Study on the Influencing Factors of Chinese Consumers’ Willingness to Participate in E-Commerce Agricultural Product Recyclable Packaging Recovery

Abstract

In recent years, the rapid development of the agricultural product e-commerce and logistics industry has generated a large amount of express packaging waste. The way in which to promote the reuse of express packaging and reduce carbon emissions has become an important issue, and the current research on consumers’ willingness to express packaging recycling is still in its infancy. This study, based on the theory of Extended Planned Behavior, constructed a factor model observing the influence of consumers’ willingness to participate in the e-commerce of agricultural products’ recyclable packaging recovery. Using partial least squares structural equation modeling to process and analyze 659 questionnaire data, the results show that policy promotion and perceived benefits had the highest degree of positive influence on recycling willingness, and there was a relatively weak positive effect of packaging performance and subjective norm on consumers’ willingness to recycle in terms of packaging performance and the subject norm of recycling willingness. The attitude towards recycling had the smallest impact on the willingness to recycle. Therefore, in order to promote the production and use of recyclable packaging, government policies should be strengthened, and consumers’ perceived benefits and subjective norm should be improved, as well as the packaging performance of recyclable packaging. This study provides new insights into Chinese consumers’ participation in the recycling and promotion of recyclable packaging for e-commerce agricultural products.

1. Introduction

E-commerce has flourished in recent years, bringing great convenience to people’s lives; more and more people tend to choose online shopping [1]. In 2021, China’s total express delivery business reached 108.3 billion pieces, exceeding 100 billion pieces for the first time. At the same time, China’s agricultural e-commerce market demand is also growing [2]. However, this has also brought about a large amount of packaging waste [3], which has caused immeasurable harm to the ecological environment [4]. Recyclable packaging can be used many times, causing little pollution to the environment, and its packaging performance is relatively better, which can better protect the agricultural products purchased by consumers, reduce carbon emissions in the supply chain of e-commerce agricultural products, and increase economic benefits; thus, promoting the use of recyclable packaging in the e-commerce logistics supply chain of agricultural products has become the current development trend. For the purchase and use of recycled packaging, most of it is currently done in developed countries [5]. However, in China, recycled packaging and its practical application are still in the development stage. The packaging waste recycling rate in the U.S. is 30%, while the relevant data in China is not yet optimistic, with less than 20% of packaging materials being recycled, which indicates that significant improvements are still needed in packaging design, marketing, and policy support. At the same time, promoting the use of recyclable packaging is expected to achieve efficient recycling of resources, protect the environment, and promote the green cycle of sustainable development of the express industry [6].

In order to achieve the “dual-carbon” goal of carbon peaking by 2030 and carbon neutrality by 2060, China has in recent years issued and implemented the “E-commerce Logistics Recyclable Packaging Management Code”, “Dimensions and technical requirements of fruit and vegetable cycling container”, and other management documents, which clearly stipulate the recyclable crates that comply with the national standards of China and are suitable for agricultural products. Agricultural products can be recycled crates. It can be seen that recyclable packaging is gradually occupying an important position in the express packaging of agricultural products, and the implementation of relevant regulations in China has also helped the production and promotion of recyclable packaging for e-commerce agricultural products. Recycled packaging means that packaging is designed, manufactured, used, and disposed of in accordance with the principles of circular development. It emits less carbon throughout the supply chain. In recent years, with the international community’s increasing concern for economic and environmental issues, the annual production of recyclable packaging has been on the rise [7]. It is also environmentally friendly, as well as human and livestock health friendly. Recyclable packaging is beneficial, safe, and healthy for individuals and organizations throughout the packaging life cycle [8]. With the use and promotion of recyclable packaging, consumers’ willingness to purchase recyclable packaging products will also be affected. An investigation found that Wal-Mart attempted to achieve packaging neutrality by 2025 through its 4Rs strategy, and that its promotion of recyclable packaging influenced consumers’ purchasing decisions and attitudes toward specific brands [9]; therefore, encouraging consumers to participate in recycling of recyclable packaging needs to be pushed in a directional way [10]. For the purchase and use of recyclable packaging, most of it is currently done in developed countries [5]. There are studies comparing the United States as a developed country with other countries in terms of recycling of packaging waste; however, China as a developing country has a different situation, and the realization of promoting recyclable packaging needs to be further researched.

In addition to e-commerce purchasing, agricultural recycling packaging can help solve environmental problems such as carbon emissions and increasing express packaging waste. Moreover, a comprehensive analysis of consumer recycling willingness in agricultural recycling packaging can promote research and development in the field of recycling packaging, improve ecological efficiency, reduce waste and pollution, and accelerate the realization of China’s “dual-carbon” goal of sustainable development policies. The study of the recycling system of agricultural products has important theoretical significance and practical value in promoting the widespread use of agricultural products recycling packaging.

