Background

A small proportion of Austrian children aged 11–13 years (27.2 % boys; 17.1 % girls) meet the general recommendations of 1 hour of moderate-to-vigorous physical activity (MVPA) per day (Felder-Puig et al. 2019; Fonds Gesundes Österreich 2020). Increasing active transport to school (ATS) is an effective health-promoting intervention to integrate physical activity into the everyday lives of children and adolescents (Villa-González et al. 2018; Denstel et al. 2015).

Approaches to promoting ATS are diverse, and focus on modifying the built environment or deploying measures to raise awareness and change behaviour among children and parents. Associations between the natural or built environment and physical activity have been presented in the literature (Masoumi 2017; Sallis et al. 2016; Smith et al. 2017). Previous studies have also shown an association between environmental characteristics and ATS (Rahman et al. 2020; Oliver et al. 2014; Murtagh et al. 2016), where children living in urban areas or having lower distances from home to school were more likely to engage in ATS (Stewart 2011).

In order to understand children's ATS, individual parameters, such as parental attitude toward ATS and control over mode of school travel or children´s motivation, should be taken into account, in addition to environmental aspects. Parental attitude, for example, has been positively associated with ATS (Ross et al. 2018). Children using active transport had parents who reported using active transport to work or to other locations or being more physically active in general (Henne et al. 2014; Tanaka et al. 2018; Garriguet et al. 2017).

Murtagh et al. (2012) tested the association between children’s relationship to their ATS and the following psychosocial constructs based on the Theory of Planned Behaviour provided by Fishbein and Ajzen (2010): behavioural intention, attitude, subjective norm, and perceived behavioural control. They showed that perceived behavioural control for walking and the car/bus habit predicted intention, while intention of car/bus habit predicted behaviour for ATS (Murtagh et al. 2012). According to another study from Murtagh (Murtagh et al. 2016) children in Ireland were more likely to engage in or continue ATS if they lived in an urban area. This can be explained by the lack of walking and cycling infrastructure, deficient street connectivity, and long commuting distances in rural areas (Murtagh et al. 2016; Hofer-Fischanger et al. 2020; Veitch et al. 2017; Pocock et al. 2019).

Most of this available evidence explores the association with behavioural or built environment factors in urban areas, but evidence on associations in rural areas is rare. As two-thirds of Austrians live in rural areas (Statistik Austria 2019), many children grow up and go to school in rural regions. Active transport behaviour seems to be different among people living in rural areas and those living in urban areas, with people living in rural areas engaging less in active transport (Carlson et al. 2018; van Dyck et al. 2011). A recently published study from Germany showed that passive modes of transport (e.g., cars) were more often used in small- or medium-sized towns (Reimers et al. 2020).

The aim of this study was to explore the association between psychosocial and environmental determinants and ATS among children living in two rural communities in Styria, Austria.

Method

Study setting

This cross-sectional study was conducted in two rural communities in Styria (Austria) in April 2017. Town A (St. Margarethen an der Raab) consists of 4,090 inhabitants, while town B (Grafendorf bei Hartberg) has 3,080 inhabitants. With about 80 inhabitants per km2, this region can be classified as very rural, in accordance with the OECD’s criterion of a population density lower than 150 inhabitants per km2 (OECD 2007). A built environment assessment of the two rural towns was conducted in 2017 and showed that sidewalks and cycling routes are rare, narrow, and often not well maintained (Hofer-Fischanger et al. 2021). In addition, the landscape is hilly, public transport options are limited, and the school bus runs three times a day (in the morning, at noon, in the afternoon). Most of the children live outside the town centre (up to 10 km), and some of the parents commute to the city of Graz for work, which is a drive of about 30 minutes (town A) or 50 minutes (town B) (Hofer-Fischanger et al. 2021).

Study design and procedure

A self-reported questionnaire was distributed to all children (n = 569) attending four schools (two primary schools and two secondary schools) and their parents. The paper–pencil questionnaire was packed in a sealed envelope together with some supplements (letter to parents, description of the survey and targets, description of the procedure and instructions for filling in the questionnaire, an informed consent, and a return envelope). One parent (mother or father) was asked to complete the questionnaire together with the child. Parents were asked to place the completed questionnaire in the return envelope and give it to the child to take to school within 1 week. Classroom teachers collected the returned envelopes with the completed questionnaires, which were picked up by project staff at each school after 1 week. Ethical approval for the study was granted by the ethics commission of the Medical University of Graz (29-463 ex 16/17).

