3.1 Baseline regression analysis
Table 2, Model (1) reports the regression results without including any control variables. Model (2) reports the OR regression results for model (1); model (3) reports the regression results with the inclusion of all control variables; and model (4) reports the OR regression results for model (3). All estimated coefficients of extracurricular physical activity on adolescents' self-education expectations are positive and pass the significance at 1% level test. This indicates that extracurricular physical exercise has a positive intervention effect on adolescents' self-education expectations, which is conducive to promoting adolescents' self-education expectations. In model (4), the OR for extracurricular physical activity was 1.2077, indicating that, all else being equal, overall, the self-education expectations of adolescents who participated in extracurricular physical activity were, on average, 20.77% points higher than those of adolescents who did not participate in extracurricular physical activity.
Regarding control variables, cognitive ability, academic performance, health status, self-confidence, and persistence all positively correlated with adolescents' self-education expectations. Regarding family characteristics variables, parental education, family economic conditions, and whether or not they are only children significantly impact adolescents' self-education expectations. The higher the parents' education level, the higher the self-education expectation of adolescents; relatively speaking, adolescents from only-child families and rich families have higher self-education expectations. The self-education expectation of adolescents attending private schools was significantly greater than that of adolescents attending public schools; the higher the quality of the school, the greater the self-education expectation of adolescents.
Table 2 Baseline regression estimation results
|
(1)
|
(2)
|
(3)
|
(4)
|
|
|
|
|
|
Extracurricular
physical exercise
|
0.4258***
|
1.5307***
|
0.1887***
|
1.2077***
|
|
(0.0273)
|
(0.0411)
|
(0.0275)
|
(0.0333)
|
|
|
|
|
|
Cognitive ability
|
|
|
0.2561***
|
1.2919***
|
|
|
|
(0.0178)
|
(0.0228)
|
|
|
|
|
|
Academic
Performance
|
|
|
0.0772***
|
1.0803***
|
|
|
|
(0.0019)
|
(.0019)
|
|
|
|
|
|
Health Status
|
|
|
0.1769***
|
1.1935***
|
|
|
|
(0.0304)
|
(0.0362)
|
|
|
|
|
|
Self-confidence
|
|
|
0.6559***
|
1.9269***
|
|
|
|
(0.0404)
|
(0.0754)
|
|
|
|
|
|
Persistence
|
|
|
0.2627***
|
1.3005***
|
|
|
|
(0.0487)
|
(0.0576)
|
|
|
|
|
|
Student Gender
|
|
|
-0.1107***
|
0.8951***
|
|
|
|
(0.0270)
|
(0.0244)
|
|
|
|
|
|
Whether or not to
live on campus
|
|
|
0.0009
|
1.0008
|
|
|
|
(0.0329)
|
(.0322)
|
|
|
|
|
|
Account Type
|
|
|
-0.0675**
|
0.9347**
|
|
|
|
(0.0314)
|
(0.0293)
|
|
|
|
|
|
Parenting
|
|
|
0.1044***
|
1.1101***
|
|
|
|
(0.0052)
|
(0.0057)
|
|
|
|
|
|
Family economic
conditions
|
|
|
-0.0332*
|
0.9868
|
|
|
|
(0.0175)
|
(0.0380)
|
|
|
|
|
|
Only one child or not
|
|
|
0.0915***
|
1.0958***
|
|
|
|
(0.0302)
|
(0.0335)
|
|
|
|
|
|
Whether the
parents live at
home
|
|
|
-0.0275
|
0.9729
|
|
|
|
(0.0332)
|
(0.0315)
|
|
|
|
|
|
Mobility
|
|
|
-0.0293
|
0.9710
|
|
|
|
(0.0356)
|
(0.0347)
|
|
|
|
|
|
Nature of school
|
|
|
-0.6384***
|
0.5281***
|
|
|
|
(0.0499)
|
(0.0270)
|
|
|
|
|
|
School Ranking
|
|
|
0.1877***
|
1.2064***
|
|
|
|
(0.0236)
|
(0.0281)
|
|
|
|
|
|
Location Type
|
|
|
-0.0195
|
0.9806
|
|
|
|
(0.0164)
|
(0.0163)
|
|
|
|
|
|
Regional
Distribution
|
|
|
0.0608***
|
1.0627***
|
|
|
|
(0.0164)
|
(0.0175)
|
|
|
|
|
|
N
|
16540
|
16540
|
16540
|
16540
|
3.2 Analysis of the effects of extracurricular physical activity on self-education expectations of adolescents with different characteristics
