The Effects of Principals’ Instructional Leadership on Primary School Students’ Academic Achievement in China: Evidence from Serial Multiple Mediating Analysis

Abstract

It is essential to reorient education in order to improve student learning outcomes and offer instruction in sustainable development. This reorientation should particularly focus on updating the educational concepts and behaviors of principals and teachers. It is a common practice to see principals’ instructional leadership (PIL) as a crucial aspect that may facilitate the development of teaching and learning in a sustainable manner. This article identified the influence path of principals’ instructional leadership on students’ academic achievement (SAA) through teachers’ teaching practice. In 2016 and 2017, survey data were gathered from 60,419 students and parents, 2232 teachers, and 506 principals from all 231 primary schools in the capital city of a central province in China. This paper uses a structural equation model to examine serial, multiple, mediating effects based on a value-added model to identify effective teachers’ professional development (ETPD). The research findings show that ETPD in Chinese primary schools includes teacher cooperation and communication but not professional guidance, innovation, and individual teaching reflection. PIL does not have a significant direct positive effect on SAA, but teaching strategies (TS), rather than ETPD, can mediate between PIL and SAA. Finally, PIL influences TS through ETPD and ultimately influences SAA. This paper argues that for providing sustainable and high-quality education, not all TPDs are effective for Chinese primary school teachers. Principals can promote SAA by encouraging teacher cooperation and communication to improve teaching strategies. Chinese principals should shift their attention more from administrative matters to the sustainable development of teaching and learning.

1. Introduction

Since the turn of the twenty-first century, to meet the challenges posed by the marketization of education, school-level competition, and public responsibility, primary school reform has called for the establishment of cooperative learning organizations to improve teaching quality and sustain students’ development [1]. In such contexts of accountability, principals are likewise concerned with enabling teaching and learning [2]. As a learning organization, the sustainable development of primary schools needs the strong instructional leadership of principals and the improvement of teachers’ teaching concepts and behaviors to sustain the development of students. The extensive review of instructional leadership articles published between 1940 and 2018 by Hallinger et al. [3] revealed that instructional leadership has not only endured but also expanded into one of the most influential paradigms guiding school leadership research, policy, and practice. It is a common practice to see principals’ instructional leadership (PIL) as a crucial aspect that may facilitate the development of teaching and learning in a sustainable manner. Quantitative empirical research on the effect of PIL on students’ academic achievement (SAA) can be loosely categorized as either direct or indirect effect studies. For the direct effect studies, several empirical studies and meta-analyses indicated that PIL had a positive effect on SAA [4,5,6,7]. Other studies, however, determined that PIL had no significant effect on SAA [8,9]. The inconsistencies among the research results mean that individual school leaders should apply the findings both with caution and with an understanding of their own particular social and school context [4]. Specifically, in China, Chen’s [10] finding that self-evaluated PIL had no significant effect on SAA in Shanghai middle schools led the author to conclude that certain Chinese middle schools had not yet realized the positive effect of PIL. To sum up, Chinese principals often concentrated on their daily administrative duties and did not plan regular engagement with teachers. Indeed, many teachers and students felt intimidated by high-status principals [11,12].

It is more widely understood that the effects of PIL on students are largely indirect. Using data from Taiwan, Pan et al. [13] found that both direct and indirect effects existed, with the indirect effect being stronger. Since 2000, researchers have sought a deeper knowledge of impact mechanisms, especially the mediating effects of PIL on SAA through teachers [6,14,15]. It means that PIL promotes students’ sustainable development by guiding teachers to reshape their teaching concepts and behaviors. This article argues that teachers, rather than principals, undertake the direct instruction of students. Scholars have mainly selected teachers’ professional development (TPD), teaching efficacy, teachers’ leadership, teaching practice, and job satisfaction, along with school culture or atmosphere, as mediating variables between PIL and SAA [16,17].