Measures

The survey was developed on the basis of two questionnaires, the validated ALPHA (Spittaels et al. 2010) for measuring physical-activity-related environmental factors and an established scale for measuring planned behaviour based on the Theory of Planned Behaviour (TPB) (Fishbein and Ajzen 2010), which was used in a previous ATS study (Murtagh et al. 2012). After the questionnaire was completed, we conducted pre-tests with 18 participants to increase the survey’s validity and reliability (Chronbach´s alpha reported in the following sections), and then prepared a final version of the questionnaire. The final questionnaire consisted of two sections, one for parents and one for children. Parents were asked to answer questions on the constructs of the TPB, environmental factors, and personal/family factors, while children were asked to answer (alone or together with parents) only the questions on planned behaviour and personal factors. Personal and family factors were measured as described in the following section, while TPB and environmental items were measured on a 4-point scale: 1 (strongly disagree), 2 (disagree), 3 (agree) or 4 (strongly agree).

Personal and family factors (covariates)

The children attended primary school or secondary school and were between 6 to 15 years old. Children were categorized according to age and school years: 6–8 years (1st and 2nd class), 9–10 years (3rd and 4th class), 11–12 years (1st and 2nd class secondary school), 13–15 years (3rd and 4th class secondary school). The educational level of parents was categorized according to the International Standard Classification of Education (UNESCO 2003) as follows: low (primary school or lower secondary school without vocational training), medium (vocational training, upper secondary school, or professional school), high (any higher level, e.g., tertiary education). The questionnaire further included questions about the child´s and parent´s gender, number of people in the household and the number of cars in the household.

Active transport characteristics

Active transport was measured as the number of times per week walking (also walking to a bus) or cycling to school or from school (also from a bus), as suggested by Murtagh et al. (2016). As in Dalton et al. (2011), parents were asked, “How does your child usually travel to and from school, on how many days per week?” Possible responses were walking, cycling, using bus, walking to the bus, traveling by car or “other”. “Walking to bus” was included because we assumed ATS to be low in the study setting, due to long distances and a hilly landscape. Participants were categorized as active if they “walked”, “cycled” or “walked to bus” a minimum of 5 times per week on their way to/from school, which is on average at least once a day. Other variables measured were distance from home to school (categorized as: less than 500 meters, 500 m to under 1 km, 1 km to under 2 km, 2 km to under 3 km, 3 km and more) and duration for walking from home to school/bus station (categorized as: less than 5 minutes, 5 to 10 minutes, 11 to 20 minutes, more than 20 minutes).

Environmental factors

The number of items from ALPHA was reduced to fit the school environment. Additionally, items including the formulation of “neighbourhood” or “living environment” were changed to “school environment”. For this study, five themes (infrastructure, maintenance of infrastructure, neighbourhood safety, aesthetics, walking and cycling network), including 11 original items and four slightly changed items, were used.

Walking or cycling infrastructure was measured with one item: “There are sidewalks in my school environment”. Maintenance of infrastructure was measured using one item: “The sidewalks in the school environment are well maintained”. Neighbourhood safety was measured with four items: “There are not enough safe places to cross busy streets in my school environment”, “Walking is dangerous because of the traffic in my school environment”, “Cycling is dangerous because of the traffic in my school environment”, “It is dangerous in my school environment during the day because of the level of crime”. The mean of the participants´ scores on these four items served as the overall measure of neighbourhood safety for use in the subsequent data analysis (α = 0.82). Aesthetics of school environment were measured with four items: “My school environment is a pleasant environment for walking and cycling”, “There are trees along the streets in my school environment”, “There are seating possibilities in my school environment” (changed from “There is litter or graffiti in the streets of my neighbourhood”) and “There is much interesting to see in the school environment” (changed from “In my neighbourhood there are badly maintained, unoccupied, or ugly buildings”). The mean of these four items served as the overall measure of attitude for use in the subsequent analyses (α = 0.79). The cycling and walking network was measured with the following item: “There are many shortcuts for walking in my neighbourhood”.

Psychosocial characteristics

Psychosocial characteristics, which were all theory of planned behaviour constructs, were used based on previous research on active school travel, as described in Murtagh et al. (2012). Cronbach alphas were calculated to assess the internal consistency of the constructs. Intention was measured using two items. The mean of participants´ scores on these two items served as the overall measure of intention for use in the subsequent analysis (α = 0.97 parents; α = 0.92 children). Attitude was measured with four items (α = 0.90 parents; α = 0.86 children), subjective norm with six items (α = 0.78 parents; α = 0.74 children) and perceived behavioural control was measured with three items (α = 0.84 parents; α = 0.90 children).

