Journal of Geographical Sciences >
Exploring zonal heterogeneities of primary school students’ commutemode choices through a geographically weighted regression model
Wu Dawei (1999), PhD Candidate, specialized in transport geography and travel behaviour. Email: wdw808@bjtu.edu.cn 
Received date: 20230326
Accepted date: 20231103
Online published: 20240424
Supported by
National Natural Science Foundation of China(71971023)
Beijing Social Science Foundation(21DTR055)
Commuting is an important part of primary school students’ travel behavior, which has been concerned for a long time. We found that the commutemode choice behavior of primary school students in the context of regional segmentation shows strong characteristics in space, but has not yet been considered in traditional research. To fill this gap, this study summarizes the commutemode choices of primary school students with different characteristics based on the Beijing School Commute Survey. And the geographically weighted regression (GWR) model is built to analyze the zonal heterogeneity of the impact of personal characteristics, family factors and school factors on the primary school students’ commute mode choices from a lowcarbon perspective. The results demonstrate that the possibility of primary school students choosing lowcarbon commuting modes is positively correlated with the grade, commuting time, commuting escort type and housing category, but is inversely related with the commuting distance and the distance from the school to the city center. The coefficient estimates of explanatory variables vary across regions. Finally, we put forward policy suggestions regarding promoting the lowcarbon commuting structure, such as developing the urban slow traffic system, which serve as a reference for policymakers.
WU Dawei , MA Lu , YAN Xuedong . Exploring zonal heterogeneities of primary school students’ commutemode choices through a geographically weighted regression model[J]. Journal of Geographical Sciences, 2024 , 34(4) : 804 833 . DOI: 10.1007/s1144202422289
Figure 1 Map of central urban area of Beijing (a) and TAZs (b) 
Figure 2 Spatial distribution of schools and number of samples 
Table 1 Part of the results of the commuting survey among primary school students 
Object ID  Mode  School  Student  Family  

Name  Longitude  Latitude  Region  Grade  Commuting time  Commuting distance  Housing category  Commuting escort type  
498**  Walk  B**  116.36**  39.9**  Xicheng  Five  AM7:30  1.167 km  Selfowned  None 
39**  Bicycle  T**  116.32**  40.0**  Haidian  Six  AM7:15  2.709 km  Tenancy  None 
304**  Car  T**  116.32**  40.0**  Haidian  Six  AM7:30  3.454 km  Selfowned  Parents 
Note: To protect the privacy of students, object ID, name, longitude and latitude of the school are not fully displayed in this table. 
Figure 3 Low carbon characteristics of commuting modes of primary school students in different regions 
Figure 4 Spine chart of primary school students’ grade (a), commuting time (b) and the commuting mode choice (Note: 1 to 11 in the vertical coordinate represent walk, bicycle, twowheel electric bicycle, threewheel electric bicycle, bus, subway, school bus, shuttle bus, motorcycle, taxi and car, respectively.) 
Figure 5 Violin chart of primary school students’ escort type (a), housing category (b) and the commuting mode choice (Note: 1 to 11 in the vertical coordinate represent walk, bicycle, twowheel electric bicycle, threewheel electric bicycle, bus, subway, school bus, shuttle bus, motorcycle, taxi and car, respectively.) 
Figure 6 Distribution of commuting distance among primary school students 
Table 2 Description of the dependent variable and explanatory variable 
Type  Attribution  Name  Definition  Assignment  Category 