In this study, recyclable packaging recovery refers to logistics packaging containers that can be put into secondary use through recycling can be considered as recycled packaging. Recycled packaging is made of durable materials such as plastic hollow sheets, wood, or metal, thus providing clear advantages in terms of effectiveness, sustainability, and safety, and is considered an important part of sustainability programs implemented in logistics segments of various industries. Consumers’ willingness to participate in e-commerce agricultural product recycling packaging recovery is influenced by a variety of factors, including recycling attitude, subjective norm, perceived benefits, packaging performance, and policy promotion. In addition, starting from the extended theory of planned behavior, this study analyzes and discusses the data results by distributing questionnaires to study the influencing factors of consumers’ willingness to participate in e-commerce agricultural product recycling packaging recovery, with the following specific research objectives:

• Determine the influencing factors of consumers’ willingness to participate in recycling packaging recovery of e-commerce agricultural products and put forward relevant hypotheses.
• Determine the extent to which each influencing factor affects consumers’ willingness to recycle and explore the internal correlation between the factors.
• Discuss how to promote and use recyclable packaging for e-commerce agricultural products.

The rest of the paper is organized as follows: “Literature Review and Hypothesis” introduces the relevant theories in the literature and puts forward the research hypotheses; “Research Methodology” shows the research methodology and the results of model calculation; “Discussion” discusses the research results; and “Conclusions and Limitations” describes the findings and shortcomings of the study.

2. Literature Review and Hypothesis

2.1. Literature Review

Theory of Reasoned Behavior: The Theory of Reasoned Behavior (TRA) is the predecessor of the Theory of Planned Behavior (TPB). The theory suggests that behavioral attitudes determine behavioral intentions, and expected behavioral outcomes and outcome assessments in turn determine behavioral attitudes [11]. Subsequent studies have developed the multi-attribute attitude theory to refine the theory of rational behavior. The rational behavior theory considers behavioral intention as a direct determinant of behavior, which is influenced by behavioral attitudes and subjective norms [12]. Since the rational behavior theory assumes that individual behavior is controlled by will, which seriously limits the wide application of the theory, in order to expand the scope of application of the theory, on the basis of the rational behavior theory, later studies added the variable of perceived behavioral control and initially proposed the theory of planned behavior.

Theory of Planned Behavior: The Theory of Planned Behavior (TPB) is a basic framework widely used in different product areas to explain or predict consumer behavior [13], which provides the basic theoretical background for this study. The Theory of Planned Behavior (TPB) is the best-known theory of attitudinal–behavioral relationships in social psychology, which posits that behavioral intentions are the most direct influence on behavior, which in turn is influenced by attitudes, subjective norms, and perceived behavioral control [14]. In terms of predicting consumer behavior, some scholars have studied the consumption habits of Taiwan’s consumers using recyclable packaging in online shopping based on TPB [15]. In addition, some scholars have used TPB to study consumers’ recycling of pharmaceutical cartons [16]. At the same time, the Theory of Planned Behavior has its own limitations, with factors such as attitudes, subjective norms, and perceived behavioral control not achieving sufficiently desirable predictive effects in some cases [17]. As a result, some scholars have made targeted extensions to the TPB [18]. For example, some scholars added new variables to the traditional TPB when studying consumers’ recycling behavior of electric bicycles [19], one study analyzed the influencing factors behind household food waste composting intentions [20], and environmental awareness plays an important role in influencing consumer attitudes [21].

2.2. Research Hypothesis

Recycling Attitude ATT: refers to consumers’ attitudes towards the value, process, and benefits of recycling behavior of e-commerce agricultural products’ recyclable packaging. Consumers generally prefer environmentally friendly packaging products [22]. If consumers recognize that e-commerce agricultural product recycling packaging is an environmentally friendly packaging product, consumers’ willingness to recycle may increase [23]. However, it has also been noted in the literature that consumer attitudes do not influence specific recycling behaviors [21]. In this paper, we measure consumers’ recycling attitudes in terms of concern (ATT1) and environmental benefits (ATT2). Therefore, based on the original TPB model, this paper proposes the following hypothesis:

H1: 

Consumers’ attitude towards recycling has a positive effect on consumers’ willingness to recycle.

Subjective norm SN: the external social pressure felt by the public to perform a certain behavior [24]. It can be categorized into two types: peremptory norms and exemplary norms [14]. Consumers’ subjective norms have a positive effect on recycling intentions; others’ behavioral decisions also have a significant demonstrative effect on consumers [25]. Meanwhile, subjective norm also has a positive influence on the green consumption behavior of the general public [26]. In this paper, we measure consumers’ subjective norms in terms of preferential policies of e-commerce platforms (SN1) and social atmosphere of packaging recycling (SN2). Based on this literature review, this study proposes the following hypothesis:

H2: 

Consumers’ subjective norm has a positive effect on consumers’ willingness to recycle.

