Development and construct validation of a questionnaire for measuring affordances for motor behavior of schoolchildren

Environments where children move about provide affordances that play a signi�cant role in their development. This research presents the Affordances for Motor Behavior of Schoolchildren (AMBS) assessment tool, which aims to assess the interdependent systems, such as home, school, and sports activities, that can in�uence 6 to 10-year-old children's motor development, motor learning, and motor competence. After establishing face validity, 259 south Brazilian families completed the questionnaire. We assessed construct validity on the whole sample with Con�rmatory Factor Analysis. The model testing showed a very good �t, and the structural model presented signi�cative loading coe�cients from the identi�ed variables to the theoretically speci�ed latent variables (factors). Signi�cant correlation values were found between factors: Home and Materials (r = 0.77), Home and School (r = 0.41), and Materials and School (r = 0.56). Our results suggest that the AMBS can assess the opportunities for action provided to children by their home, the materials in it, and their school.


Introduction
Children's motor development is affected by different levels of environmental in uences ranging from proximal (immediate) to distal ones.Bronfenbrenner (1979) considered these ecological settings as interdependent systems, where each one ts inside the other, and termed them from a micro to a macro level, starting from the child and family environment to the cultural in uences speci c of different countries.The rst level of in uence is the microsystem, which corresponds to the child's immediate surroundings (e.g., home, neighborhood, daycare center, school, sports environments), where proximal processes occur.These proximal interactional processes of development relate also directly to the Gibsonian ecological approach.To Gibson (1979), each environment has materials, spaces, surfaces, actions, events, and people that provide the child possibilities for action (i.e., affordances), according to his or her action capabilities.Consequently, the perception of the environment and its features guides the child's movements and, reciprocally, action facilitates the detection of those features, specifying other affordances (Adolph & Hoch, 2019;Flôres et al., 2019;Heft, 2012).These affordances exist on the child's microsystems, but they are affected by the other layers of environmental in uence.For example, culture (as a macrosystem) in uences the parental practices and perceptions about what children need, which constrain, among other things, the type of toys that parents buy (Bradley & Corwyn, 2005).
As children grow, their immediate surroundings expand and the interactional activity in and with those environments progressively becomes more complex (Bronfenbrenner & Evans, 2000;Bronfenbrenner & Morris, 2006).After entering primary school, children begin regularly to attend places other than home.
The major focus of research has been in infancy and early childhood and in the home setting, as we can notice when analyzing the different tools that have been developed to evaluate the affordances present children's environments, namely the HOME (Bradley & Caldwell, 1984), the AHEMD-SR (Rodrigues et al., 2005), the AHEMD-IS (Caçola et al., 2011), and the AHEMD-IS to daycare setting (Müller et al., 2017).
Unlike the HOME inventory, which is not speci cally focused on affordances for motor development, the AHEMD-SR and the AHEMD-IS use a comprehensive inventory to assess the existent physical affordances that are conducive to enhancing motor development in the home environment.Some other tools were developed to analyze separate contexts of the child's life.The System for Observing Fitness Instruction Time (SOFIT), which was designed to assess different variables associated with children's activity levels and opportunities to become physically t in physical education (McKenzie et al., 1992;Pope et al., 2002); the Nutrition Physical Activity Self-Assessment for Child Care (NAPSACC), which aimed to analyze the school environment about the physical structure, food and educational policies (Ward et al., 2008); and the Physical Activity Neighborhood Environment Scale (PANES), which was developed to assess the neighborhood environment walkability and recreation facilities, related to the support physical activity for children (Sallis et al., 2010).
Although the above-mentioned tools cannot capture all affordances that exist in children's microsystems, they have proved to be valuable to assess the opportunities for action offered by different environments.As Gibson stated, the richest and most elaborate affordances are provided by other people (Gibson, 1979) and this type of affordances (i.e., social affordances) is di cult to be captured by inventory-type tools.In fact, to capture all the affordances that exist in a child's microsystems is a massive task, because affordances depend not only on the features of the environment but also on the child's characteristics, so they are dynamic, they might arise and dissolve over time and along with development.Thus, our purpose was to build a tool capable of capturing a speci c part of the affordances that exist in different children's microsystems (i.e., those provided by the physical environment) as detected by the children's families.
As far as we know, there is a lack in the literature related to an assessment tool capable of analyzing the 6-to 10-year-old children's physical affordances in different microsystems.Thus, our work represents an extension of previous tools (Rodrigues et al., 2005;Caçola et al., 2011) to an older age group contextual settings, aiming to create and establish the validity of a parental self-reporting research questionnaire, named Affordances for Motor Behavior of Schoolchildren (AMBS), which assesses the quality and quantity of the physical affordances in the 6-to 10-year-old children microsystems.Speci cally, we intend to validate the AMBS as a reliable tool to analyze the home (and its materials) and school environments, with the purpose of better understanding the potential of those environments for optimizing children's motor behavior.