In China, since the Ministry of Education promulgated The Professional Standards for Principals of Compulsory Education Schools in 2013, instructional leadership for learning has become a significant dimension in the appointment and assessment of Chinese primary school principals. Although the attention in this sector has been gradually increasing [18], few studies have examined this effect, especially the impact mechanisms of PIL, on SAA in the Chinese context. Furthermore, the subject of the majority of studies on Chinese PIL are middle school principals, and the studies utilize large-scale survey databases such as PISA and TALIS. Few articles have focused on the effect of PIL on SAA in Chinese primary schools. To fill this research gap, this paper addresses the following question: how does PIL of Chinese primary schools promote SAA through the behaviors of teachers?

As the leaders of learning organizations, primary school principals should apply leadership toward learning, which is the core of school improvement and students’ sustainable development. The concept of PIL has transitioned from a narrow interpretation to a broad understanding [7]. In the early stage of its promotion, PIL mainly referred to engaging in teaching activities directly, such as attending lectures or maintaining classroom discipline [19]. Since the 1990s, the understanding of PIL has expanded to include all the factors that can influence teaching and learning. From the management and system perspective, principals, as facilitators, need to pay attention to teaching activities and TPD to sustain the development of students [20,21]. Marks and Printy [22] shifted the concept of PIL from a principal-centered practice to a shared practice, emphasizing the empowerment of teachers to participate in the decision-making related to instruction.

This paper emphasizes the notion that the PIL of Chinese primary schools should define and communicate shared goals and empower teachers to participate in decision-making related to instruction. Principals should also supervise instruction and guide teachers to find and solve problems in teaching practice to sustain the professional competence of teachers. For instance, principals should lead teachers to improve teaching methods, adopt new teaching strategies, and make teaching innovative as well as provide teachers with support, guidance, evaluation, and feedback. They should provide teaching and learning resources and support teachers in participating in TPD to improve teaching quality. Therefore, in this paper, the PIL of Chinese primary schools includes the ability to make democratic decisions, encourage teaching innovation, and support TPD.

This paper aims to identify the impact mechanisms between PIL of Chinese primary schools and SAA through the behaviors of teachers. As PIL in this paper includes encourage teaching innovation and support TPD, we focus on the mediating roles of TPD and teaching strategies (TS) in the context of China. Some studies have argued that the prevalence of teacher participation in school-based, job-embedded, professional development activities was the reason for Shanghai’s success in PISA [2]. However, there is a remarkable scarcity of large-scale, quantitative research into the effectiveness of TPD in Chinese primary schools [23]. Chinese TPD activities are often an institutionalized or even mandated practice [24], bringing the effectiveness of TPD into question. For example, some teachers have reported that collaboration among colleagues is only at the instrumental level; the content of school-based learning activities is often formalized and ignores teachers’ individual needs; and teachers tend to uncritically follow the ideals or suggestions promoted by experts [25,26]. Therefore, this paper seeks to identify effective TPD and examine the mediating effect of effective TPD between PIL and SAA in Chinese primary schools. As previous studies have indicated that PIL had a positive, indirect effect on classroom instruction, such as teaching strategies (TS), through the mediation of TPD [27,28,29], we also examined the mediating effect of TS and any serial multiple mediating effects. The theoretical framework is shown in Figure 1.

2. Literature Review

2.1. Literature on the Indirect Effect of PIL on SAA

2.1.1. Effective Professional Development for Teachers as a Mediator

Studies show that TPD promotes teachers’ knowledge, skills, and instructional practices [30,31] and student learning [32,33]. However, not all TPDs are conducive to the improvement of SAA [34,35]. Some studies have also found that formal teaching training or on-the-job training had no significant effect on SAA [36,37]. The need to identify effective TPDs in Chinese primary schools in order to promote SAA remains.

Regarding the relationship between PIL and TPD, some studies have affirmed a positive relationship between PIL and TPD in Chinese primary schools and middle schools [38,39]. When principals paid attention to the construction of the teaching and learning environment, supported TPD, encouraged teacher cooperation, and gave teachers timely and effective teaching feedback, this was found to be conducive to SAA [6,40]. Zhang and Hu [41] discovered that, in primary and secondary schools in China, principals could improve teachers’ professional quality by encouraging cooperative behaviors which then had a significant effect on SAA. However, whether effective TPD can be a mediator between PIL and SAA in Chinese primary schools still needs to be examined. Therefore, we propose Hypothesis 1:

Hypothesis 1:

Effective TPD plays a mediating role between PIL and SAA in Chinese primary schools.