Extended TPB variables: In general, the original TPB framework provides the theoretical foundation, while this paper further determines the extent of influence of relevant influencing factors by adding new variables in addition to the original model. Therefore, we added the 3 variables of packaging performance, perceived benefits, and policy promotion to the extended model. Meanwhile, for consumers’ recycling willingness RW to participate in recyclable packaging of e-commerce agricultural products, the three dimensions of willingness to participate (RW1), economic subsidy (RW2), and convenience (RW3) were introduced to measure consumers’ recycling willingness.

Packaging performance PP: refers to the various performance properties of packaging; there are two main aspects: one is the natural function, that is, to protect the goods [27], and the other is the social function, i.e., it acts as a medium for the goods to expand the market. Product packaging is an important attribute in consumers’ behavior of choosing products [28]. In China, there are problems with packaging logistics, such as food spoilage, environmental pollution, and waste of resources [29]. It has been suggested that some specific packaging characteristics can facilitate or hinder packaging recycling [30]. And the enhancement of packaging performance may make consumers more dependent on packaging [31], and they may be more willing to leave quality recyclable packages for their own use than to send them back to the recycling process. In this paper, we look at the shelf-life performance (PP1) [32], environmental performance (PP2), willingness to self-use (PP3), and food safety performance (PP4) to measure the packaging performance of recyclable packaging. Therefore, the following hypothesis is proposed in this study:

H3: 

The packaging performance of recyclable packaging has a negative impact on consumers’ willingness to recycle.

Perceived benefits (PB): refers to the consumer’s desire to realize a certain customer value in the transaction process. The higher the perceived benefits brought to consumers by participating in the recycling of agricultural products, the more willing consumers are to actively participate in the recycling of packaging [33]. Some studies have found that increasing consumers’ psychological ownership can increase their willingness to recycle [34]. In this paper, the perceived benefits to consumers are measured in terms of economic benefits (PB1), psychological satisfaction (PB2), and environmental benefits (PB3). Therefore, the following hypothesis is proposed in this study:

H4: 

There is a significant positive effect of perceived consumer benefits on consumers’ willingness to recycle.

Policy promotion PPG: refers to the promotion of the guidelines and policies formulated by the state and political parties, so that they can be quickly understood by the general public and turned into conscious actions of the masses. Some studies have found that information positioning and media type positively influence consumers’ willingness to recycle. Governments and companies are also able to promote recyclable packaging by revising their management systems [35]. Therefore, government policy implementation and media policy promotion may have a significant positive impact on consumers’ willingness to recycle [36]. In this paper, the government’s policy promotion is measured in terms of the introduction of the policy (PPG1) and the need for promotion (PPG2). Therefore, the following hypothesis is formulated in this study:

H5: 

There is a significant positive effect of government policy promotion on consumers’ willingness to recycle.

The final system of influencing factor indicators is shown in

Table 1. Impact factor indicator determination.

Latent Variable	Specified Variable	Notation
Recycling attitude	Level of concern	ATT1
	Environmental benefits	ATT2
Subjective norm	E-commerce platform incentives	SN1
	Packaging recycling social climate	SN2
Recycle willingness	Willingness to participate	RW1
	Economic subsidies	RW2
	Convenience	RW3
Packaging performance	Shelf life	PP1
	Environmental performance	PP2
	Willingness to use	PP3
	Food Safety Performance	PP4
Perceived benefits	Economic gain	PB1
	Psychological satisfaction	PB2
	Environmental benefits	PB3
Policy promotion	Introduce policies	PPG1
	The need for promotion	PPG2

.

2.3. Theoretical Model

In the field of e-commerce agricultural product recycling and packaging recovery, the factors affecting consumers’ willingness are quite complex, being influenced by internal and external aspects. Based on the perceived risk theory, pro-environmental behavior theory, and SERVQUAL theory, we studied the factors affecting Chinese consumers’ participation in e-commerce agricultural product recycling and packaging recovery. Based on the above discussion, this paper proposed ATT, SN, PP, PB, PPG, and RW and constructed a theoretical model of the influencing factors of consumers’ willingness to participate in the recycling of e-commerce agricultural products’ recycling packages, as shown in Figure 1.

3. Research Methodology

The purpose of this study was to explore the factors influencing consumers’ willingness to participate in the recycling of e-commerce agricultural products’ recyclable packaging. The data collection method used is an online questionnaire (see Supplementary Materials). The questionnaire of this study is divided into three parts: introduction of the concept of e-commerce agricultural products’ recyclable packaging, demographic characteristics of the respondents, and the items of measurement.