Participants
Pilot study -For the initial construction of the AMBS, there was the need to pilot a rst version of the inventory.A convenience sample of 20 south Brazilian families was recruited using the University website and Facebook page.These 20 families represented a variety of socioeconomic, and educational levels.
Validation of the AMBS -Participants were recruited from twelve schools in south Brazil.Four hundred and fty-three (n = 453) families with children aged 6-to-10 years old (7.94 ± 1.23) were invited to participate in the study and two hundred and fty-nine (n = 259) consented to participate and completed the questionnaire (143 boys and 116 girls).This number of participants is adequate according to a rule of thumb of 10 cases per variable (Byrne, 2016;Nunnally & Bernstein, 1967), and accommodates a power of 0.89 for a model such as ours (Wolf et al., 2013).
All parents received a letter explaining the purpose of the study and asking for the signed informed consent for their participation.Parents answered through an online version of the AMBS, in which all the answers were sent directly to the cloud database.The research was approved by the university ethics committee.
A total of 259 parents answered the nal version of the AMBS, 72 (27.80%) were from families with children aged between 6 and 7 years, 56 (21.62%) were from families with children aged between 7 and 8 years, 67 (25.87%) were from families with children aged between 8 and 9 years, and 64 (24.71%) were from families with children aged between 9 and 10 years.Regarding the monthly income of families, 9.30% earned R$1000 or less (under $10000 per year), 19.30% earned between R$1001 and R$2000 (between $10001 and $15000 per year), 17.4% earned between R$2001 and R$3000 (between $15001 and $25000 per year), 8.50% earned between R$3001 and R$4000 (between $25001 and $35000 per year), 7.70% earned between R$4001 and R$5000 (between $35001 and $5000 per year), and 37.8% earned R$5001 or more ($50001 or more).Most parents that participated in the study (57.9% of fathers and 44.8% of mothers) had completed high school or less.Most children attended school in the afternoon (n = 132; 51,0%), 25.8% (n = 67) attended it in the morning, and 23.2% (n = 60) attended full-time school.
Regarding the type of housing, 59.1% of the children (n = 153) lived in houses and 40.9% (n = 106) lived in apartments.Table 1 shows the data regarding extracurricular activities that children attend to.Twenty families were asked to complete the AMBS twice with a 1-week interval to establish reliability.(2.7%)

Procedures
The present study was developed in separate phases, concerning the initial development of the assessment tool, and the construct validity of the AMBS.

