What Influences the Public to Work as Crowdshippers Using Cargo Bikes? An Extended Theory of Planned Behavior

Abstract

Driven by the green and low-carbon transformation of urban logistics, the integration of crowdsourced delivery and green transportation is considered an important pathway to achieving sustainable last-mile delivery. This study focuses on urban crowdsourced delivery using cargo bikes and develops an extended behavioral model based on the Theory of Planned Behavior (TPB). The model systematically examines the key factors influencing the public’s behavioral intention (BI) to participate as crowdshippers. While retaining the core structure of TPB, the model incorporates external variables—perceived risk (PR), policy support (PS), and infrastructure conditions (IC)—to improve its explanatory power and applicability to real-world delivery scenarios. A questionnaire survey was conducted in South Korea, yielding 600 valid responses. The results indicate that usage attitude and perceived behavioral control exert significant positive effects on BI. PR has a significant negative effect on both attitude and BI. PS indirectly enhances BI by improving attitudes, whereas IC primarily influences BI by strengthening the public’s sense of control. This study not only expands the theoretical explanatory power of the TPB model in the context of green crowdsourced delivery but also provides empirical evidence for policymakers and platform operators.

1. Introduction

With the increasing demand for last-mile delivery, urban logistics is under mounting pressure. Traffic congestion, carbon emissions, and safety hazards associated with conventional internal combustion engine freight vehicles have become major obstacles to the development of sustainable urban logistics systems [1,2]. To address these challenges, governments and industries are actively exploring greener and more efficient transportation modes. Cargo bikes have been recognized as a promising alternative for last-mile delivery in urban areas due to their low carbon emissions, operational flexibility, and ability to reduce traffic congestion [3,4].

Despite the environmental and operational advantages of cargo bike delivery, its low load capacity necessitates more frequent trips, thereby increasing the demand for delivery personnel. Given the high labor costs of hiring full-time couriers and the current shortage of available drivers, these structural limitations continue to hinder the widespread adoption of cargo bike delivery in the mainstream logistics market.

To address this challenge, crowdsourced delivery platforms can serve as a viable alternative. Supported by the platform economy, this approach significantly enhances delivery flexibility and efficiency [5]. The concept of crowdsourced delivery was first introduced by Howe [6], who described it as the transfer of tasks traditionally performed by employees to unspecified members of the public in an open manner. Participants in crowdsourced delivery generally fall into four categories: parcel delivery companies that outsource delivery tasks, platform operators who connect participants and manage operations, drivers who carry out deliveries, and end customers who receive the goods. A schematic representation of these relationships is provided in Figure 1.

Unlike traditional distribution, the crowdsourcing model does not rely on hired professional drivers but instead uses platforms to match delivery tasks with voluntary public participants (i.e., crowdshippers). These participants are not salaried employees but independent delivery service providers, characterized by self-selected tasks, flexible hours, personal equipment, and pay-per-order arrangements [5,7].

Meanwhile, combining platform-based crowdsourced delivery with cargo bikes could help reduce operating costs and improve service coverage. Additionally, this approach does not require increasing the number of vehicles or motorcycles in urban areas, which often leads to traffic congestion and environmental problems as the service expands. However, some questions remain, such as whether crowdshippers will participate as carriers and whether cargo bikes are a suitable mode of transportation for general last-mile delivery.

Most previous research has focused on consumers’ or residents’ attitudes toward cargo bikes or other green transportation modes [8,9]. In addition, although some studies have applied the Technology Acceptance Model (TAM) and the Theory of Planned Behavior (TPB) to explain people’s acceptance of green transportation tools [10,11], no systematic study has yet combined TPB with contextual variables, such as policy and infrastructure, to develop an explanatory framework for the willingness to adopt cargo bikes in crowdsourced delivery contexts.

This study focuses on the Korean public and constructs an extended TPB model. Key external variables, including perceived risk (PR), policy support (PS), and infrastructure conditions (IC), are incorporated into the analysis. The objective is to examine how these factors influence the public’s willingness to participate as crowdfunding users and to use cargo bikes. Based on this framework, the study proposes the following three specific research questions (RQs):

RQ1: 

Are the public willing to participate in crowdsourced delivery services using cargo bikes?