3.1. Data Collection

The focus of this study is to explore the factors influencing consumers’ willingness to participate in e-commerce agricultural product recycling and packaging recovery behaviors, and therefore Harbin, Heilongjiang Province, China, was chosen as the data collection site for the following reasons:

• Heilongjiang province is the top province of grain production in China [37]. According to China’s National Bureau of Statistics, in 2017, Heilongjiang province produced about 60.188 million tons of grain, accounting for 9.74% of the total grain production in the country. The per capita level of arable land is also much higher than that of other regions, and correspondingly, Heilongjiang Province has a large and representative volume of agricultural e-commerce transactions.
• Harbin, as the capital of Heilongjiang Province, has a relatively developed economy, a higher level of public consumption, relatively more consumers of e-commerce agricultural products, and a relatively developed recycling economy. The overall quality of consumers and their awareness of packaging recycling is more advanced.
• Harbin is part of the “14th Five-Year Plan” period of being a “no waste city”. This involves the implementation of a recycling development model, relying on express delivery enterprises to establish express packaging recovery, utilizing a recycling cooperation model, strict control of excessive packaging, and use of secondary packaging.

The Question-Star application is an online questionnaire collection program that is more commonly used in China and is often used for data collection. The questionnaire in this paper uses a 5-point Likert scale in order to maintain a high level of reliability and validity [38], from 1 (totally disagree) to 5 (totally agree); after data collection and screening of valid questionnaires, a total of 848 questionnaires were collected, of which 659 were valid questionnaires, with a valid questionnaire rate of 77.7%. A table of the statistical characteristics of the respondents’ demographics is shown in

Table 2. Statistical characteristics of respondents’ demographics.

	Options	Sample Size	Percentage (%)
Gender	male	231	35.1
	female	428	64.9
Age	under 20	28	4.2
	21–30 years	100	15.2
	31–40 years	169	25.6
	41–50 years	216	32.8
	51–60 years	128	19.4
	60 years and over	18	2.7
Educational attainment	elementary and secondary (and below)	125	19.0
	high school	116	17.6
	junior college	171	25.9
	undergraduate course	222	33.7
	master’s degree or above	25	3.8
Careers	retirees	51	7.8
	party and government organizations	49	7.4
	students/enterprises	208	31.6
	nationalized business	109	16.5
	foreign/joint venture/private enterprise	44	6.7
	freelance	198	30.0
Family size	1 person	31	4.7
	2–3 persons	471	71.5
	4–5 persons	140	21.2
	6–7 persons	14	2.1
	8 or more	3	0.5
Gross monthly household income	less than USD 417.9	194	29.4
	USD 417.9–696.5	188	28.5
	USD 696.5–975.1	108	16.4
	USD 975.1–1253.7	67	10.2
	USD 1253.7–1532.3	45	6.8
	USD 1532.3 or more	57	8.6

.

The following measures were used to ensure the representativeness and validity of the sample:

• We selected adults aged 18 and above as the survey respondents because this group has the ability to make independent consumption decisions and is the main consumer group of e-commerce agricultural products. Respondents were asked to have at least 1 e-commerce agricultural product purchase experience in the past 6 months as a way to ensure that the survey respondents have actual experience with e-commerce agricultural product consumption and can provide valuable feedback. In the specific process of questionnaire distribution, we mainly used the Questionnaire Star platform. Social media such as WeChat and Weibo were utilized to release the questionnaire link with a brief introduction of the study and participation guidance to attract the target population to participate.
• When designing the questionnaire, in order to avoid sample selection bias, we clearly informed the respondents of the purpose, significance, and general flow of the study at the beginning of the questionnaire, and we emphasized the anonymity of the questionnaire filling in and the confidentiality of the data, so as to eliminate the respondents’ concerns and improve the enthusiasm of participation and the authenticity of the answers. At the same time, through the setting of the questionnaire star platform, each IP address can only fill out the questionnaire once to prevent the impact of repeated responses on the data quality.

When using partial least squares structural equation modeling for research, a reasonable sample size should be ten times the latent variables [39]; there are 21 measurement items in this study, and a sample size > 210 is considered reasonable. The sample size in this study was 659, which is more significant than 210, indicating that our sample size is reasonable.

In terms of the basic characteristics of the respondents, 64.9% of the respondents were female compared to 35.1% of the respondents being male; the age distribution was dominated by the age groups with stronger purchasing power of 31–40 years old (25.6%) and 41–50 years old (32.8%), and the distribution of education was dominated by the undergraduates (33.7%) and junior colleges (25.9%). The distribution of monthly household income levels was relatively even, with respondents whose occupation was student/institution accounting for 31.6% of the total sample, followed by freelancers, etc., and staff of state-owned enterprises, accounting for 30.0% and 16.5% of the total sample, respectively.