Initial Development of the AMBS
Stemming from the previous work that has focused on the in uence of speci c environments on children's development (Abbott & Bartlett, 1999;Bradley et al., 2000;Gabbard et al., 2008;Rodrigues et al., 2005;Sallis et al., 2010), and grounded on the assumption that the child's development is in uenced by multiple contexts or systems, particularly as children grow older (Bronfenbrenner, 1995;Lerner, 2006), an extensive literature search was undertaken (Flôres et al., 2019).This review of literature aimed at characterizing the availability of affordances for motor development in the microsystem's settings that 6to 10-year-old children frequently attend (i.e., home, school, and leisure environments).As a result, an initial version of the AMBS was created in Portuguese, as a parent self-report assessment tool with seven sections, and 111 questions or items.
To assess content validity, the questionnaire was sent to three experts in child motor development (agreed participation) for critical review to assess the main sections and questions.The experts (two Portuguese and one Brazilian) were instructed to express their opinion regarding the wording and the relevance of the questions, recommending changes in questions and sections when needed.Experts' suggestions for change in the formulation of the questions were included, and only questions and sections positively rated by all experts were considered in the next version of the AMBS.Suggested new questions or sections made by any of the experts were again rated by the other experts on a second and nal round.
After the contribution of the experts, one section was removed (i.e., general data), one was added (i.e., characterization of house items), and two were renamed (i.e., house and house items to home characterization; and sports environments to extracurricular activities).Likewise, of the 111 questions in the initial version, 45 were deleted (e.g., parents' general data, number of garages, number of classmates, number of dance rooms, Pilates classes, etc.), and seven new ones were added (e.g., number of cars or motorcycles, number of computers, etc.) resulting on a revised version of the AMBS with 73 questions grouped in 7 sections.
This AMBS revised version was then piloted to assess face validity with 20 south Brazilian families, representing a variety of socioeconomic, and educational levels.Parents or guardians were asked to answer the questionnaire while pointing out di culties or making suggestions for corrections.They were also asked about other opportunities for action that could be missing in the present form of the AMBS.All items were classi ed as having a good comprehension by all parents, but due to the suggestions made in this pilot study, the authors re ned the wording of 2 items and changed 1 picture in the questionnaire.
The nal version of the AMBS questionnaire consists of 73 questions organized in seven sections: Child Characteristics (7 questions), Extracurricular Activities (6 questions), Family characterization (7 questions), Home characterization (10 questions), House Items (7 questions), Child Play Materials (22 questions), and School (14 questions).The questionnaire contains dichotomous questions, Likert-type scales, and description-based queries (used to collect general information to characterize the sample and participants).When relevant, gures were used in Likert scales to illustrate the alternatives (e.g., type of play materials) (see http://www.questionarioambs.com/ambs20/calculadora/).
An AMBS application (App) for Android, IOS and computer, was built to allow digitally answering the questionnaire.All answers of the submitted questionnaires on the different platforms are saved on an online database.
A rst exploratory analysis of the 259 families' data aimed to aggregate the items (questions) into meaningful variables that could represent speci c types of motor affordances.Forty-eight of the 73 questions were grouped according to common content into 11 variables grouped in three categories or factors and representing the characteristics of the school and home environments that children attend.The remaining questions (25) were used to collect information regarding the child characterization (7 questions), extracurricular activities (6 questions), family characterization (7 questions), and school characterization (4 questions).The correlation matrix for the 48 items was inspected for consistency and discrimination properties according to intended common content.The allocation of the items into group variables was changed if it was not positively correlated with the other questions within the variable or if it had a higher relationship to other variable questions.As a result of this process, three questions were reallocated to a different variable.The 11 variables were then tentatively associated with one of three broad categories or factors (Home, Materials, and School) according to a derived theoretical model for motor behavior affordances of school children (see Table 2).Reliability was established by the analysis of the intraclass correlation coe cient (ICC) for the variables.The results ranged from 0.79 to 0.90, providing support for use of the self-report version of AMBS as a reliable tool (see Table 3).

Data Analysis
The structural validity of this theoretical model of the AMBS was tested by robust con rmatory factor analysis (CFA) using the maximum likelihood estimation method, performed in AMOS software version 24.0.According to the theoretical argument, variables' unique loadings into latent factors (categories), and full correlational paths between latent factors were speci ed.The model goodness of t was assessed by the index Chi-Square/df, the Comparative Fit Index (CFI), Tucker Lewis Index (TLI), and the Root Mean Square Error of Approximation (RMSEA).Values of Chi-Square/df < 2, CFI and TLI ≥ 0.90, and an RMSEA ≤ 0.06 are considered indicative of good model t (Brown, 2006;Hu & Bentler, 2009).Modi cation indices (MI) were analyzed, and variables were considered for modi cation from their initial path to another factor, or for deletion when MI suggested that such procedure resulted in a signi cant improvement of the model t within the theoretical framework.