2.1.2. Teaching Strategies as a Mediator

Recent research has confirmed that PIL is positively related to teaching practice, including TS. For instance, Vanblaere and Devos [42] explored that PIL was related to teachers’ participation in de-privatized practices and reflective dialogues. Chen and Guo [43] stated that PIL had a significant influence on TS in Chinese primary schools. Xia et al. [44] presented TS as an important mediator between principals’ instructional leadership evaluated by teachers and teaching quality in Chinese middle schools. However, the measurement and estimation method, sample size, and representation in Xia et al.’s paper need to be improved, and whether TS can be a mediator between PIL and SAA in Chinese primary schools still needs to be examined. Therefore, we propose Hypothesis 2:

Hypothesis 2:

TS plays a mediating role between PIL and SAA in Chinese primary schools.

2.2. Literature on the Serial Multiple Mediating Effects between PIL and SAA

Some papers reported positive relationships between TPD and teaching practices such as TS. For instance, Ozdemir and Pusmaz [28] discovered that web-based TPD had a positive effect on mathematics teachers’ TS of problem solving in middle school. Ke et al. [23] found that collective lesson planning and teacher collegiality had significant effects on the teacher efficacy and teaching strategies of Chinese secondary school teachers. Previous studies also indicated that PIL had a positive indirect effect on classroom instruction through the mediation of TPD, such as teachers’ collaboration [27,28,29]. For instance, Bellibaş [45] found that, in primary and secondary schools in Turkey, there is an indirect relationship between PIL and teachers’ instructional practices, including TS, with the full mediation roles of shared or collaborative practices among teachers as one TPD behavior. According to Bach et al. [46], in primary and secondary schools in Germany, the principal’s innovation-oriented leadership is positively associated with teachers’ instructional improvement, having a small direct effect, and is mediated by instructional collaboration and individualized staff development.

Few studies analyzed the serial multiple mediating effects between PIL and SAA via TS and TPD in China. Xia et al. [44] reported that the PIL of Chinese middle schools had a positive effect on teaching quality by influencing TS and the construction of learning communities, with TS and learning communities playing serial multiple mediating roles. They suggested that the construction of a learning community improved TS; however, the measurement and estimation method, sample size, and representation of that study need to be improved. Whether effective TPD and TS play serial multiple mediating roles between PIL and SAA in Chinese primary schools still requires examination. Therefore, we propose Hypothesis 3:

Hypothesis 3:

The PIL of Chinese primary schools influences effective TPD, which influences TS and, finally, influences SAA.

3. Materials and Methods

3.1. Procedure and Participants

This paper was supported by a large joint project conducted by the Collaborative Innovation Center of Assessment for Basic Education Quality and the Education Bureau of a central provincial capital city. The research team used a general survey to investigate fifth-grade students of all 231 primary schools in the capital city of a central province in China. The students took Chinese, mathematics, and science tests. Additionally, 60,419 students and parents, 2232 teachers, and 506 principals completed a questionnaire in 2016 and 2017 to elicit information primarily about principals’ instructional leadership, teachers’ professional development, and teaching strategies. Among the sample of students, 46.3% are female, 2.5% are from an ethnic minority, 41.9% are from rural areas, and 6.2% are from incomplete families, such as single-parent families. About 20 percent of parents have a bachelor’s degree or above. Among the sample of teachers, 86.8% are female, 83.2% have a bachelor’s degree or above, 41.6% have senior titles, and about half of the teachers have more than 10 years of teaching experience. As for the sample of principals, 49.0% are female, 87.2% have a bachelor’s degree or above, and 35.6% have been principals or vice-principals for 10 years or more.