Given the significant individual differences among different types of adolescent students, does extracurricular physical activity have a heterogenous effect on self-education expectations of adolescents with different characteristics exhibit heterogeneity? We grouped the samples according to adolescents' cognitive ability, academic performance, health status, self-confidence level, and persistence to answer this question. Then we analyzed the heterogeneity of the effects of extracurricular physical activity on the self-education expectations of different groups of adolescents with different characteristics. In the grouping based on cognitive ability, the third quartile of the student's cognitive ability standard score served as the dividing line between the lower cognitive ability group, the middle cognitive ability group, and the higher cognitive ability group. In the academic performance grouping, we used the average standardized scores of students in language, mathematics, and English as the basis. Those with scores of 65 and below were considered the poor academic performance group, those with scores greater than 65 and less than 75 were considered the moderate academic performance group, and those with scores greater than 75 were considered the better academic performance group. In the grouping based on health status, healthy students were kept in one group, and students who were less healthy and unhealthy were kept in one group. In the grouping based on self-confidence, students who were more confident and very confident were kept in one group, and students who were less confident and less confident were kept in one group. In the grouping based on perseverance, students with good and very good perseverance were kept in one group, and students with less good and bad perseverance were kept in one group. The grouping estimation results are shown in Tables 3 and 4[1]。
Table 3 Heterogeneity analysis (I)
|
Cognitive ability
|
Academic Performance
|
|
relatively
low
|
Moderate
|
high
|
Poor
|
Moderate
|
better
|
Extracurricular physical exercise
|
0.2469***
|
0.2054***
|
0.0757
|
-0.2468***
|
0.2212***
|
0.0703
|
|
(0.0629)
|
(0.0346)
|
(0.0677)
|
(0.0729)
|
(0.0374)
|
(0.0959)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Cognitive ability
|
|
|
|
0.3188***
|
0.3309***
|
0.2288***
|
|
|
|
|
(0.0513)
|
(0.0239)
|
(0.0571)
|
|
|
|
|
|
|
|
Academic
Performance
|
0.0722***
|
0.0798***
|
0.0843***
|
|
|
|
|
(0.0037)
|
(0.0022)
|
(0.0051)
|
|
|
|
|
|
|
|
|
|
|
Health Status
|
0.2295***
|
0.1795***
|
0.1348*
|
0.2479***
|
0.2270***
|
0.0167
|
|
(0.0682)
|
(0.0383)
|
(0.0738)
|
(0.0812)
|
(0.0410)
|
(0.1063)
|
|
|
|
|
|
|
|
Self-confidence
|
0.6876***
|
0.6625***
|
0.8850***
|
0.6456***
|
0.7455***
|
0.7786***
|
|
(0.0773)
|
(0.0493)
|
(0.1144)
|
(0.0870)
|
(0.0511)
|
(0.1668)
|
|
|
|
|
|
|
|
Persistence
|
0.3598***
|
0.1774***
|
0.4142***
|
0.1308
|
0.3290***
|
0.3096**
|
|
(0.0844)
|
(0.0558)
|
(0.1421)
|
(0.1108)
|
(0.0602)
|
(0.1460)
|
|
|
|
|
|
|
|
Other Variables
|
Control
|
Control
|
Control
|
Control
|
Control
|
Control
|
N
|
3168
|
9956
|
3416
|
4896
|
7761
|
3883
|
Table 3 shows that the effect of extracurricular physical activity on adolescents' self-education expectations is significantly different in cognitive ability and academic performance. Regarding cognitive ability, extracurricular physical activity significantly increased the self-education expectations of adolescent students with moderate and lower cognitive ability levels but had no significant effect on students with higher cognitive ability. Further calculation and comparison of the ORs of the regression coefficients showed that the effect of extracurricular physical activity on the self-education expectation with the lower cognitive ability group was greater than that with the middle cognitive ability group. Extracurricular physical activity significantly increased the self-education expectations of adolescents with moderate academic performance. However, it significantly decreased the self-education expectations of adolescents with