RQ2: 

What role does perceived risk (PR) play in shaping the public’s attitude toward using cargo bikes (ATU) to participate in crowdsourced delivery?

RQ3: 

How do external contextual variables indirectly influence the public’s intention to become crowdshippers?

The main contributions of this research are threefold. First, by incorporating PS, IC, and PR into the TPB framework, the explanatory power of the structural equation modeling (SEM) approach is enhanced in the context of green crowdsourced delivery. Second, the study addresses a gap in the literature regarding the behavioral mechanisms of the public as crowdshippers using cargo bikes. Third, the research highlights the crucial role of institutional incentives and infrastructure improvements in promoting public participation, offering policy recommendations for both policymakers and logistics platforms to enhance engagement in green crowdsourced delivery.

The remainder of this study is organized as follows: Section 2 reviews the relevant literature; Section 3 presents the proposed research model; Section 4 reports the analysis results; Section 5 discusses the findings; and Section 6 concludes the study.

2. Literature Review

2.1. Public Participation in Crowdsourced Delivery

Compared with traditional logistics modes, crowdsourced delivery enhances distribution flexibility by enabling crowdshippers to perform last-mile tasks. This service also provides the general public with opportunities to earn part-time income.

Table 1. Summary of research on public participation in crowdsourced delivery.

Author	Research Region	Methodology	Perspective
Le and Ukkusuri [12]	United States	Survey-based analysis	Potential crowdshippers
Cao [13]	China	Expert interviews, case studies	Public participation policy
Upadhyay, et al. [14]	India	SEM	Willingness to participate in crowdsourcing
Nguyen, et al. [15]	Vietnam	SEM	Intention to keep working as a crowdshipper
Fessler, et al. [16]	Denmark	TPB	crowdshippers
Mohri, et al. [17]	Austrailia	Discrete choice model	Crowdshippers using parcel lockers
Mohri, et al. [18]	Austrailia	Discrete choice models	Crowdshippers using parcel lockers

summarizes previous research on public participation in crowdsourced delivery.

Numerous studies have shown that individuals’ basic demographic characteristics significantly influence their likelihood of participating in crowdsourced delivery. Commonly examined variables include age, driving experience, and income, which are often used to describe the structural characteristics of potential crowdshipper groups [16,19]. Punel et al. [20] found that men, low-income individuals, and full-time employees are more likely to use crowdsourced delivery services. Le and Ukkusuri [12] further reported that individuals with driving ability and flexible working hours are more likely to become high-frequency crowdshippers. Fessler et al. [19] observed that students exhibited the highest willingness to participate due to their strong time flexibility. Maleki et al. [21] confirmed that young people with flexible working hours who seek to supplement their income are most likely to become high-frequency crowdshippers.

Delivery behavior is also strongly influenced by individuals’ psychological perceptions and intrinsic motivations. Several studies have shown that platform transparency and the level of user trust significantly increase the public’s willingness to participate. Work autonomy is another key driver encouraging public engagement in crowdsourced delivery [15,22]. Additionally, individuals with a strong sense of social identity are more likely to participate as crowdshippers. Fessler et al. [19] found that attitude, subjective norm (SN), and perceived behavioral control (PBC) can significantly predict whether individuals choose to work as crowdshippers.

Economic motivation is also a key factor driving public participation in crowdsourced delivery. Le and Ukkusuri [12] found that different types of delivery tasks significantly influence the public’s willingness to participate and the associated time–income trade-off. In particular, tasks with shorter delivery distances and more predictable rewards are more likely to attract part-time workers. Similarly, Le et al. [23] noted that such tasks are especially appealing to part-time workers due to their predictability and low time costs.