Dependent variable  Classified variable  Commuting mode  Primary school students’ commutemode choices  1  Walk 
2  Bicycle  
3  Twowheel electric bicycle  
4  Threewheel electric bicycle  
5  Bus  
6  Subway  
7  School bus  
8  Shuttle bus  
9  Motorcycle  
10  Taxi  
11  Car  
Explanatory variable  Classified variable  Grade  Grades of primary school students  4  Grade four 
5  Grade five  
6  Grade six  
Classified variable  Commuting time  Primary school students’ commuting time  2  Commuting time from 6:00 to 6:59  
3  Commuting time from 7:00 to 7:59  
4  Commuting time from 8:00 to 8:59  
5  Commuting time after 9:00  
Classified variable  Commuting distance  Primary school students’ commuting distance  [1, 20)  Commuting distance range from 1 to 20 km  
Classified variable  Distance from the city center  Distance from the city center to school  [1.6, 53.9]  Distance from the city center to school range from 1.6 to 53.9 km  
Classified variable  Commuting escort type  Primary school students’ commuting escort type  1  Parents  
2  None  
3  Elderly  
Classified variable  Housing category  Housing categories of primary school students’ families  1  Selfowned  
2  Tenancy  
3  Lodge 
Note: The center of Beijing is Tiananmen Square at coordinates (39.9087N, 116.3974E). 
Table 3 Comparison of fitting effect between OLS and GWR 
AICc  R^{2}  

OLS  112456.9  0.2733 
GWR  112234.8  0.4933 
Note: R^{2} of the GWR is the optimal value. 
Figure 7 Spatial distribution of standard residuals for the GWR model 
Table 4 Estimated Value of the GWR Model for the commuting mode of primary school students 
Variable  MEAN  MIN  LQ  MED  UQ  MAX 

Grade  0.0767  0.2053  0.0781  0.0732  0.0697  0.0026 
Distance from the city center  0.0423  0.0409  0.0355  0.0431  0.0504  0.0820 
Commuting time  1.4454  2.2218  1.4709  1.4489  1.4062  1.1439 
Commuting escort type  1.5022  1.7208  1.5189  1.4931  1.4809  0.8113 
Commuting distance  0.3081  0.2225  0.3057  0.3061  0.3076  0.3336 
Housing category  1.0908  1.2200  1.1242  1.0991  1.0803  0.0617 
Figure 8 Spatial distribution of average coefficients of the grade (a) and commuting distance (b) for the GWR model 
Figure 9 Spatial distribution of average coefficients of commuting time (a) and the distance from the city center (b) for the GWR model 
Figure 10 Spatial distribution of average estimation coefficients of commuting escort types (a) and housing categories (b) for the GWR model 
Table 5 The division and assignment of commuting modes based on the sample size 
Commuting mode  Category  Assignment 

Walk  Strong low carbon  1 
Bicycle  Slightly strong low carbon  2 
Twowheel electric bicycle  
Threewheel electric bicycle  Slightly weak low carbon  3 
Bus  
Subway  
School bus  
Motorcycle  
Shuttle bus  
Taxi  
Car  Weak low carbon  4 
Figure 11 Spatial distribution of average estimation coefficients of commuting escort types in classification methods based on the sample type (a) and the sample size (b) 
Table A1 VIF values of explanatory variables of OLS model 
Variable  Coef.  Pvalue  VIF 

Grade  0.057  0.047  1.021 
Distance from the city center  0.034  0.000  1.024 
Commuting time  1.311  0.000  1.135 
Commuting escort type  1.542  0.000  1.039 
Commuting distance  0.317  0.000  1.159 
Housing category  1.181  0.000  1.037 
Table A2 Results of spatial autocorrelation test on explanatory variables 
Variable  Moran’s I  Zscore  Pvalue 

Grade  0.094  103.306  0.000 
Distance from the city center  0.651  2520.812  0.000 
Commuting time  0.152  170.284  0.000 
Commuting escort type  0.044  49.300  0.000 
Commuting distance  0.177  197.302  0.000 
Housing category  0.287  320.557  0.000 
Table A3 The legend names and descriptions of figures 
Figure  Legend name  Description 