3.2. Measures

3.2.1. Tests of Students’ Academic Achievement

The test items for Chinese, mathematics, and science were produced based on the requirements set by the National Curriculum Standard (2011 version). For instance, the mathematics test items measure students’ understanding of number sense and algebra, graphs and geometry, statistics and probability, and students’ cognitive competence to know, understand, grasp, and apply knowledge. Educational experts and school teachers worked together in developing those items. The instruments went through the processes of interviews with 6 students, a pilot with 30 students, another pilot with 300 students, and independent reviews by experts in content areas. All these processes had good construct validity. An item response theory model was used to analyze the psychometric properties of the items. Detailed information on the validity of these tests was reported by the Collaborative Innovation Center for Chinese Basic Education Quality (2016 and 2017).

3.2.2. The Principals’ Instructional Leadership Scale

The measurements of PIL in PISA, TIMSS, and TALIS were used as references. Based on China’s Professional Standards for Principals of Compulsory Education Schools, the principals’ instructional leadership scale (Appendix A) used in this paper included three dimensions: five items for ‘democratic decision-making’ (DDM, e.g., providing opportunities for teachers to participate in teaching reform), four items to measure ‘encouraging teaching innovation’ (ETI, e.g., encouraging teachers to try new teaching methods and take them into practice), and five items to measure ‘supporting teachers’ professional development’ (SPD, e.g., inviting experts and providing effective professional guidance and help for teachers).

Teachers were asked to respond to the items on a Likert scale from 1 (strongly disagree) to 5 (strongly agree), with higher scores indicating higher levels of PIL. The coefficients of internal consistency reliability (Cronbach’s alpha) of responses to the scale and the sub-scales DDM, ETI, and SPD were 0.986, 0.978, 0.976, and 0.967, respectively; that is, all larger than 0.9. Taking DDM, ETI, and SPD as latent variables, the results of a simple CFA, not a second-order CFA, provided evidence of the construct validity of the scales. All the factor loadings of the items from the T-E scale were between 0.798 and 0.976. CFI and TLI of this scale were 0.987 and 0.979, all larger than 0.9. RMSEA was 0.078, which indicates adequate model fit [47].

3.2.3. The Teachers’ Professional Development Scale

TPD refers not only to the improvement of specific abilities obtained through on-the-job training but, also, to progress with respect to teaching goals, teaching skills, and the ability to cooperate with colleagues [48]. The measurement of teachers’ professional development in the TIMSS and NAEP were used as references. The teachers’ professional development scale (Appendix B) included three dimensions: three items to measure ‘professional guidance or innovation’ (PGI, e.g., attend experts’ lectures), three items to measure ‘teacher cooperation and communication’ (TCC, e.g., share teaching experiences and discuss problems with colleagues), and two items to measure ‘individual teaching reflection’ (ITR, e.g., obtain teaching knowledge or reflect on teaching problems on my own).

Teachers were asked to respond to the items on a Likert scale from 1 to 5 indicating the number of times they participated in TPD; the higher the scores are, the more frequently the teachers engaged in TPD. The Cronbach’s alpha of this scale and the sub-scales PGI, TCC, and ITR were 0.910, 0.716, 0.914, and 0.863, respectively; that is, all larger than 0.7. Taking PGI, TCC, and ITR as latent variables, the results of a simple CFA, not a second-order CFA, provided evidence of the construct validity of this scale. The factor loadings of the items from this scale were between 0.413 and 0.946. CFI and TLI of this scale were 0.986 and 0.973; RMSEA was 0.076, which indicates adequate model fit [47].

3.2.4. The Teaching Strategy Scale

The measurements of TS in PISA, TIMSS, and NAEP were references. The teaching strategy scale (Appendix C) was developed under the theoretical framework of Wang and Wo [49] and included three dimensions: five items to measure ‘individualized instruction’ (II, e.g., teachers provide us individual advice for learning), three items to measure ‘participatory instruction’ (PI, e.g., teachers organize group discussions in the class), and five items to measure ‘guided inquiry’ (GI, e.g., teachers encourage us to make a hypothesis and test it with various methods).