Statistical analysis

First, we tested for possible clustering of our outcome ATS within schools, using multilevel logistic regression, and did not find any clustering. Subsequently, the association between ATS and psychosocial characteristics (child and parent) and environmental factors was calculated using single-level logistic regression models, with active mode or non-active mode of transport to/from school as the dependent variable. Models were adjusted for variables that might influence the associations: age (child), gender (child), educational level (parent) and number of cars in the household. Psychosocial characteristics and environmental determinants that were significantly associated with the outcome variable (p < 0.05) were included in the hierarchical multiple regression models, predicting the odds of active or non-active mode for walking or cycling to/from school (calculated stepwise). In model 1, the independent variables were psychosocial characteristics of parents (intention, attitude, subjective norm, and perceived behavioural control). Psychosocial characteristics and behaviour of children were entered as additional independent variables in model 2. In model 3, walking and cycling network as an environmental factor and distance and duration were added. Odds ratios (OR) and confidence intervals (CI) are reported. The regression coefficients are displayed with their 95% confidence interval (95% CI). All analyses were performed using IBM SPSS Statistics for Windows, Version 26.

Results

Sample characteristics

In total, 467 questionnaires (response rate: 82.07%) were returned from schools, and 382 were included in the analysis (85 were excluded due to missing values on psychosocial, environmental or active transport characteristics), of which 186 were active (they walked, cycled, or walked to bus a minimum of 5 times per week). Table 1 shows sample characteristics. Most participants were female, especially among the parents, and 77% of the participants had two or more cars in the household. Table 2 shows that about half of the participants engage in some kind of ATS. Only about one third of the participants live within 2 km or less. Most of the children commuted to school using the bus (58.6%) and walking to the bus (33.5%) and some went by car (21.5%), while walking (14.4%) and cycling (2.6%) were used less frequently.

Table 1 Personal factors of the study population
Full size table
Table 2 ATS characteristics of the study population
Full size table

Results bivariate analysis

Table 3 shows the associations between ATS and psychosocial and environmental factors from bivariate analyses. Children of parents with more points on the psychosocial characteristics (ORs 1.5 to 1.9) and with more points on their own psychosocial characteristics (ORs 2.0 to 2.4) were more likely to engage in ATS. With each point increase in intention, attitude, subjective norm, or perceived behavioural control, the probability of engaging in ATS doubled. Most environmental factors (infrastructure, maintenance of infrastructure, neighbourhood safety, or aesthetics) show no statistically significant association with ATS. However, with each point increase in reported walking and cycling network, children’s probability of engaging in ATS increased by 44%. No statistically significant association was found between ATS and age and gender of the children, the educational level of parents, or the number of cars in the household.

Table 3 Results of bivariate analysis for psychosocial, environmental and other active transport characteristics
Full size table

Results of multivariate analysis

Table 4 presents best-fit models for predicting ATS. The probability of children engaging in ATS increases with each unit increase in parents’ intention (OR 1.6, p < 0.001), parents’ perceived behavioural control (OR 1.3; p < 0.05), and children’s perceived behavioural control (OR 1.8, p < 0.001). Children’s perceived behavioural control was the strongest predictor for ATS on the individual level. Children who walked between 5 to 10 minutes to school or to the bus were 4.9 times (p < 0.001) more likely to engage in ATS than those who walk more than 20 minutes.

Table 4 Results of multiple regression models
Full size table

Discussion

This study examined determinants associated with active commuting to school in rural communities, including psychosocial, environmental, and other influencing determinants on active transport. The results of bivariate analyses showed significant associations between ATS and psychosocial characteristics of parents and children (intention, attitude, social norm and perceived behavioural control), one environmental item (walking and cycling network), and with distance and duration of walking. However, in the multiple regression model, only the parents’ intention, children’s perceived behavioural control, and duration of walking were significantly associated with ATS. Odds for active school transport were higher among children who had higher perceived behavioural control, walked between 5 to 10 minutes to school, and whose parents had a higher intention for active school transport.

While the association between built environment and levels of active transport has been shown for urban areas (Evenson et al. 2006; Oliver et al. 2014), the current study provides little empirical evidence for this association in rural areas. “Walking and cycling network” in the sense of having many shortcuts for walking and cycling in the school environment was the only environmental variable associated with active commuting to school in bivariate analyses. We also found that the environmental characteristics derived from ALPHA (Spittaels et al. 2010) were automatically eliminated when distance and duration were incorporated into the model. It is noticeable that children with a walking time of 5 to 10 minutes to school/bus have a higher likelihood of active transport than children with a walking time of less than 5 minutes. We explain the result as follows: the short walking time (< 5 minutes) indicates that this category probably includes many children who walk to the bus.

