger children reported equal levels of Fascination, perhaps indicating divided interest in physical and cognitive stimuli, girls and older children may have indicated a preference for cognitive stimuli when they reported a higher Fascination score for the library. Although all children who participated in this study would have reached the concrete operational thought stage (Piaget 1980), perhaps older children have developed a preference or capacity for more cognitive stimuli than the younger children. The results reported here for both the total sample and by sex and age group indicate the PRCS-C is able to distinguish between environments, a key criterion for validity. Results also indicate possible sex and age differences when considering perceived restorative component levels of environments, particularly Fascination.

Examination of a total perceived restorativeness score, by adding scores of the factors together, was conducted. The assumption, however, that each factor contributes equally to a restorative environment and thus to restoration of attentional capacity does remain a theoretical one. The total perceived restorativeness scale (i.e., all items), yielded strong internal consistency and was able to distinguish between the two environments, with the playground scoring higher on restoration potential than the school library, as predicted by ART (Kaplan and Kaplan 1989; Kaplan 1995) and prior research (Hartig et al. 1997a; Hartig et al. 1997b; Hartig et al. 1991; Herzog et al. 2002; Herzog et al. 2003; Korpela et al. 2002; Laumann et al. 2001). This finding was replicated for both boys and girls and both younger and older children.

Group Differences

Although sample size did not allow for factor analysis to be conducted by sex nor age, factor scores by sex were compared as well as by age groups, as a broad indicator whether the PRCS-C could differentiate between groups of people’s reports of perceived restorativeness and to assess any developmental differences. Across the five factors for the two environments (i.e., ten comparisons), two differences were revealed between boys and girls. First, boys reported higher Being Away- Psychological scores for the playground, indicating that boys feel freer and less constrained than girls when on the playground. Girls’ lower scores on this measure may be influenced by the fact that they find the playground less interesting than the boys do, as reflected in the girls’ lower scores for playground Fascination. Second, boys also reported higher Being Away- Physical scores for the library than girls, indicating that boys perceive a greater difference between the library and classroom than girls. Results revealed no differences between the younger and older children on the five factors for the two environments. As the children were from the same developmental stage (Piaget 1980), it is not surprising these direct comparisons did not reveal differences between the younger and older children; it would suggest younger and older children in this limited age range (i.e., 8-11 years) experience these environments as more similar than boys and girls do.

When the total perceived restorativeness scores for boys and girls were compared for both environments, boys reported the playground as having higher restoration potential than girls. As the playground would contain more natural elements than the library, and prior research suggests a stronger association between girls and natural environments, than boys (Bagot and Gullone 2003; Faber Taylor et al. 2001, 2002), this could be considered an unexpected finding. Prior research, however, reports a relationship between boys and natural elements in play areas away from home (Faber Taylor et al. 2001), one example of which is school playgrounds. Boys and girls reported similar restoration potential for the library. When results for younger and older children were compared, both age groups reported similar levels of total perceived restorativeness for the two environments. That is, the playground was reported to have similar restoration potential for both younger and older groups. This was also reported with the library environment.

Overall, results indicate the PRCS-C is able to detect differences in the factor scores across groups of people, as would be expected when evaluating restoration potential of environments with different groups. Combining results for the comparisons across environments (i.e., comparing playground and library scores) and between groups (i.e., comparing results between boys and girls, and between younger and older children), indicates girls report the library to be more like their classroom and more interesting than the playground, and report the playground to have lower restoration potential than reported by boys. The only developmental difference was reported within environment comparisons with older children reported the library to be more interesting than the playground, whereas younger children reported the two environments to be of similar interest.