Out of the total respondents, about 63.7% of the respondents said that they are the main members of the household who shop for e-commerce agricultural products. This suggests that most of the consumers who participated in the survey have a greater influence on purchasing decisions in their households, and therefore their attitudes and behaviors may play a key role in household consumption patterns. As high as 84.8% of the respondents indicated that they were very aware of recycled packaging for e-commerce agricultural products. This high percentage indicates that consumers generally have a high awareness of recycled packaging, which may be attributed to media publicity, platform education campaigns, or consumer self-education. A total of 89.8% of consumers in the data said they support the use of recycled packaging for e-commerce agricultural products. This not only reflects consumers’ positive attitude towards environmentally friendly packaging solutions but may also indicate a general concern for sustainability and environmental issues. A total of 90.2% of the survey respondents support participating in recyclable packaging recovery for e-commerce agricultural products equipped with recycled packaging. This percentage is even higher than the support for recyclable packaging recovery itself, indicating that when consumers understand and support recyclable packaging, they are also more willing to actively participate in related recycling activities. The results of the survey show that consumers not only have a good understanding of and support for recyclable packaging, but also most of them are willing to participate in recyclable packaging recovery.

Table 3. Consumer recycling intention survey form.

	Options	Sample Size	Percentage (%)
Are you a key member of the household shopping for e-commerce produce?	Yes	422	63.7
	No	237	36.3
Do you know much about recycled packaging for e-commerce agricultural products?	Yes	559	84.8
	No	100	15.2
Do you support the adoption of recycled packaging for e-commerce agricultural products?	Yes	592	89.8
	No	67	10.2
Are you in favor of participating in recyclable packaging recovery for e-commerce agricultural products?	Yes	594	90.2
	No	65	9.8

demonstrates the results of the survey about consumers’ willingness for recyclable packaging recovery of e-commerce agricultural products.

3.2. Structural Equation Modeling by Partial Least Squares Method

Partial Least Squares Structural Equation Modeling (PLS-SEM) is a traditional SEM treatment often used to explore the relationship between dependent and latent variables and then search for new theories or models [40]. PLS-SEM has the advantage of being able to be used under a variety of conditions, regardless of sample size and data normality; therefore, it can derive initial hypotheses in studies with small samples [41]. This technique is also frequently used in assessing consumer willingness in [42]. In this paper, SPSS 26, Smart PLS 4 software for data analysis, and the PLS-SEM were used to test the validity of the proposed hypotheses on the factors influencing the willingness of consumers to participate in the recycling and packaging recovery of e-commerce agricultural products. The structural equation model in the PLS-SEM hypothesis testing was constructed by checking the path coefficients and R² values in the model containing 21 items in 6 structures, as well as explaining the strength and direction of the influence relationship between variables. In this paper, the 3000 times resampling technique was used to calculate the data statistics.

3.3. Analysis of Results

The results of the model measurements in this paper are as follows: in terms of reliability, both Cronbach’s alpha and composite reliability values were greater than 0.7, indicating that the reliability was acceptable [43]. The results are shown in

Table 4. PLS-SEM model confidence analysis.

	Cronbach’s Alpha	Composite Reliability (CR)
ATT	0.704	0.870
SN	0.724	0.878
PP	0.861	0.906
PB	0.820	0.893
PPG	0.840	0.926
RW	0.872	0.922

, where it can be obtained that the Cronbach’s alpha value was 0.963, while the Cronbach’s alpha values based on the standardized terms were all greater than 0.7, and the composite reliabilities (CR) were all greater than 0.7, which indicates that the data had a high degree of internal consistency and were highly reliable.

The results of KMO and Bartlett’s test are shown in

Table 5. KMO and Bartlett’s test for the model.

KMO and Bartlett’s Test	KMO Quantity of Sample Suitability	Bartlett’s Test of Sphericity	Degrees of Freedom (df)	Significance (p-Value)
Value	0.966	7603.138	120	0.000

, comparing the value of the sampling suitability test (KMO) with 0 and 1: the closer to 0, the lower the validity; the closer to 1, the higher the validity. The overall KMO coefficient was more than 0.8, and the significance (p-value) was less than 0.01, which indicates that the model had high validity.

Average variance (AVE) is used to measure convergent validity [44]. The threshold of AVE should preferably be 0.5 or greater, and the loadings for convergent validity should be higher than 0.7.

Table 6. Convergent validity of the model.