Students were asked to respond to the items on a Likert scale from 1 (strongly disagree) to 5 (strongly agree), with higher scores indicating higher levels of TS used. The Cronbach’s alpha of this scale and the sub-scales II, PI, and GI were 0.987, 0.943, 0.947 and 0.989, respectively; that is, all larger than 0.9. Taking II, PI, and GI as latent variables, the results of a simple CFA, not a second-order CFA, provided evidence of the construct validity of this scale. The factor loadings of the items from this scale were between 0.782 and 0.988. CFI and TLI of this scale were 0.989 and 0.983; RMSEA was 0.076, which indicates adequate model fit [47].

3.3. Analytic Approaches

3.3.1. Value-Added Model (VAM)

VAM, taking schools as the analysis unit, was used to estimate the effect of the TPD of Chinese primary school teachers on SAA to identify the effective TPD. VAM can solve the problem of omitted variables to a certain extent, stripping out the time-dependent variables that impact SAA [50]. The VAM is as follows:

$$Y_t={\beta }_0+{\beta }_1{Y}_{t-1}+{\beta }_{2}TP{D}_t+{\beta }_3{C}_t+\epsilon$$

(1)

$Y_t$ and $Y_{t-1}$ indicate the average scores in Chinese literacy, mathematics, and science at the school level in 2017 and 2016, respectively. The standardized scores of primary school students in Chinese, mathematics, and science were aggregated at the school level. $TP{D}_t$ indicates the items of the scale of TPD responded to by Chinese primary school teachers in 2017, such as attending others’ lectures and discussing with colleagues after class. $C_t$ represents control variables, including the student–teacher ratio [51] and the average school-level socioeconomic status (SES) of students in 2017. SES consisted of three indicators: (a) the highest professional level of both parents; (b) the highest level of education of both parents; and (c) the household assets. The standardized scores of primary school students’ SES were aggregated at the school level. The proportion of teachers with a bachelor’s degree [52,53], the average teaching years, the quadratic term of the average teaching years [54], and the teacher’s average working hours per day [55] were also included in $C_t$.

3.3.2. Structural Equation Model (SEM)

A simple SEM, not a second-order SEM, was used to estimate the effect paths of PIL of Chinese primary schools on SAA and test whether effective TPD and TS play serial multiple mediating roles. The SEM is as follows:

$$X_2=P_{21}{X}_1+{\epsilon }_2$$

(2)

$$X_3=P_{31}{X}_1+P_{32}{X}_2+{\epsilon }_3$$

(3)

$$X_4=P_{41}{X}_1+P_{42}{X}_2+P_{43}{X}_3+P_{4m}{C}_m+{\epsilon }_4$$

(4)

$X_1$ indicates the latent variable of PIL, including three observed variables: DDM, ETI, and SPD. $X_2$ indicates the latent variable of effective TPD. The included observed variables are derived from the following analysis. $X_3$ indicates the latent variable of TS, including three observed variables: II, PI, and GI. $X_4$ indicates the latent variable of SAA in 2017, including three observed variables: the average Chinese, mathematics, and science scores of primary school students at the school level. $C_m$ indicates control variables, including SAA in 2016, student–teacher ratio, average SES, proportion of teachers with a bachelor’s degree, average teaching years and its quadratic term, and average working hours. $P_{ij}$ indicates the path coefficients. ${\epsilon }_i$ indicates the residuals (i = 2,3,4; j = 1, 2, …, m).

4. Results

4.1. Descriptive Statistics

The statistical software Stata 13.0 was used to analyze the data. Descriptive statistics of the major variables used in VAM and SEM are shown in

Table 1. Descriptive Statistics of Variables Used in the VAM and SEM.

Variables	Mean	SD	Max	Min	N
Students’ Academic Achievement (SAA) in 2017	500.000	92.760	786.720	307.000	277
Democratic Decision-Making (DMM)	4.333	0.485	5.000	2.667	277
Encouraging Teaching Innovation (ETI)	4.488	0.398	5.000	3.063	277
Supporting TPD (SPD)	4.347	0.467	5.000	2.733	277
Professional Guidance or Innovation (PGI)	2.904	0.484	4.667	1.000	277
Teacher Cooperation and Communication (TCC)	4.145	0.567	5.000	2.000	277
Individual Teaching Reflection (ITR)	4.147	0.565	5.000	2.167	277
Individualized Instruction (II)	3.632	0.356	4.779	2.864	277
Participatory Instruction (PI)	3.594	0.401	4.892	2.667	277
Guided Inquiry (GI)	3.781	0.376	4.921	2.771	277
SAA in 2016	500.000	93.596	753.919	251.400	264
Student–Teacher Ratio	20.725	5.866	40.875	1.000	243
Average SES	−0.151	0.403	0.916	−1.101	277
Proportion of Teachers with a Bachelor’s Degree	0.815	0.185	1.000	0.000	277
Average Teaching Years	10.897	4.312	20.000	1.500	277
Average Working Hours	8.681	0.940	12.375	6.000	277