Results
Descriptive values for all items and variables are shown in Table 2.As seen in Fig. 1

Discussion
Our starting premise was that different immediate contexts (or microsystems) (Bronfenbrenner, 1995;Bronfenbrenner & Ceci, 1993) and their objects, places, surfaces, events, and other people (Gibson, 1979) can invite, permit, or inhibit progressively more complex child-environment interactions.These opportunities for action, or affordances, are intimately tied to the features of the environment, but also to the child's action capabilities, which change along with the development.Ward (1978, p.85) points out that "children will play everywhere and with anything" during their childhood, and "they will play wherever they happen to be", thus a city (and all their available microsystems) needs to make the whole environment accessible to them, because "whether invited to or not, they are going to use the whole environment".Nowadays, these thoughts continue to be important in the study of the environmental contexts experienced by children.Thus, our goal was to create a parental self-reporting research assessment tool to assess the quality and quantity of the characteristics (affordances) in the home, school and extracurricular activities for children ranging from 6-to 10-years.
The current study revealed that the AMBS was able to detect a common structured organization of potential affordances in the children's microsystems comprising different groups of categories: Home, Materials, and School, representing a meaningful structure associated with the children's environment.This is a relevant nding because it is the rst time that a shared structure for motor affordances provision is detected between the home and the school environment, probably meaning that decisions and characteristics of the opportunities for children engaging in motor challenges are also bonded at a community level.The AMBS seems to have the potential to evaluate and discriminate among different community pro les, according to the theoretically-driven organization for motor affordances.The hypothetical theoretical con guration seems to be promising in representing movement affordances available for children in their microsystems, resulting in the validation of the AMBS construct.
The AMBS intends to detect affordances for movement in the different settings.These affordances detections are mediated by the family's intentional evaluation of the conditions and materials according to the knowledge of their child's motor competence (Silva et al., 2017).In fact, it is this working ecological t between the existing settings or materials and the child's motor behavior that allows identifying the motor affordances as signi cant ones.According to our results, affordances perceived by families in the home, school, and the materials provided to children are related to each other (see Fig. 1).The strongest correlation was found between Home (its physical characteristics) and the available Materials as it would be expected, but there is a moderate and signi cant relationship between School and Materials, and School and Home.This relationship probably shows a community-linked association regarding the organization of spaces and the provision of movement affordances, both in-home and school settings, but also can be indicative of how families are stable when detecting movement opportunities for their children.The fact that the variables-factor coe cients for School are high but not perfect, shows that affordances provided in the school differ from school to school, but they can also differ within the same school depending on the perception of the physical settings as meaningful affordances for the child.
Within the Materials category, all the different variables show a high and positive loading into the latent factor, showing that the different materials are provided within the home family.Nevertheless, educational toys and manipulative materials showed the strongest loading values indicating that these speci c types of materials are the ones more present in all the houses, probably because parents feel that affordances that are conducive to stimulating ne motor skills are particularly important for school-aged children.Also, for the School factor, the relevance of Space for Movement is lowest than the other two, meaning that this characteristic varies more from school to school.