Evaluation of Measure

The findings reported here are important for three reasons. First, they indicate the measure can distinguish between different types of environments. Second, they establish that the measure can detect differences across different groups of people. Both are key criteria when determining the validity of a measure of restorative environments. Third, the overall results are in keeping with predictions (i.e., the library should score lower than the playground as it would more closely resemble a classroom than outdoors), supporting the fact that the PRCS-C is functioning as would be expected. The finding that Fascination scored higher in libraries than playgrounds is intriguing and is attributed to the girls’ and older children’s responses. Although too early to determine the implications of these differences, it does add to the small body of research reporting sex differences within children’s environmental research (Bagot and Gullone 2003, Faber Taylor, et al. 2001, 2002), and indicates that sex differences should be considered where sample size allows.

Limitations and Future Research

Although deemed a fair size and adequate by the ten participants per item rule (Tabachnick and Fidell 1996), the sample of 225 participants has potentially influenced the factor structure outcomes through low eigen values for some of the factors. It is argued that a larger sample would yield strengthened results, an issue to be addressed by future research.

The key limitation regarding the outcome of the measure is the Extent factor. It contains only two items and had a low internal consistency result. Both items address the Scope component of Extent, and did not cover Connectedness. Future work would do well to consider additional items for this component.

The current study’s methodology focused on adapting items drawn from adult measures. The inclusion of qualitative methods such as focus groups and semi-structured interviews could further understanding of children’s perceptions and experiences of their environments. In turn, additional items not previously representative of the five factors in adult work could be included and further examination of the factor structure reported by children be conducted.

Understanding the development of restoration and the role of restorative environments across the lifespan requires measures suitable for each developmental stage. Work with those in early adulthood is underway, and this initial study indicates that children should be considered in restorative environment research, specifically emphasizing late childhood (i.e., aged 8 to 11). The importance and distinction of restorative environments for different developmental stages is yet to be investigated. Although younger and older children generally reported similar results in this study, the sample did not fully cover any developmental period. The inclusion of two or more developmental stages would more likely yield differences. With a measure of the restorative components available, an evaluation of the restoration potential of children’s environments can be made and thus contribute to the theoretically proposed components of the restoration of attentional capacity in children. Consideration of the relationship between the restorative components individually or combined and ART with individual factors such as preferences and attitudes also awaits future research.

A measure of the level of naturalness within environments would have been of interest in this study. Obviously school playgrounds have a higher level of natural elements than school libraries, and results reported support the proposal that restorative environments are more likely to be natural environments. A measure of levels of nature, however, would have facilitated the determination of whether the measure could distinguish between familiar, everyday outdoor environments on that basis. This was beyond the scope of this initial study, but is highlighted as a direction for future research.

Implications and Recommendations

With the development of a measure that allows children to report levels of restorative components, a number of exciting avenues have been opened. First, the PRCS-C provides an opportunity for children to evaluate the perceived restorativeness of environments they frequent. This allows researchers investigating the role of the physical environment in children’s lives to include perceived restorativeness in studies they conduct. It also provides an opportunity for children to be involved in and contribute to such research. Any environment in which children spend time could be considered, including hospitals, day care centers, holiday program facilities and community playgrounds. Second, and more specifically, the restorative components play a key role in ART. With the development of the PRCS-C, ART can be evaluated from a child’s perspective; that is, children’s attentional capacity pre- and post- spending time in environments with varying levels of restorative components can be examined.

Future ART work could use the PRCS-C to evaluate the perceived restorativeness of specific environments and possible attentional advantages for children. For example, children who spend play periods on school grounds which comprise the five restorative components may exhibit academic and social benefits in the classroom (Faber Taylor et al. 2001). For those children with attentional difficulties, seeking out environments identified to be high in restoration potential would provide a possible resource for strengthening their attentional capacity.

The initial sex and age differences reported in restorative components, combined with prior work reporting sex and age differences in relationships with nature, indicate that separate consideration should be given to boys’ and girls’ environments. For example, separate playground areas or varied grade level access may be considered in co-education schools; single-sex schools might consider different types of playgrounds. The role of other areas such as the library or art room during classroom breaks could also be considered. Consideration of sex and age differences does extend to ART as a whole.