Numerous studies have shown that potential participants often have significant concerns regarding crowdsourced delivery. These concerns generally fall into several categories. First, the public frequently worries that liability is not clearly defined (e.g., in cases of lost or damaged packages) [24,25]. Second, they are typically concerned about the risk of traffic accidents [26,27]. Because participants are not professionals, compliance-related concerns also arise [28]. Furthermore, if the delivery task lacks insurance coverage and a clear liability mechanism provided by the platform, these concerns are significantly heightened [26].

Platform design and institutional support have also been found to play a key role in influencing the public’s willingness to become crowdshippers. Cebeci et al. [29] emphasized that a platform’s reputation mechanism is critical for building user trust. Arriagada et al. [30] found an interactive effect between user experience and platform reputation. Shuaibu et al. [31] further highlighted the importance of institutional guarantees, particularly infrastructure configurations, including non-motorized lanes and relevant policies.

2.2. Cargo Bike Adoption

As a green delivery tool, cargo bikes have increasingly been recognized as an alternative for last-mile delivery. Against the backdrop of growing urban traffic congestion, the adoption of cargo bikes has attracted widespread attention from policymakers and platform operators [32,33].

Table 2. Summary of research on cargo bike acceptance.

Author	Research Region	Methodology	Perspective
He, et al. [10]	China	TPB, Risk Homeostasis Theory, Risk-Adaptation Theory	Riders
Ye [34]	China and Estonia	SEM	Consumers
Giglio and Maio [35]	Southern Italy	Innovation Diffusion Theory (IDT)	Consumers
Mehmood and Zhou [36]	China	TPB, Goal-Framing Theory, Value–Belief–Norm theory	Eco-conscious tourists
Garcia-Pajoy, et al. [37]	Colombia	SEM, Km2 Methodology	Stakeholders including citizens, wholesalers, and retailers
Kapousizis, et al. [38]	Austria, Belgium, Germany, Greece, Netherlands	Unified Theory of Acceptance and Use of Technology 2 (UTAUT2)	Consumers
Zabiulla, et al. [11]	India	TPB	Non-electric bicycle users
Qian and Shi [39]	China	TPB	Riders

summarizes recent research on the acceptance of cargo bikes.

Several studies have found that willingness to adopt cargo bikes varies significantly across different genders, age groups, and occupational backgrounds [8,40,41]. Marincek et al. [33] reported that young urban residents with cycling experience are more likely to accept cargo bikes for short-distance deliveries. Betancur Arenas et al. [41] further highlighted that individuals with higher education and stronger environmental awareness are more inclined to adopt this type of green transportation. Additionally, Philipsen et al. [42] found that education level and environmental motivation are important factors that significantly enhance public acceptance of electric cargo bikes.

Similar to the factors influencing the public’s willingness to participate in crowdsourced delivery, PR is also a key barrier to cargo bike adoption. Several studies have reported that the public generally has concerns about cycling safety. The main issues include high traffic volume, narrow intersections, and uncertainty regarding vehicle performance. The risk associated with adverse weather conditions further exacerbates these concerns [9,43]. Additionally, studies have identified negative social image and cycling fatigue as major limiting factors, particularly in environments where safety and comfort are not assured [44].

External contextual variables are critical in determining whether cargo bikes can move from concept to practice. Chatziioannou et al. [9] found that both physical and institutional factors influence the public’s willingness to use cargo bikes. Patella et al. [45] further noted that institutional arrangements, such as government policy guidance and clear platform responsibilities, can significantly enhance public trust. Mehmood and Zhou [36] emphasized the leading role of the government in infrastructure investment and promotion, arguing that accessibility of the traffic environment and institutional guarantees jointly shape PBC.

2.3. Structural Equation Modeling

With the widespread adoption of green distribution tools, understanding the mechanisms underlying individual acceptance and adoption has gradually emerged as a key topic in the fields of transportation and behavioral sciences.