Variable	Measurement Item	Factor Loadings	VIF	AVE	R2
ATT	ATT1	0.855	1.420	0.771	0.435
	ATT2	0.901	1.420
SN	SN1	0.868	1.475	0.783	0.430
	SN2	0.902	1.475
PP	PP1	0.868	2.378	0.708	0.668
	PP2	0.853	2.171
	PP3	0.769	1.753
	PP4	0.872	2.393
PB	PB1	0.859	1.920	0.735	0.572
	PB2	0.878	2.023
	PB3	0.835	1.660
PPG	PPG1	0.927	2.102	0.862	0.756
	PPG2	0.930	2.102
RW	RW1	0.876	2.055	0.797	0.734
	RW2	0.905	2.633
	RW3	0.896	2.498

shows the results of Average Extracted Variance (AVE), where all the values were higher than 0.5. It also shows the loadings of the variables, all of which were higher than 0.7. Therefore, all the items mapped well to the model, and the model had good convergent validity.

The covariance test can be analyzed by determining the Variance Inflation Factor (VIF value), which tests the linear relationship between variables [45]. When the VIF value is greater than 5, the problem of covariance exists. In this study, the maximum VIF value in the model was 2.633 (RW2) and the minimum VIF value was 1.420 (ATT1, ATT2), as shown in Table 6. The variables in the model had low VIF values, and there was no problem of multicollinearity.

The R² value is an important indicator of the predictive power of the path model, and it is important to ensure that the R² value is high enough to describe the structural equation model in this paper; thus, the R² value should be greater than 0.26 in order to have sufficient validity to explain the model. The R² values for the data studied in this paper are shown in Table 6, wherein they met the requirements of this paper for the predictive power of the structural equation model.

Distinguishing validity measures the extent to which one construct is different from the others [46]. Criteria suggest that particular variables should be related to their own items rather than to other variables [47]. For this reason, the values on the diagonal, i.e., the square root of AVE, as shown in

Table 7. Results of correlation analysis between variables.

	ATT	PB	PP	PPG	RW	SN
ATT	0.878
PB	0.718	0.857
PP	0.750	0.818	0.842
PPG	0.660	0.757	0.699	0.928
RW	0.677	0.790	0.744	0.792	0.893
SN	0.718	0.710	0.683	0.656	0.699	0.885

, should be greater than the correlation between the constructs. Table 7 shows the correlation matrix, thus confirming the discriminant validity. The dimensions in this study have good discriminant validity and are within acceptable limits.

This study analyzed the degree of fit of PLS-SEM, verified the significance of the research hypothesis, and obtained the following results.

This study used the canonical fit index (NFI) [48] and the absolute fit index (SRMR) to test model fit [49]. The value of SRMR was less than 0.08, and the value of NFI was greater than 0.8, indicating that the model fit the data well [50]. According to the test results, the SRMR value of the model was 0.055 (<0.08); the NFI value was 0.808 (>0.8); and the results of the model indicator test were all within a reasonable range, as shown in

Table 8. Model fitting test results.

	Benchmark Value	Saturated Model
SRMR	<0.08	0.055

, which indicates that the data of the research model can be focused on one variable, and the model was well fitted.

The path coefficients of the model of this study are shown in

Table 9. Path coefficients and effect sizes in the model.

H	Path	Path Coefficients	Indirect Effects	p-Values
H1	ATT-RW	0.022		0.559
H2	SN-RW	0.148		0.000
H3	PP-RW	0.145		0.001
	PB-PP	0.818		0.000
H4	PB-RW	0.263		0.000
	PPG-ATT	0.660		0.000
H5	PPG-RW	0.380		0.000
	PPG-PB	0.757		0.000
	PPG-SN	0.656		0.000
	PPG-PB-PP		0.619	0.000
	PPG-ATT-RW		0.014	0.562
	PPG-PB-PP-RW		0.090	0.001
	PPG-SN-RW		0.097	0.000
	PPG-PB-RW		0.199	0.000
	PB-PP-RW		0.119	0.001

, and the summary of the hypothesis testing is shown in

Table 10. Results of hypothesis testing.

Grade	Research Hypothesis	Result
H1	Consumer attitudes toward recycling have a positive impact on consumers’ willingness to recycle	not supported
H2	Consumers’ subjective norm has a positive impact on consumers’ willingness to recycle	supported
H3	Packaging performance of recyclable packaging has a negative impact on consumers’ willingness to recycle	not supported
H4	Consumers’ perceived benefits have a significant positive impact on consumers’ willingness to recycle	supported
H5	There is a significant positive impact of government policy promotion on consumers’ willingness to recycle	supported

.