lower academic performance while not affecting the self-education expectations of those with better academic performance. Students with higher cognitive ability and better academic performance may already have higher self-education expectations; therefore, the effect of extracurricular physical activity on the self-education expectations of these adolescents is not significant. The above findings also partially confirm the findings of studies that "cognitive ability and academic performance are significantly and positively related to adolescents' self-education expectations" (Domina et al., 2011; Wu, Yun, 2018; Liu, 2020) [32-34]. For adolescents with poor academic performance, the amount of energy and time they spend on studying may be low, and weekend sports participation may further reduce their study time on the one hand and divert their study energy and interest on the other hand, making it difficult for them to set high educational expectations.
Table 4 Heterogeneity analysis (II)
|
Health Status
|
Self-confidence
|
Persistence
|
|
Health
|
Unhealthy
|
More
confident
|
Lack of
confidence
|
better
|
Poor
|
|
|
|
|
|
|
|
Extracurricular physical exercise
|
0.1671***
|
0.2354***
|
0.1886***
|
0.1018
|
0.1972***
|
0.1408
|
|
(0.0335)
|
(0.0519)
|
(0.0307)
|
(0.0604)
|
(0.0304)
|
(0.0845)
|
|
|
|
|
|
|
|
Cognitive ability
|
0.2446***
|
0.2718***
|
0.2353***
|
0.3503***
|
0.2439***
|
0.3026***
|
|
(0.0210)
|
(0.0348)
|
(0.0194)
|
(0.0478)
|
(0.0194)
|
(0.0475)
|
|
|
|
|
|
|
|
Academic
Performance
|
0.0757***
|
0.0808***
|
0.0800***
|
0.0630***
|
0.0781***
|
0.0716***
|
|
(0.0021)
|
(0.0035)
|
(0.0020)
|
(0.0044)
|
(0.0020)
|
(0.0049)
|
|
|
|
|
|
|
|
Health Status
|
|
|
0.1965***
|
0.0748
|
0.1556***
|
0.2806***
|
|
|
|
(0.0343)
|
(0.0712)
|
(0.0334)
|
(0.0820)
|
|
|
|
|
|
|
|
Self-confidence
|
0.7723***
|
0.5545***
|
|
|
0.7021***
|
0.5989***
|
|
(0.0515)
|
(0.0623)
|
|
|
(0.0432)
|
(0.1009)
|
|
|
|
|
|
|
|
Persistence
|
0.2129***
|
0.3192***
|
0.2405***
|
0.2925***
|
|
|
|
(0.0580)
|
(0.0716)
|
(0.0509)
|
(0.0965)
|
|
|
|
|
|
|
|
|
|
Other Variables
|
Control
|
Control
|
Control
|
Control
|
Control
|
Control
|
|
|
|
|
|
|
|
N
|
12486
|
4054
|
13941
|
2599
|
14036
|
2504
|
Table 4 reports the estimated results of the effects of extracurricular physical activity on self-education expectations of adolescents with different health states, levels of self-confidence, and levels of persistence. Regarding health status, extracurricular physical activity significantly increased the self-education expectations of adolescents with different health statuses. Further analysis showed that the regression coefficients of the core explanatory variables were 0.2354 > 0.16,71, and the corresponding OR values were 1.2654 > 1.1819, indicating that the effect of extracurricular physical activity on the self-education expectation of unhealthy adolescents was significantly greater than that of healthy adolescents. In terms of self-confidence, extracurricular physical exercise can significantly enhance the self-education expectation of more confident adolescents and the enhancement effect on the self-education expectation of adolescents who lack self-confidence is not significant. In terms of persistence, extracurricular physical exercise significantly increased the self-education expectation of students with better persistence. In contrast, the effect of raising self-education expectations among adolescents with poor persistence was not significant. Table 2 also shows that students' self-confidence and persistence significantly and positively affected their self-education expectations. However, students who lacked self-confidence and had poor persistence tended to have difficulty in establishing higher self-education expectations and exogenous stimuli such as extracurricular physical exercise had limited effects on raising self-education expectations of such groups.