Among these methods, the Theory of Rational Action (TRA) proposed by Fishbein and Ajzen [46] is widely used to predict individuals’ behavioral intentions (BI) in social psychological contexts. Davis [47] further developed the TAM based on TRA, introducing two key variables—perceived usefulness and perceived ease of use—to explain how individuals’ cognitive evaluation of technology influences their willingness to adopt it. Subsequently, Ajzen [48] proposed the TPB, which further incorporates PBC to account for the likelihood of individual behavior in the presence of external constraints.

Venkatesh and Davis [49] proposed the TAM2 model, incorporating exogenous variables such as SN, result visibility, and social influence [50]. Subsequently, Venkatesh et al. [51] integrated TAM, TPB, and the IDT to develop the UTAUT model, identifying four core dimensions: performance expectancy, effort expectancy, social influence, and facilitating conditions. To better capture the complex behavioral characteristics of the public, Venkatesh et al. [52] further proposed the UTAUT2 model, introducing variables such as hedonic motivation, price value, and usage habits [53].

Through the literature review of the theories, it can be found that although there are many models that can explain individuals’ willingness or adoption behavior. The TPB model can offer greater structural flexibility for complex behavioral situations involving external institutional environments. Therefore, this study extends the TPB framework to investigate the public’s willingness to participate in crowdsourced delivery and adopt cargo bikes.

2.4. Research Gaps and Contributions

After reviewing previous research on crowdsourced delivery and cargo bikes, several limitations can be identified:

First, existing research on crowdsourced delivery primarily focuses on platform mechanisms, consumer behavior, or delivery efficiency. Few studies explicitly position crowdshippers as key actors in platform operations and systematically examine their willingness to participate, behavioral mechanisms, and influencing factors.

Second, although some studies have applied technology adoption models to explore the public’s acceptance of green transportation, most focus on riders or consumers. No study has analyzed the practical applicability of such green tools from the perspective of crowdshippers who actually perform delivery tasks.

Third, crowdsourced delivery services typically integrate multiple characteristics, such as platform economy, shared logistics, and individual employment. A single TAM model is insufficient to account for key external factors and cannot fully explain BI or participation pathways. Therefore, it is necessary to develop a more explanatory extended model to address these complexities.

To fill the above research gaps, this study constructs an extended path model based on the TPB model. The contributions of this study are included as follows. This study not only focuses on the behavioral psychological mechanism at the individual level, but also pays more attention to how external institutional and urban environmental factors affect the public’s green crowdsourcing delivery behavioral intentions through indirect paths. Through this framework, this paper aims to provide new theoretical ideas and empirical support for the behavioral modeling and policy optimization of green urban logistics.

3. Research Design

3.1. Model Description

As noted above, this study extends the TPB model to examine behavioral mechanisms. Given that crowdsourced delivery behavior is highly individualized, SN exerts relatively weak influence on behavioral intentions. Existing research has indicated that its predictive power is limited in informal labor contexts [10]. Consequently, this study excludes the SN variable and retains ATU, PBC, and BI. The specific variables are presented in

Table 3. Latent variables and their corresponding measurement items.

Latent Variable	Observation Variable	Variable ID	Reference
Policy support	Subsidy for purchasing e-cargo bike	PS 1	Carracedo and Mostofi [54]
	Subsidies for charging station installation	PS 2	Feng, et al. [55]
	Incentives and tax relief for using e-cargo bike	PS 3	Nosratzadeh, et al. [56]
	National training and support programs for drivers	PS 4	Masterson [57], Malik, et al. [58]
Infrastructure condition	Bike lane	IC 1	Dalla Chiara, et al. [59]
	Parking space for e-cargo bike	IC 2	Nürnberg [60]
	Charging station	IC 3	Chatziioannou, et al. [9]);
	Traffic safety environment	IC 4	Vasiutina, et al. [4]
Perceived risk	Physical injury in accidents	PR 1	Hu, et al. [61]
	Cargo theft and missing shipments	PR 2	Burges [62]
	Adverse weather conditions	PR 3	Zimmermann and Palgan [63]
	Legal or regulatory uncertainty	PR 4	Raimondi, et al. [64], Malik, et al. [58]
Attitude toward using	Positive perception	ATU 1	Vasiutina, et al. [4]
	Perceived idea attractiveness	ATU 2	Heinrich, et al. [65], Mikkelsen and Wernberg [66]
	Perceived suitability	ATU 3	Llorca and Moeckel [67], Papaioannou, et al. [68]
	Perceived environmental–social Value congruence	ATU 4	Chatziioannou, et al. [9]
	Perceived lifestyle compatibility	ATU 5	Mikkelsen and Wernberg [66]
Perceived behavioral control	Perceived infrastructure facilitation	PBC 1	Hess and Schubert [69]
	Perceived regulatory facilitation	PBC 2	Zimmermann and Palgan [63]
	Perceived environmental readiness	PBC 3	Vasiutina, et al. [4]
	Perceived procedure and incentive clarity	PBC 4	Nosratzadeh, et al. [56]
Behavioural intention	Conditional participation intention	BI 1	Narayanan, et al. [40]
	Platform registration intention	BI 2	Beyioku [70]
	Flexible work intention	BI 3	Gruber and Narayanan [8]
	Future participation intention	BI 4	Hess and Schubert [69]