Conclusion

The current study reports the initial development and psychometric validation of the PRCS-C and provides a number of interesting findings. First, with items drawn from adult work, the adult factor structure was replicated and children reported a similar structure in the restorative components of their environments. Second, the measure distinguishes between two environments that are both familiar to children and in close proximity. Third, a small number of sex and age differences on the perceived restorative components and total perceived restorativeness are reported. Fourth, overall, playgrounds have higher restoration potential than libraries, consistent with prior research and theoretical proposals of the benefits of natural environments.

The results provide initial support for the PRCS-C and continued development and application of the measure in future research is warranted. It is hoped the PRCS-C provides an opportunity for restorative environment researchers to consider children in their work, and indeed, for child environment researchers to consider perceived restorativeness as a factor.

Endnotes

1. This paper was awarded Third Prize in the 2003 CYE Graduate Student Paper Award for Excellence in Research competition.

2. The author would like to thank Dr. Karin Laumann and Terry Hartig who generously allowed items from their adult scales to be adapted for this measure, the principals and pupils of participating schools for their interest and time, and the CYE reviewers for their comments and feedback.

Kathleen Bagot is a part-time Ph.D. Candidate in the Department of Psychology, Monash University, Australia. Her research interests include the relationship between children and natural environments with a focus on beneficial cognitive and affective outcomes in line with principles of Positive Psychology. Her thesis is supervised by Felicity C.L. Allen (Monash University) and Frances E. Kuo (University of Illinois).

References

Anastasi, A. and S. Urbina (1997). Psychological Testing. New Jersey: Prentice-Hall, Inc.

Bagot, Kathleen L. and Eleonora Gullone (2003). "Development and Psychometric Evaluation of the Positive Affect Schedule for Youth (Pas-Y)." Behaviour Change 20(2): 63-75.

Cobb, E. (1969). "The Ecology of Imagination in Childhood." In Shepard, P. and D. McKinley, eds. The Subversive Science: Essays toward an Ecology of Man. Boston: Houghton Mifflin, 97-131.

Driver, Jon (2001). "A Selective Review of Selective Attention Research from the Past Century." British Journal of Psychology 92(1): 53-76.

Faber Taylor, Andrea, Frances E. Kuo and William C. Sullivan (2001). "Coping with ADD: The Surprising Connection to Green Play Settings." Environment and Behavior 33(1): 54-77.

Faber Taylor, Andrea, Frances E. Kuo and William C. Sullivan (2002). "Views of Nature and Self-Discipline: Evidence from Inner City Children." Journal of Environmental Psychology 22: 49-63.

Grahn, Patrik, et al. (1997). "Ute Pa Dagis (Outdoors at Daycare)." Stad och Land (City and Country) 145.

Hartig, Terry, Florian G. Kaiser and Peter A. Bowler (1997a). Further Development of a Measure of Perceived Environmental Restorativeness. Uppsala: Uppsala University.

Hartig, Terry, et al. (1997b). "A Measure of Restorative Quality in Environments." Scandinavian Housing and Planning Research 14: 175-197.

Hartig, Terry, Marlis Mang and Gary W. Evans (1991). "Restorative Effects of Natural Environment Experiences." Environment and Behavior 23(1): 3-26.

Herzog, Thomas R., et al. (1997). "Reflection and Attentional Recovery as Distinctive Benefits of Restorative Environments." Journal of Environmental Psychology 17: 165-170.

Herzog, Thomas R., Hong C. Chen and Jessica S. Primeau (2002). "Perception of the Restorative Potential of Natural and Other Settings." Journal of Environmental Psychology 22: 295-306.

Herzog, Thomas R. and Theresa A. Gale (1996). "Preference for Urban Buildings as a Function of Age and Nature Context." Environment and Behavior 28(1): 44-72.

Herzog, Thomas R., Colleen P. Maguire and Mary B Nebel (2003). "Assessing the Restorative Components of Environments." Journal of Environmental Psychology 23: 159-170.

Kaplan, Rachel (2001). "The Nature of the View from Home: Psychological Benefits." Environment and Behavior 33(4): 507-542.