3.3 Robustness test
To examine whether the above findings are reliable, further robustness tests are conducted as follows: First, Variable replacement: Since sports interest is highly correlated with weekend physical activity, students who are interested in sports have a high probability of participating in physical activity on weekends. Therefore, the original model is re-estimated by substituting "whether or not they are interested in sports" for the primary explanatory variable "whether or not they participate in physical exercise"; And then, variable and method substitution: the time spent on physical exercise on weekends (continuous variable) is used as the primary explanatory variable, and the self-education expectation (continuous variable) is used. The effect of extracurricular physical activity on adolescents' self-education expectation was re-estimated using an OLS model using weekend physical activity time (a continuous variable) as the primary explanatory variable and self-education expectation (converted into specific years) as the explanatory variable. Due to the space limitation, only the robustness tests of the model (2) are shown in Tables 5 and 6. Table 5 shows that the estimated coefficients of the primary explanatory variables are statistically significant at the 1% level for all the samples, consistent with the estimation results of model (3) in Table 2. The subgroup samples show that the direction and significance of the effects of extracurricular physical activity on self-education expectations of lower, moderate, and higher cognitive ability adolescents and self-education expectations of poor, moderate, and better academic performance adolescents remain consistent with the results in Table 3, indicating that the estimates in Tables 2 and 3 are relatively robust.
Table 5 Robustness test (I)
|
Total utility
|
Cognitive ability
|
Academic Performance
|
|
Full sample
|
relatively
low
|
Moderate
|
high
|
Poor
|
Moderate
|
better
|
Extracurricular physical exercise
|
0.0612***
|
0.0749**
|
0.0584***
|
0.0443
|
-0.0344*
|
0.0842***
|
0.0530
|
|
(0.0200)
|
(0.0357)
|
(0.0225)
|
(0.0460)
|
(0.0178)
|
(0.0242)
|
(0.0819)
|
|
|
|
|
|
|
|
|
Cognitive ability
|
0.4506***
|
|
|
|
0.6280***
|
0.6002***
|
0.3893***
|
|
(0.0349)
|
|
|
|
(0.1067)
|
(0.0470)
|
(0.1066)
|
|
|
|
|
|
|
|
|
Academic
Performance
|
0.1340***
|
0.1331***
|
0.1393***
|
0.1336***
|
|
|
|
|
(0.0035)
|
(0.0071)
|
(0.0040)
|
(0.0090)
|
|
|
|
|
|
|
|
|
|
|
|
Health Status
|
0.3245***
|
0.4681***
|
0.2954***
|
0.2317*
|
0.5192***
|
0.4050***
|
0.0113
|
|
(0.0599)
|
(0.1414)
|
(0.0746)
|
(0.1375)
|
(0.1700)
|
(0.0819)
|
(0.1991)
|
|
|
|
|
|
|
|
|
Self-confidence
|
1.3047***
|
1.2938***
|
1.2418***
|
1.5814***
|
1.1865***
|
1.4264***
|
1.3986***
|
|
(0.0786)
|
(0.1568)
|
(0.0953)
|
(0.2101)
|
(0.1788)