. The extended TPB model was then constructed to test the path relationships illustrated in Figure 2.

3.2. Hypothesis Design

Under the consideration of the adoption characteristics of green transportation tools in crowdsourced delivery, key external variables such as PS, IC, and PR are introduced into the TPB model. There are several reasons for the proposal hypotheses. First, the institutional environment is an important external prerequisite for public adoption behavior. Existing studies have also pointed out that policy guidance has a positive impact on the public’s green travel decisions [10,71]. In the context of crowdsourced delivery, clear regulatory support, tax incentives, and transportation priority policies may enhance the public’s positive attitude towards green crowdsourced delivery.

Therefore, the following hypothesis is proposed:

H1: 

PS has a significant positive impact on ATU.

The completeness of IC not only affects the convenience of use but also influences the public’s subjective judgment regarding the behavior [32,72]. Therefore, it is proposed:

H2: 

IC has a significant positive impact on PBC.

PR may weaken the public’s positive attitude toward the behavior, or directly reduce BI [11,73]. Therefore, it is proposed that:

H3: 

PR has a significant negative impact on ATU.

H4: 

PR has a significant negative impact on BI.

According to the TPB, an individual’s behavioral attitude not only directly affects BI but may also indirectly influence it by enhancing their subjective evaluation of PBC [48]. Therefore, it is proposed that:

H5: 

ATU has a significant positive impact on PBC.

Furthermore, the TPB model emphasizes that BI is jointly influenced by ATU and PBC, a relationship confirmed by numerous studies [10,74]. Therefore, it is further proposed that:

H6: 

ATU has a significant positive impact on BI.

H7: 

PBC has a significant positive impact on BI.

4. Results of Empirical Survey

4.1. Characteristics of Respondents

This research collected data through questionnaire surveys to test the proposed hypotheses. An Institutional Review Board (IRB) exempted approval was obtained from the Incheon National University. The questionnaire was distributed online in South Korea. Prior to distribution, a pre-test was conducted, and based on feedback regarding the clarity and comprehension of the questions, the scale expressions and question order were revised. Ultimately, a valid sample of 600 respondents (N = 600) was obtained as the basis for analysis. A detailed description of the respondents is provided in

Table 4. Demographic characteristics of respondents.