below. The average student–teacher ratio of the sample was 20.725. The average proportion of teachers with a bachelor’s degree was 81.5%. The average teaching years and working time per day for teachers were 10.89 years and 8.681 h, respectively. The mean of the average score of primary school students participating in the 2017 and 2016 project surveys was 500.000 for both, but the standard deviations were 92.760 and 93.596, respectively. The mean of items contained in each dimension of each scale represented the score of each dimension. Among the three dimensions of PIL, the encouraging teaching innovation score (ETI, M = 4.488) was relatively high. Among the three dimensions of TPD, the scores of teacher cooperation and communication (TCC, M = 4.145) and individual teaching and reflection (ITR, M = 4.147) were high. Among the three dimensions of TS, the guided inquiry score (GI, M = 3.781) was relatively high.

4.2. The Effects of TPD of Chinese Primary Schools on SAA

Effective TPD mainly refers to the behaviors that improve teachers’ knowledge, skills, attitudes, and emotions, guide teachers’ teaching practice, improve their teaching quality, and promote students’ development [56]. The results of the effect of TPD on SAA to identify the effective TPD of Chinese primary schools are shown in

Table 2. The Effects of Three Dimensions and Sub-items of TPD on SAA.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
PGI	0.021
	(6.056)
TCC		0.060 *
		(5.334)
ITR			0.038
			(5.344)
PGI-1				0.0004
				(5.330)
PGI-2					−0.001
					(4.888)
PGI-3						0.047
						(4.405)
TCC-1							0.055 *
							(4.444)
TCC-2								0.054 *
								(5.434)
TCC-3									0.055
									(4.899)
ITR-1										0.039
										(4.584)
ITR-2											0.031
											(5.504)
Control Variables 1	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES	YES
F	114.02 ***	115.92 ***	114.61 ***	113.73 ***	113.73 ***	115.19 ***	115.71 ***	115.51 ***	115.45 ***	114.64 ***	114.34 ***
N	231	231	231	231	231	231	231	231	231	231	231
Adjusted R2	0.797	0.800	0.798	0.797	0.797	0.799	0.800	0.799	0.799	0.798	0.798

¹ Control Variables include SAA in 2016, student–teacher ratio, average SES, proportion of teachers with a bachelor’s degree, average teaching years and its quadratic term, and average working hours. The standardized coefficients are shown in the table; * p < 0.1, *** p < 0.01.

. Among the three dimensions of TPD, only teacher cooperation and communication (TCC) had a significant, positive effect on SAA (p < 0.1). Neither professional guidance or innovation (PGI) nor individual teaching and reflection (ITR) showed a significant effect (p > 0.1). Therefore, in primary schools in China, TPD such as teacher cooperation and communication (TCC)—including attending others’ lectures, engaging in discussions with colleagues after class, and sharing teaching experiences and discussing problems with colleagues—are more conducive to improving SAA.

4.3. The Effect Paths of PIL of Chinese Primary Schools on SAA

To test the effect paths of PIL of Chinese primary schools on SAA, an SEM was conducted (Figure 2). The estimation method was maximum likelihood with missing values and the standard error type was bootstrap. The SEM had a mediocre fit with the data (χ² = 301.907; df = 88; p < 0.001; CFI = 0.949; TLI = 0.936; RMSEA = 0.094). The SEM (P-E) also had a good fit with the data (χ² = 268.570; df = 88; p < 0.001; CFI = 0.952; TLI = 0.940; RMSEA = 0.086) [57]. The model showed that PIL had no significant positive direct effect on SAA ($\mathsf{\beta } =0$.043, p > 0.1).