Kaplan, Rachel and Stephen Kaplan, eds. (1989). The Experience of Nature. Cambridge: Cambridge University Press.

Kaplan, Stephen (1983). "A Model of Person-Environment Compatibility." Environment and Behavior 15(3): 311-332.

Kaplan, Stephen (1995). "The Restorative Benefits of Nature: Toward an Integrative Framework." Journal of Environmental Psychology 15: 169-182.

Kaplan, Stephen (2001). "Meditation, Restoration, and the Management of Mental Fatigue." Environment and Behavior 33(4): 480-506.

Kellert, Stephen R. (2002). "Experiencing Nature: Affective, Cognitive and Evaluative Development in Children." In Kahn, Peter H., Jr. and Stephen R. Kellert, eds. Children and Nature: Psychological, Sociocultural and Evolutionary Investigations: MIT Press, 211-261.

Korpela, Kalevi, Marketta Kytta and Terry Hartig (2002). "Restorative Experience, Self-Regulation, and Children's Place Preferences." Journal of Environmental Psychology 22: 387-398.

Korpela, Kalevi, et al. (2001). "Restorative Experience and Self-Regulation in Favorite Places." Environment and Behavior 33(4): 572-589.

Kuo, Frances E. (2001). "Coping with Poverty: Impacts of Environment and Attention in the Inner City." Environment and Behavior 33(1): 5-34.

Kuo, Frances E., Magdalena Bacaicoa and William C. Sullivan (1998). "Transforming Inner-City Landscapes: Trees, Sense of Safety, and Preference." Environment and Behavior 30(1): 28-59.

Kuo, Frances E. and William C. Sullivan (2001). "Aggression and Violence in the Inner City: Effects of Environment Via Mental Fatigue." Environment and Behavior 33(4): 543-571.

Laumann, Karin, Tommy Garling and Kjell Morten Stormark (2001). "Rating Scale Measures of Restorative Components of Environments." Journal of Environmental Psychology 21: 31-44.

Moore, R.C. (1986). Childhood's Domain: Play and Place in Child Development. London: Croom Helm.

Nabhan, G.P. and S. Trimble (1994). The Geography of Childhood: Why Children Need Wild Places. Boston: Beacon Press.

Piaget, Jean (1980). Six Psychological Studies. Brighton, Sussex: Harvester.

Piontkowski, Dorothy and Robert Calfee (1979). "Attention in the Classroom." In Hale, Gordon A. and Michael Lewis, eds. Attention and Cognitive Development. New York: Plenum Press, 297-330.

Rockstroh, Sybille and Karl Schweizer (2001). "The Contributions of Memory and Attention Processes to Cognitive Abilities." The Journal of General Psychology 128(1): 30-42.

Ruff, Holly Alliger and Mayr Klevjord Rothbart (1996). "Attention in Learning and Performance." In Ruff, Holly Alliger and Mayr Klevjord Rothbart, eds. Attention in Early Development: Themes and Variations. New York: Oxford University Press, 155-173.

Sobel, David (1990). "A Place in the World: Adults' Memories of Childhood's Special Places." Children's Environments Quarterly 7(4): 5-12.

Tabachnick, B.G. and L.S. Fidell (1996). Using Multivariate Statistics. New York: Harper Collins College Publishers.

Wells, Adrian and Gerald Matthews (1994). "Attention: Basic Conceptual and Theoretical Issues." In Wells, Adrian and Gerald Matthews, eds. Attention and Emotion: A Clinical Perspective. East Sussex: Lawrence Erlbaum Associates Ltd., Publishers, 19-44.

Wells, Nancy M. (2000). "At Home with Nature: Effects of 'Greenness' on Children's Cognitive Functioning." Environment and Behavior 32(6): 775-795.

Williams, Kathryn J. (1998). The Enchanted Forest: Transcendent Experience in Forest Environments. Psychology. Monash University, Melbourne.