The results of our study show that children whose journeys to school are shorter in either distance or duration are more likely to choose active transportation, which fits with the current evidence in the literature, where distance to school has been found to be the strongest predictor for active commuting to school (Carver et al. 2011; Porskamp et al. 2019; Molina-García et al. 2020). Previous studies showed a lower likelihood of choosing an active mode of transport if the target location is not within 800 m of walking distance (Pocock et al. 2019; Veitch et al. 2017). In our study population, this is seen in the relatively low percentage of children walking (14.4%) or cycling (2.6%) to school, as only about 10% lived within a distance < 1 km to school. In a study conducted in 2007 in Norfolk (UK), where about half of the participating children resided within 1600 m of their schools, 62% walked and 12% cycled. However, in that study, only 20% of children had a home address classified as rural (Carver et al. 2014). Previous studies have also shown a lower proportion of people travelling on foot in rural areas (22.1%) than in urban areas (49.5%) (Panter et al. 2008). However, comparisons between countries are complex because of significant differences in landscape environment and infrastructural determinants.

In addition to the abovementioned distance to school, four other criteria influence travel mode to school, according to Mitra (2013): 1) safety concerns, 2) street connectivity, 3) pedestrian facilities and aesthetics, and 4) social connectedness. Street connectivity and pedestrian facilities and aesthetics mainly describe the infrastructure for active transport. Previous studies have found a lack of safe and maintained sidewalks and cycling routes or few connections between streets outside the town centres as impeding determinants on active transport in rural areas (Hofer-Fischanger et al. 2020; Carver et al. 2014; Frost et al. 2010; Hansen et al. 2015).

However, little research has been conducted on the distance to/from the bus and the existing infrastructure for walking along the way. The results of our study suggest that in rural areas, the infrastructure conditions in the immediate vicinity of children's homes, for example the availability of safe walking routes to bus stops or weather-protected bus stops, are a key factors that influence walking. Furthermore, children´s long travel-times on rural bus routes (Hansen et al. 2015) and early arrival times at school hinder engaging in school bus transport and so negatively influence walking to the bus.

Regarding safety concerns and social connectedness, studies (Panter et al. 2008; Porskamp et al. 2019; Faulkner et al. 2009) have shown that family members and friends influence the travel mode to school (parental and social support, mobility choices and restrictions). Especially parents are considered as ‘gatekeepers’, who control their children´s mobility outside the home more than ever before (Carver et al. 2014). Researchers describe parents today as becoming more protective due to a decrease in overall social trust (Porskamp et al. 2019) and the fast-moving traffic on rural streets (Hansen et al. 2015). Thus, very young children are only allowed to walk to school when the school is located close to home or if a parent or other guardian accompanies them to school (Faulkner et al. 2009).

Beside this, previous studies conducted in more urban and suburban areas have found clear correlations between the number of cars available in the household and the choice of an active mode of transport (Panter et al. 2008; Fishman et al. 2015; Dieleman et al. 2002). This correlation was not present in the current study, possibly due to the generally high number of cars available in the rural households (3/4 of study participants had two or more cars). However, further investigation is needed to determine whether the number of cars per household is associated with ATS in rural areas.

Murtagh and colleagues recommended investigating the influence of the psychosocial characteristics of children living in rural areas, as the published data were based on an urban sample (Murtagh et al. 2012). This was done in the present study. All psychosocial items, based on parents’ and childrens’ TPB (intention, attitude, social norm, and perceived behavioural control) showed significant results in bivariate analysis. Furthermore, the intention of parents and perceived behavioural control of children showed significant influences on ATS in the multiple regression model. We found that children´s perceived behavioural control might be of greater importance than the perceived behavioural control of their parents, as perceived behavioural control of parents did not show significant results after the inclusion of the children’s psychosocial items. Murtagh and colleagues also identified children’s perceived behavioural control as an independent predictor of children´s intentions to actively travel to school (Murtagh et al. 2012). According to the multiple regression, neither attitude nor subjective norm seem to be a significant influencing factor on ATS in our study.