Regarding indirect effects, effective TPD did not play a mediating role between PIL and SAA. PIL showed a significant, positive effect (β = 0.417, p < 0.01) on effective TPD, but effective TPD had no significant effect on SAA (β = −0.007, p > 0.1). The joint test of the product of the path coefficients using bootstrapping was −0.003 (p > 0.1), so Hypothesis 1 was rejected. TS played a role as a mediator between PIL and SAA. PIL had a significant, positive effect (β = 0.197, p < 0.01) on TS, and TS also had a significant, positive effect (β = 0.093, p < 0.1) on SAA. The joint test of the product of the path coefficients using bootstrapping was 0.018 (p < 0.1), so Hypothesis 2 was confirmed. Effective TPD had a significant, positive effect (β = 0.313, p < 0.01) on TS. The joint test of the product of the serial multiple mediating effect path coefficients that PIL-T influenced effective TPD, then influenced TS, and, ultimately, influenced SAA using bootstrapping was 0.012 (p < 0.1); thus, Hypothesis 3 was confirmed.

5. Discussion

Our study found that, for the three dimensions of TPD, only teacher cooperation and communication (TCC)—rather than professional guidance or innovation (PGI) or individual teaching and reflection (ITR)—had a significant positive effect on SAA. It means that increasing the frequency and quality of communication and cooperation among teachers is conducive to the sustainable development of students. TPD can be divided into independent and non-independent professional development activities. Independent professional development refers to spontaneous and self-directed activities led by the teachers themselves, whereby the learning content, goals, and methods are controlled by the teachers [48,58]. Dependent professional development refers to the activities teachers participate in passively to meet the requirements of school rules, especially teacher training activities organized and implemented by schools or relevant education departments. Teachers’ cooperation and communication activities, such as attending others’ lectures and engaging in discussions with colleagues after class or sharing teaching experiences and discussing problems with colleagues, belong to independent professional development activities, and we found they were effective. Teachers could reflect on their teaching, summarize their experiences, improve their teaching skills, gain peer respect and appreciation, and achieve self-realization through these activities. Non-independent professional development activities, such as attending experts’ lectures, participating in the teaching and research activities organized by the school or other departments, and preparing and evaluating lessons collectively with the teaching and research group, were usually not established specifically for the teaching problems encountered by individual teachers. Therefore, they had little effect on improving individual teachers’ teaching practice and SAA. This finding is in line with the findings that teacher collaboration is an important characteristic of successful schools and a significant predictor of school improvement and students’ sustainable development [59,60]. Some studies have also argued that successful TPD is collaborative and sustainable and have suggested alternative ways for teachers to learn [61]. Compared with ITR, TCC, including attending others’ lectures and discussing with colleagues after class and sharing teaching experiences and discussing problems with colleagues, can improve SAA through the exchange and sharing of knowledge and experience among teachers to improve teaching skills. On the other hand, it is also possible to improve SAA by providing teachers with positive emotional support, reducing their loneliness within the organization, making teachers more motivated, reducing their work pressure or work burden, boosting their work morale, and improving their work efficiency [1,60].

From the perspective of social capital, the reasons for the effectiveness of TCC to cause SAA more than PGI, such as participating in project research or the teaching and research activities organized by the school or other departments, may be related to the foundational mechanism of the relationship. Social capital measured by teachers’ cooperative behaviors can demonstrate two mechanisms: organizational intervention and spontaneous formation. Informal teacher cooperation such as TCC corresponds to the spontaneous exchange of information, opinions, ideas, and mutual assistance among teachers according to their needs, such as communicating their confusions or problems with regard to teaching, and jointly seeking solutions. When teachers can obtain support from other teachers, they are likely to have a high sense of trust and commitment to the relationship network, which is conducive to both the sustainable development of teachers and students. With formal teacher cooperation such as PGI, the composition, form, and content are usually regulated by schools or other departments. Under formal cooperation, teachers mainly cooperate in accordance with the organization’s goals and interests, with the result that their personal commitment and work enthusiasm are often not high. Therefore, sustainable school improvement is limited [62]. Fear of authority, an important characteristic of East Asian culture, might cause teacher cooperation in East Asia to be top-down, mandatory, or passive [63]. Centralized training under traditional Chinese bureaucratic management may make teachers more inclined to simply fulfil the tasks assigned by school leaders rather than act spontaneously and consciously in the teaching community. This approach may generate resistance among teachers and it may be likely to increase teachers’ work burden due to frequent meetings and discussions, additional tasks, and corresponding responsibilities, while reducing teachers’ teaching time or other student-related activities, potentially leading to competition, anxiety, and even conflict among teachers [60].