Variable	Group	Frequency	Percentage
Gender	Male	300	50.00
	Female	300	50.00
Age	20~29 years	85	14.17
	30~39 years	97	16.17
	40~49 years	117	19.50
	50~59 years	127	21.17
	60~69 years	112	18.67
	70~79 years	62	10.33
Education level	High school graduate or below	119	19.83
	Junior college graduate	75	12.50
	University graduate	338	56.33
	Graduate school graduate	68	11.33
Job type	Office worker/Professional	274	45.67
	Skilled/manual or elementary worker	77	12.83
	Self-employed	49	8.17
	homemaker	89	14.83
	Student (undergraduate/graduate)	25	4.17
	Unemployed	71	11.83
	Other	15	2.50
Driver license	Yes	534	89.00
	No	66	11.00
Average yearly income (KRW)	~20 million	133	22.17
	20 million~40 million	202	33.67
	40 million~60 million	141	23.50
	60 million~80 million	73	12.17
	80 million~100 million	31	5.17
	100 million~	20	3.33
Region	Seoul Metropolitan City	153	25.50
	Incheon Metropolitan City	33	5.50
	Busan Metropolitan City	35	5.83
	Daegu Metropolitan City	19	3.17
	Daejeon Metropolitan City	18	3.00
	Gwangju Metropolitan City	13	2.17
	Ulsan Metropolitan City	13	2.17
	Sejong Special Self-Governing City	5	0.83
	Jeju Special Self-Governing Province	4	0.67
	Gyeonggi Province	171	28.50
	Gangwon Province	14	2.33
	North Chungcheong Province	15	2.50
	South Chungcheong Province	14	2.33
	North Jeolla Province	20	3.33
	South Jeolla Province	16	2.67
	North Gyeongsang Province	24	4.00
	South Gyeongsang Province	33	5.50

.

4.2. Reliability and Validity Tests

To ensure the reliability and validity of the questionnaire, a preliminary assessment of the scale was conducted. As shown in

Table 5. KMO and Bartlett’s test of sphericity.

Kaiser–Meyer–Olkin Measure of Sampling Adequacy		0.942
Bartlett’s test of sphericity	Approx. Chi-square	11,217.655
	Df	300
	Sig	<0.001

, the Kaiser–Meyer–Olkin (KMO) value is 0.942, well above the excellent threshold of 0.9 [75]. The results of Bartlett’s sphericity test indicate a chi-square value of 11,217.655 with 300 degrees of freedom and a significance level of p < 0.001, further confirming that the data are significantly correlated and suitable for factor analysis.

4.3. Confirmatory Analysis

To further verify the convergent and discriminant validity of the measurement model, this study conducted confirmatory factor analysis (CFA). As shown in

Table 6. Cronbach’s alpha values for latent variables.

Latent Variable	N of Variable	Cronbach’s Alpha Based on Standardized Items	Cronbach’s Alpha
PS	4	0.904	0.903
IC	4	0.879	0.879
PR	4	0.774	0.770
ATU	5	0.873	0.873
PBC	4	0.921	0.921
BI	4	0.949	0.949

and

Table 7. Convergent validity and construct reliability.

Latent Variable	Cronbach’s α	AVE	CR
PS	0.903	0.774	0.932
IC	0.879	0.727	0.914
PR	0.768	0.744	0.853
ATU	0.864	0.698	0.902
PBC	0.921	0.800	0.941
BI	0.949	0.799	0.941

, the Cronbach’s alpha coefficients for all latent variables exceed the commonly accepted threshold of 0.76, with BI exhibiting the highest value (0.949). Additionally, Table 7 shows that the average variance extracted (AVE) for each construct is greater than 0.5, and the composite reliability (CR) exceeds 0.85, meeting the convergent validity criteria proposed by Fornell and Larcker [76]. Furthermore, as shown in

Table 8. Discriminant validity.

	PS	IC	PR	ATU	PBC	BI
PS	0.880
IC	0.804	0.853
PR	0.343	0.508	0.862
ATU	0.798	0.598	0.167	0.835
PBC	0.728	0.614	0.242	0.804	0.895
BI	0.445	0.224	−0.008	0.580	0.578	0.894

, the square root of the AVE for each dimension is greater than its correlation with other latent variables, indicating that the scale demonstrates good discriminant validity.

4.4. Hypothesis Testing

Subsequently, the seven research hypotheses were tested. As shown in

Table 9. A summary of hypothesis testing results.