Figure 1  Map of the central urban area of Beijing and TAZs  Location of the central urban area of Beijing and the division results of TAZs 
Figure 2  The sample size  The sample size contained at each dot 
Figure 3  The commuting mode coefficients  The commuting mode coefficients of primary school students in different regions 
Figure 7  The standard residuals  The standard residuals for the GWR model 
Figure 8  The coefficients for the model  The coefficients of the grade and commuting distance for the model 
Figure 9  The coefficients for the model  The coefficients of the distance from the city center and commuting time for the model 
Figure 10  The coefficients for the model  The coefficients of commuting escort types and housing categories for the model 
Figure 11  The coefficients for the model  The coefficients of commuting escort types for the model 
[1] 

[2] 
Beijing Municipal Education Commission, 2022. Regulations of Beijing on Safety Management of Primary and Middle School Kindergarten for trial implementation. http://jw.beijing.gov.cn/jyzx/ztzl/bjjyxx/jyxxzcfg/202207/t20220706_2764772.html.

[3] 
Beijing Xiaoshengchu Network BXN, 2022. Basic information on the number of primary school graduates and the number of students in all grades in Beijing in the 20212022 academic year. https://www.xschu.com/xiaoshengchu/17/44263.html.

[4] 
Beijing Youshengxiao Network BYN, 2022. Number of primary schools and classes in each grade in Beijing in the 20212022 school year. https://www.ysxiao.cn/c/202203/51422.html.

[5] 

[6] 

[7] 

[8] 

[9] 

[10] 

[11] 

[12] 

[13] 

[14] 

[15] 

[16] 

[17] 

[18] 

[19] 

[20] 

[21] 

[22] 

[23] 

[24] 

[25] 

[26] 

[27] 

[28] 

[29] 

[30] 

[31] 

[32] 

[33] 

[34] 

[35] 

[36] 

[37] 

[38] 

[39] 

[40] 

[41] 

[42] 

[43] 

[44] 

[45] 

[46] 

[47] 

[48] 

[49] 

[50] 

[51] 

[52] 

[53] 

[54] 

[55] 

[56] 

[57] 

[58] 

[59] 

[60] 

[61] 

[62] 

[63] 

[64] 

[65] 

[66] 

[67] 

[68] 

[69] 

[70] 

[71] 

[72] 
Netease, 2022a. Haidian builds the urban slow traffic system, and the first bicycle dedicated road will be developed in the south. https://www.163.com/dy/article/HB7KKBQS0552UKSD.html.

[73] 
Netease, 2022b. Ranking of top 100 cities in China’s GDP in 2021. https://www.163.com/dy/article/ H1QK7JUM0552NMG9.html.

[74] 

[75] 

[76] 

[77] 

[78] 
People’s Government of Beijing Municipality PGBM, 2022. Main data of the seventh national population census in Beijing. http://www.beijing.gov.cn/gongkai/shuju/sjjd/202105/t20210519_2392877.html.

[79] 

[80] 

[81] 

[82] 

[83] 

[84] 

[85] 

[86] 

[87] 

[88] 

[89] 

[90] 

[91] 

[92] 

[93] 

[94] 
Tencent, 2022. What are the highquality schools in Chaoyang District—Elementary School. https://new. qq.com/rain/a/20220107A05Q5L00.

[95] 

[96] 

[97] 

[98] 

[99] 

[100] 

[101] 

[102] 

[103] 

[104] 
Wikipedia, 2022. Beijing Haidian Experimental Primary School. https://zh.wikipedia.org/wiki/%E5%8C%97% E4%BA%AC%E5%B8%82%E6%B5%B7%E6%B7%80%E5%8C%BA%E5%AE%9E%E9%AA%8C%E5%B0%8F%E5%AD%A6.

[105] 

[106] 

[107] 

[108] 

[109] 
YNET, 2022. Dongcheng district of Beijing promotes the reconstruction of the old city: Old hutongs usher in a new life. https://t.ynet.cn/baijia/32753188.html.

[110] 

[111] 

[112] 

[113] 

[114] 

[115] 

[116] 

/
〈  〉 