In addition, we found that effective TPD did not play a role as a mediator between PIL and SAA in primary schools in China. This may be because effective TPD, namely, teachers’ cooperation and communication, is generally generated spontaneously in the daily teaching work of teachers rather than initiated by principals’ leadership, command, or administration. The influence of PIL on TPD varies across types of learning activities and countries [64]. From the principals’ perspective, having weak relationships between PIL and TPD means that principals do not have enough space to exert influence on TPD. Considering Chinese government-driven teacher policies, TPD may rely more on individual teachers, collective norms, or policy-guided systems than on principals’ efforts or guidance. In addition, as existing research suggests [65], if there are divergences between national policies and individual school contexts, principals may discourage TPD if they do not align with school organizational goals and school focus. On the other hand, the measurement of PIL in our study includes the dimension of supporting teachers’ professional development, but most of the items—such as inviting experts, providing effective professional guidance and help to teachers, and encouraging teachers to participate in teaching and training activities—belong to dependent and ineffective TPD, under which it is difficult to improve SAA and sustain the development of students.

As for TS, we found that it played a role as a mediator between PIL and SAA in primary schools in China. This result is consistent with the research on the PIL of Chinese middle schools [44]. Therefore, to sustain SAA, Chinese primary school principals should pay more attention to giving teachers correct and effective evaluations to help teachers solve teaching problems and improve their teaching strategies. Since the implementation of the new curriculum reform of basic education in China, the individualized instruction, participatory instruction, and guided inquiry strategies have been widely recognized by teachers and has been extensively used in daily teaching. Thus, they have been an important strategy for teachers to improve SAA [66]. Using PISA China data, Li and Chu [67] confirmed that Chinese teachers’ emphasis on TS such as independent thinking, cooperative inquiry, and the discussion of problems based on real situations has played an important role in the effectiveness of curriculum reform in China.

As for the serial multiple mediating effects of effective TPD and TS, this paper confirmed that PIL influenced TS through effective TPD and, finally, influenced SAA. This finding is in line with the findings that effective TPD is conducive to the improvement of TS. For instance, Lee and Kim [64] argued that mentoring, peer observation, and coaching activities are regarded as effective TPDs in that they rely on interactions and the co-development of expertise between teachers and focus on teachers’ individual needs to improve instructional practices. Therefore, in terms of supporting TPD, Chinese primary school principals should create a cooperative atmosphere for teachers and attach importance to the improvement of teachers’ teaching strategies. They can encourage teachers to share teaching experiences and help each other to make teaching innovation.

Finally, it should be pointed out that this paper has some limitations in terms of the measurement and estimation methods used. First, TS and TPD are extensive and varied. However, the scale adopted in this paper only examined some kinds of TS and TPD. Second, as for SAA in the previous period—namely, the average students’ academic scores at the school level controlled in the VAM and SEM models—the possible influencing factors that did not change with time were removed, but there were other errors of omitted variables that changed over time. Follow-up research should use more effective measurement tools for TS and TPD and explore more possible impact paths for the relationship between PIL and SAA to sustain teaching quality and students’ development. Thirdly, since the COVID-19 outbreak, most primary and secondary school students in China study at home, making it difficult to carry out large-scale paper questionnaire surveys of principals, teachers, students, and their parents; further, so as to build a large tracking data in school level, we used a data of a central provincial capital city in 2016 and 2017. Future studies can use newer data to conduct more in-depth studies.