Since all psychosocial items of parents and children showed a significant association with ATS, these parameters are particularly important for further research and practice. Future initiatives should focus on raising awareness to strengthening control beliefs of parents and children and should emphasise the importance of active transport as an opportunity for physical activity in everyday life. There is a need to reinforce children's planned behaviour of walking or cycling to school already in the first year of school or even before in kindergarten. When barriers, such as lack of time or feelings of inability about ATS, arise, children need skills to overcome these issues. For example, if the route to school is hilly, parents and/or teachers could suggest e-bike possibilities for children to manage hills more easily. Programs should also target working with teachers, parents, and the further social environment (grandparents, sisters and brothers), as studies showed the influence of family members and friends on ATS (Panter et al. 2008; Porskamp et al. 2019; Faulkner et al. 2009). Further research is needed to identify effective approaches for changing the predictors of ATS.

To sum up, for practical application and political actions, a safe and well-maintained walking and biking infrastructure, well-connected walking and biking routes, or schoolwide policies (Hollein et al. 2017) and programs for safe routes to school (Safe Routes to School Partnership 2021), including best opportunities for engaging in school bus transport (short travel-times for children, adequate arrival times at school, sheltered bus stops) or e-bike possibilities for children to manage hills more easily would promote ATS in rural communities, even if distances are longer. Programmes to promote ATS should strengthen the control beliefs of parents and children and should include parents in order to increase their intentions for ATS. In the future, researchers should investigate environmental and infrastructural determinants from home to bus stops in more detail and explore transport chains of children to and from school to gain more clarity concerning determinants of ATS in rural areas.

Strengths and limitations

The major strengths of this study include the involvement of schools located in less researched rural areas of Styria, Austria and the high participation rate (82%). Furthermore, both children and their parents participated in the study, which also combined environmental and psychosocial factors influencing ATS. The fact that we measured walking to the bus stop separately can also be seen as a strength, as some previously conducted studies did not include this active pathway in their design.

The limitations of this study include the fact that data were only collected during autumn. Due to seasonal variations in children´s school travel behaviour, the frequency of walking could be lower in autumn than in spring or summer. In addition, as there is no “gold standard” for the measurement of active transport, previous studies have measured ATS differently and sometimes imprecisely, e.g., “usually” walking or biking to/from school (Carver et al. 2014; Murtagh et al. 2016). We measured active transport in times per week walking or cycling to school/to the bus as reported in the questionnaire, since a more rigorous measurement was recommended for health researchers (Lu et al. 2014). Participants were categorized as active if they “walked”, “cycled” or “walked to bus” a minimum of 5 times per week on their way to/from school. This is rather high compared to other studies, some of which categorized participants as active when walking to school once a week (Dalton et al. 2011). Therefore, the effects on active transport may have been underestimated. Active transport could be assessed more objectively if it had been measured with accelerometers. Furthermore, due to possible variations in bus routes, the duration of walking to school can differ from the duration of walking from school, especially to or from bus. This may have influenced the results of the study.

Variables for active transport show that 58.6% of the children took the bus to school, but only 33.5% said they walk to the bus. The data does not show how the other 25% get to the bus. Here the authors suspect that participants did not fill in the questionnaire completely correctly and overlooked filling in the field "walking to the bus" when they filled in "using" bus. It is therefore even possible that the item "walking to the bus" was underestimated.

Psychosocial determinants were not analysed as described in previous studies (Murtagh et al. 2012; Zhang et al. 2020). The assessed parameters were chosen as independent variables predicting ATS. Furthermore, the items in the ALPHA questionnaire may not be ideal for measuring the environment in rural areas. By making changes to individual items, we have tried to adapt the questionnaire to the measurement needs in rural areas. However, we would recommend developing a separate questionnaire that better measures the built environment in rural areas. While previous studies have suggested that interventions that promote walking and cycling as enjoyable activities or together with friends may be a key factor to encourage walking and cycling as a modal choice (Veitch et al. 2017; Mitra 2013), we did not specify social or safety aspects regarding ATS in our work.

Conclusion

Intention, attitude, social norm, and perceived behavioural control of parents and children showed a significant association with ATS. Furthermore, a good walking and cycling network is related to more ATS. However, we assume that distance between home and school is the predominant determinant on ATS in rural areas. Therefore, future initiatives to encourage ATS in rural areas should focus on strengthening control beliefs of parents and children, and should include parents in order to increase parents´ intentions for ATS through awareness-raising measures. Additionally, research should explore the effects of active transport promotion in the context of public transport and school buses and should focus on the conditions of the infrastructure in the immediate vicinity of the children's homes and near bus stops. Above all, innovative programs to shorten distance (more collection points and school bus stops) or duration for walking and cycling to school (shorter bus routes, e-bike possibilities for children to manage hills more easily, sheltered bus stops) should be considered in spatial and community planning.