The p-value less than 0.05 indicates that the data were statistically significant, and the path coefficient of recycling attitude on recycling willingness was 0.022, which was almost nonexistent, being inconsistent with the hypotheses of this paper. H1 was negated, in contrast to the path coefficient of subjective norm (0.148), which had a higher degree of impact and a more significant impact on recycling willingness. H2 was verified, packaging performance (0.145) had a stronger influence on recycling willingness, and the two were positively correlated, which is inconsistent with H3, perceived benefits (0.263) had a more positive influence on recycling willingness. The analysis supports H4, with the path coefficient of policy promotion (0.380) being the highest on recycling willingness, and the two were positively correlated in a significant way, which supports H5. In summary, H2, H4, and H5 were supported, and H1 and H3 were negated.

According to the results of the analysis, the positive influence of the above five influencing factors on the willingness to recycle was as follows in descending order: PPG > PB > SN > PP > ATT, among which PPG had the strongest influence, SN and PP had almost the same degree of influence, and the positive influence of ATT on RW was extremely weak. The structural equation model and path coefficients are shown in Figure 2.

4. Discussion

By analyzing the results, consumers’ willingness to recycle the recyclable packaging of e-commerce agricultural products in this study was influenced by five factors: policy promotion, perceived benefits, subjective norm, packaging performance, and recycling attitude. Comparing the effects of various variables on consumers’ willingness to recycle, it was found that policy promotion, perceived benefits, and packaging performance were the main influencing factors. Attitude toward recycling had a weak positive effect on the willingness to recycle.

The results of hypothesis validation indicate that policy promotion significantly affects consumers’ willingness to recycle. Recycling of recyclable packaging is closely related to environmental protection issues and can reduce resource waste. When people receive more policy promotion, they are more likely to participate in e-commerce agricultural products recycling of recyclable packaging. Previous studies have emphasized the positive impact of policy and media publicity on consumers’ willingness to recycle [51] and policy promotion as a major factor influencing willingness to recycle [52]. This is consistent with the results of this study that government decrees related to the promotion of recyclable packaging are conducive to strengthening consumers’ willingness to recycle by enhancing publicity in various social media. And there is almost no effect of recycling attitude on recycling willingness, which is inconsistent with some previous studies [53], but it has also been mentioned that recycling attitudes do not map to specific recycling behaviors [54].

The relationship between perceived benefits on willingness to recycle is positive and significant. This is consistent with the findings of previous studies [55]. Consumers’ perceived benefits are greatly influenced by economic benefits, psychological satisfaction, and environmental benefits, and increasing the economic benefits of consumers’ participation in recyclable packaging recovery and enhancing the psychological satisfaction and environmental benefits brought by recycling behaviors are conducive to enhancing consumers’ perceived benefits, with past studies having also found that consumers’ economic benefits will increase their willingness to recycle [56], increasing consumers’ concern for recyclable packaging [57]. The environmental benefits of enhanced recycling behavior are conducive to enhancing consumers’ perceived benefits.

Subjective norm has a positive and positive effect on willingness to recycle [58]. Among them, the contribution of e-commerce platform preferential policies and packaging recycling social atmosphere to subjective norm are both significantly positively correlated, indicating that consumers’ subjective norm is more influenced by the preferential policies of e-commerce platforms and packaging recycling social atmosphere. And subjective norm has a positive impact on consumers’ willingness to recycle, which is consistent with the results of previous studies [59] E-commerce platforms to strengthen preferential policies and enhance the social atmosphere of recyclable packaging recovery are conducive to strengthening consumers’ subjective norm [60]. The strengthening of consumers’ subjective norm can bring positive benefits for the increase in consumers’ recycling willingness.

Packaging performance significantly affects consumers’ willingness to recycle, which is an opposite result to the hypothesis of this study, and there are few previous studies on the link between packaging performance and consumers’ willingness to recycle [61]. Packaging performance has mostly been used only to study issues related to packaging and economics [62], and in the circulation of e-commerce agricultural recyclable packaging, recyclable packaging with higher shelf-life performance, environmental performance, and food safety performance may be more important to consumers than putting these packages into the recycling process, and perhaps they would prefer to leave them behind. However, the results of the study show that this is not the case, and that the improvement of recyclable packaging performance can have a very strong positive impact on the increase in consumers’ willingness to recycle. This may be because although the high performance of recyclable packaging increases consumers’ willingness to keep them, there is a limit to the number of packages that can be retained by each household, and most of the recyclable packages will still be involved in the logistics cycle. Therefore, it is necessary to study how to improve the performance of recyclable packaging. Packaging designers should pay attention to the application of green materials in recyclable packaging [61]. Improving the preservation and environmental and food safety performance of recyclable packaging is beneficial to improve the packaging performance of recyclable packaging. In order to increase consumers’ willingness to recycle, recyclable packaging producers can also improve the overall quality of the output packaging accordingly to gain more economic benefits.