|
(0.1011)
|
(0.3135)
|
|
|
|
|
|
|
|
|
Persistence
|
0.5647***
|
0.9394***
|
0.3875***
|
0.6667***
|
0.3879*
|
0.7038***
|
0.6205**
|
|
(0.0921)
|
(0.1714)
|
(0.1056)
|
(0.2574)
|
(0.2257)
|
(0.1159)
|
(0.2693)
|
|
|
|
|
|
|
|
|
Other Variables
|
Control
|
Control
|
Control
|
Control
|
Control
|
Control
|
Other Variables
|
|
|
|
|
|
|
|
|
_cons
|
3.9793***
|
3.2549***
|
3.3677***
|
5.0036***
|
10.7627***
|
12.4868***
|
16.1769***
|
|
(0.3275)
|
(0.6705)
|
(0.3841)
|
(0.8557)
|
(0.6096)
|
(0.3027)
|
(0.7486)
|
|
|
|
|
|
|
|
|
N
|
16540
|
3168
|
9956
|
3416
|
4896
|
7761
|
3883
|
|
|
|
|
|
|
|
|
|
|
The results in Table 6 show that the direction and significance of the effects of extracurricular physical exercise on self-education expectations of adolescents with different health status groups, adolescents with different self-confidence groups, and adolescents with different persistence groups are consistent with the results in Table 4, indicating that the results in Table 4 are relatively robust. The estimation results in Tables 3 – 6 verified the reliability of the findings of this study, including the heterogeneity of the effects of extracurricular physical exercise on self-education expectations of adolescents with different characteristics.
Table 6 Robustness tests (continued)
|
Health Status
|
Self-confidence
|
Persistence
|
|
Health
|
Unhealthy
|
More
confident
|
Lack of
confidence
|
Stronger
|
Poor
|
|
|
|
|
|
|
|
Extracurricular physical exercise
|
0.0514**
|
0.0838**
|
0.0627***
|
0.0446
|
0.0629***
|
0.0508
|
|
(0.0202)
|
(0.0351)
|
(0.0186)
|
(0.0502)
|
(0.0189)
|
(0.0470)
|
|
|
|
|
|
|
|
Cognitive ability
|
0.4208***
|
0.5269***
|
0.4094***
|
0.6774***
|
0.4327***
|
0.5506***
|
|
(0.0398)
|
(0.0675)
|
(0.0367)
|
(0.0949)
|
(0.0370)
|
(0.0911)
|
|
|
|
|
|
|
|
Academic
Performance
|
0.1305***
|
0.1420***
|
0.1382***
|
0.1109***
|
0.1357***
|
0.1237***
|
|
(0.0039)
|
(0.0064)
|
(0.0036)
|
(0.0083)
|
(0.0036)
|
(0.0089)
|
|
|
|
|
|
|
|
Health Status
|
|
|
0.3552***
|
0.1720
|
0.2909***
|
0.5305***
|
|
|
|
(0.0654)
|
(0.1437)
|
(0.0641)
|
(0.1591)
|
|
|
|
|
|
|
|
Self-confidence
|
1.4244***
|
1.1298***
|
|
|
1.3417***
|
1.1019***
|
|
(0.0972)
|
(0.1206)
|
|
|
(0.0820)
|
(0.1942)
|
|
|
|
|
|
|
|
Persistence
|
0.4755***
|
0.6994***
|
0.5227***
|
0.7007***
|
|
|
|
(0.1072)
|
(0.1375)
|
(0.0940)
|
(0.1914)
|
|
|
|
|
|
|
|
|
|
Other Variables
|
Control
|
Control
|
Control
|
Control
|
Control
|
Control
|
|
|
|
|
|
|
|
_cons
|
4.5673***
|
3.3183***
|
4.7846***
|
7.0097***
|
3.8985***
|
4.3315***
|
|
(0.3656)
|
(0.5996)
|
(0.3410)
|
(0.8188)
|
(0.3311)
|
(0.8067)
|
|
|
|
|
|
|
|
N
|
12486
|
4054
|
13941
|
2599
|
14036
|
2504
|
[1] Due to the limitation of space, only some of the explanatory variables are listed, which can be obtained from the authors if needed, as follows.