Potential Use Of The Ambs
The prospective use of the AMBS in children's motor behavior and development studies seems promising, particularly because although previous studies have assessed the importance for motor development of the different contexts that children attend, usually, only a setting is studied at a time.Miquelote, Santos, Caçola, Montebelo, and Gabbard (2012) found that home provides great resources to promote motor and cognitive skills in young children, and it is associated with gross and ne motor performance.Despite the positive results, the authors do not report the importance of other environments experienced by children, such as the home of grandparents (or relatives in general) and daycare centers.Ferreira et al. (2019) investigated the association between sport participation and motor competence in 6-to 10-year-old children.The results showed that sports participation was associated with motor competence and play a relevant role in this process.Again, the authors cannot explain the role of other environments, such as home or school, in the development of motor competence.Despite that, most of the research that studied the school period focused on physical activity during recess, playgrounds, or physical education classes, not taking into account the materials, people, and other affordances provided (Dowda et al., 2005;Frago-Calvo et al., 2017;Mills & Burnett, 2017).Goyen and Lui (2009) aimed to determine the prevalence of developmental coordination disorder (DCD) in ''apparently normal'' extremely premature or extremely low birthweight schoolchildren at 8 years of age.The results showed that "apparently normal" infants are at risk of motor dysfunction in their school years.Nevertheless, there was not an assessment of the environments and contexts experienced by these children.Several other research follows the same pattern (Flôres et al., 2015;Goyen & Lui, 2002;Herrmann, Heim, et al., 2019;Herrmann, Seelig, et al., 2019;Logan et al., 2014).Thus, a more contextualized perspective can provide investigations beyond the child's immediate context to examining the in uence of wider environments and the opportunities offered for the children (Gabbard & Krebs, 2012;Nobre et al., 2020).
The AMBS shows the potential to evaluate and discriminate affordances among different microsystems that children attend to.Future studies that use both the AMBS and a validated assessment of motor development or motor competence can help us to increase our understanding of the relationship between children's regular microsystems and motor behavior.Furthermore, our results suggest that AMBS is a reliable and valid tool to be used as an inventory for assessing affordances in the proximal settings, with clinical and research applications.
AMBS future research should examine the different contexts that children attend to, and further relate them to their levels of motor learning, motor competence, and motor development.That is, relating AMBS results to motor competence or motor development assessments (e.g., Motor Competence Assessment, Test of Gross Motor Development − 2) should provide a better understanding of how children are developing, learning, and acquiring lifespan motor competence within the environment.Information from the AMBS may be particularly bene cial to supplement motor information for children with lower levels of motor competence since this assessment tool might help to identify areas of strength and weakness in the environments the child attends to, which can help to develop appropriate interventive strategies.

Limitations
While our results suggest that the AMBS can be a valuable tool to inform about the opportunities for action provided to children by their home, the materials in the home, and their school, it still has some limitations.First, all families participating in this study were from southern Brazil, so the country's cultural diversity might not be fully represented.Second, the AMBS is answered by parents (self-report), which means that the physical environments could not be veri ed.This fact implies that the responses provided are related to parental perception and, sometimes, can have a discrepancy between the real environment and the perceived (Silva et al., 2017).Third, although an effort was made to assess the affordances in different microsystems, it is quite di cult to have an assessment tool that evaluates all the environments the child attends.For example, in the AMBS, the extracurricular activities are only identi ed (see Table 1), but no detailed assessment is made regarding the opportunities for action that exist in those environments.Even within each microsystem, the full range of affordances is di cult to capture.For instance, social affordances that emerge from the interaction with other people are not captured by the AMBS.To minimize this limitation, the AMBS also collects information about the family structure (e.g., number of children and adults in the household).Finally, although the AMBS assesses the opportunities for action provided by different environments, it does not allow for assessing the child's real interaction within those environments, since children might not always use the affordances that are available to them.

Figures
, the diagram of the structural model included paths from Home to the Inside Spaces and Outside space; to the Materials to Sedentary materials, Pretend play toys, Educational toys, Manipulative materials, and Stability materials; and to the School to the Space for movement, Free time for movements, and Sedentary free time.The testing of this AMBS model returned a Chi-Square of 91.24; p < .000,showing a very good t to data, as assessed by the index Chi-Square/df = 2.225, CFI = .956;TLI = .942and RMSEA = .059.The structural model showed signi cant loading coe cients ranging from 0.51 to 0.89 from the identi ed variables to the theoretically speci ed latent factors.Furthermore, signi cant correlation values were found between factors: Home and Materials (r = 0.77), Home and School (r = 0.41), and Materials and School (r = 0.56).Analysis of the modi cation indices did not result in any change of the model speci cation that could enhance model t, showing a good adherence of the theoretical model to the real data.

Table 1
Data regarding extracurricular activities that children attend to.

Table 2
Descriptive values of the items and variables.