In previous studies, consumer recycling attitude was described as a major factor influencing consumers’ willingness to recycle. However, in the extended TPB model of this study, consumers’ willingness to participate in recycling packaging recovery of e-commerce agricultural products seemed to be less affected by recycling attitude than expected, probably because consumers’ participation in recycling behavior of recycled packaging is itself more significantly affected by other factors, such as policy advocacy, subjective norm, and perceived benefits, relative to recycling attitude.

5. Conclusions and Limitations

5.1. Conclusions

In this study, policy promotion, perceived benefits, and packaging performance are introduced into the study of consumers’ willingness to participate in e-commerce agricultural product recycling and packaging recovery, and an extended TPB model is proposed to study the influencing factors of consumers’ willingness to participate in e-commerce agricultural product recycling and packaging recovery and the degree of their influence. The main conclusions are as follows:

• The government’s policy promotion has the most important role in promoting consumers’ willingness to participate in e-commerce agricultural product recycling packaging recovery, followed by consumers’ perceived benefits. For the willingness to recycle, the stronger the government’s policy promotion is, the stronger consumers’ willingness to recycle; along with the improvement of the perceived benefits, the consumers’ willingness to recycle is also improved.
• Consumers’ subjective norm and the packaging performance of recyclable packaging are also positively correlated with consumers’ willingness to participate in the recycling behavior of e-commerce agricultural products.
• Consumers’ attitude towards recycling has the least impact on willingness to recycle.

5.2. Practical Significance

This study argues that promoting the use of recyclable packaging requires giving full play to the guiding role of the government, increasing the publicity and interpretation of relevant policies issued by the government and other relevant policies by the media, and improving the understanding of consumers of the policies related to recyclable packaging and recovery, as well as creating a strong social atmosphere in which consumers are actively involved in recyclable packaging and recovery, alongside the protection of the urban ecological environment, in order to increase the perceived benefits and subjective norm.

This study initially thought that the improvement of packaging performance of recyclable packaging has a certain negative impact on consumers’ willingness to recycle, and that excessive packaging quality may reduce consumers’ desire to recycle and keep them for their own use; however, according to the results of the study, the improvement of packaging performance has a significant positive impact on consumers’ willingness to recycle, and the recyclable packaging manufacturers can improve the packaging performance of their output recyclable packaging in order to increase consumers’ willingness to recycle, which will bring more profits to the company and also have a good positive effect on the social ecology.

The promotion and use of recyclable packaging in the long run will certainly bring about the enhancement of the overall social benefits, reduce the carbon emissions of express logistics, and increase consumer awareness of environmental protection, not only in line with China’s national policy, but also to make up for the current widespread use of ordinary disposable packaging on the express logistics industry in the future of the shortcomings.

5.3. Limitations

This study has the following shortcomings:

• During the period when this study was conducted, the promotion of recyclable packaging recovery for e-commerce agricultural products was still in the initial stage, and the use of recyclable packaging was also in the stage of preliminary attempts in terms of policy and the development of production specifications by enterprises, which made the data collection of this study difficult, and although the sample size of the model of this study is reasonable, there may still be a slight prediction bias.
• The background of this study is based on the use and promotion of recyclable packaging recycling for agricultural e-commerce under China’s dual-carbon policy. At present, recyclable packaging is not yet widely used, and recyclable packaging is relatively unique in terms of its own performance and means of recycling, so that after recyclable packaging is popularized in the future, the conclusions and recommendations of this study may not be fully applicable to the recyclable packaging recycling model that will generate new changes in the future.
• In this study, the questionnaire on the willingness to recycle packaging for e-commerce agricultural products was only distributed in Harbin, Heilongjiang Province, and the sample in this region could not fully represent the characteristics of consumers nationwide, which limits the general applicability of the study’s conclusions. By distributing the questionnaire through Questionnaire Star, participating consumers may themselves be more interested in environmental protection or packaging recycling, resulting in a higher response from the sample in favor of recycling. Furthermore, only policy promotion, perceived benefits, and packaging performance were additionally introduced, while no other variables were introduced. In future studies, one can try to introduce new variables and study the internal correlation between them.
• This study analyzed the influencing factors of e-commerce agricultural products consumers’ willingness to participate in recycling packaging recovery, but did not go further to study the actual recycling behavior of consumers, and although the willingness to recycle is considered to be an important predictor of consumers’ recycling behavior, there is still a discrepancy between these two variables, so future research can try to continue to improve the TPB extension model of this study, including the addition of some other relevant variables, or by other methods (e.g., binary logistics regression analysis, hierarchical analysis, and factor coding) to study consumers’ willingness to participate in the recycling of e-commerce produce recycling packaging.