Hypothesis	Path	$\mathbf{Estimate} \left(\mathit{\beta }\right)$	S.E.	C.R.	p-Value	Testing Result
H1	PS → ATU	0.839	0.046	−3.084	***	Supported
H2	IC → PBC	0.204	0.046	17.963	***	Supported
H3	PR → ATU	−0.119	0.045	4.958	**	Supported
H4	PR → BI	−0.17	0.047	14.452	***	Supported
H5	ATU → PBC	0.685	0.067	−4.169	***	Supported
H6	ATU → BI	0.321	0.102	4.417	***	Supported
H7	PBC → BI	0.354	0.102	4.879	***	Supported

Note: ** p < 0.01, *** p < 0.001.

and Figure 3, all hypotheses were supported.

PS has a significant positive impact on ATU, and IC also has a significant positive effect on PBC. These results confirm the incentive effect of external institutional variables in the extended TPB model.

PR has a significant negative impact on ATU and BI, indicating that PR can weaken the public’s willingness to participate in crowdsourced delivery, particularly in the absence of prior experience or safety guarantees.

The three core paths in the TPB model are also supported. ATU has a significant positive effect on PBC, and both ATU and PBC have significant positive effects on BI. These findings suggest that the public’s positive perception of using cargo bikes can not only enhance their confidence in behavioral control but also increase their willingness to participate in crowdsourced delivery.

5. Discussion and Implications

5.1. Role of Perceived Risk

The results indicate that PR has a significant negative impact on the public’s decision to adopt cargo bikes for crowdsourced delivery, inhibiting both BI and ATU. This finding aligns with the results of Savas-Hall et al. [77] regarding adoption intentions for new services, further confirming the critical role of PR in technology adoption.

Moreover, PR can be understood as a psychological expectation bias. First, the public’s primary concern typically relates to traffic safety. Cargo bikes often share urban roads with motor vehicles and non-motorized vehicles, and some users perceive a higher risk of accidents during peak hours or under complex road conditions [63]. Adverse weather conditions, such as rain or low visibility, further amplify safety concerns. Second, the public is also concerned about unclear policies and regulations, including restrictions on delivery areas, driving rights, and the allocation of liability in accidents [78]. Such perceptions of institutional uncertainty can significantly reduce individuals’ sense of control and expected certainty, thereby weakening their motivation to participate.

PR may also influence attitude formation and decision-making regarding behavioral intention through indirect pathways. Arciniega-Rocha et al. [79] noted that when users perceive high levels of risk, their BI can be negatively affected, even if the tool is easy to use. In other words, the presence of risk perception can diminish the public’s positive evaluation of the ease of use and potential benefits of new tools, thereby indirectly suppressing both attitudes and behavioral intentions.

Similarly, Neves et al. [80] proposed that an increase in the perceived risk of a particular technology or behavior can offset users’ judgments regarding controllability and expected benefits. Consistent with these findings, PR not only weakens individuals’ positive attitudes toward the use of cargo bikes but also directly reduces their intention to participate in crowdsourced delivery. This suggests that, in emerging fields such as green crowdsourced delivery, the moderating effect of risk perception may be more influential than technological advantages or platform-based incentives.

5.2. Policy and Infrastructure Impact

In this study, PS primarily refers to supportive measures, including vehicle purchase subsidies and charging facility assistance. PS helps reduce the public’s perception of uncertainty during the adoption of cargo bikes, thereby enhancing ATU. More importantly, PS can also convey positive institutional encouragement at the psychological level. Through PS, the public is more likely to adopt cargo bikes as a socially accepted and legitimate form of crowdshipping. Consistent with existing research, the public generally tends to respond to behavioral orientations endorsed by the government, particularly in emerging employment scenarios such as crowdsourced delivery that have not yet been fully institutionalized [10].

IC directly affects the public’s evaluation of their own delivery capabilities, thereby enhancing PBC. In this study, IC encompasses dedicated cycling lanes, parking areas, charging stations, and a safe traffic environment. A well-developed physical environment not only improves delivery efficiency and safety but also alleviates users’ concerns about traffic conflicts and vehicle wear. Existing studies have emphasized that the accessibility and convenience of green transportation tools are core factors influencing their continued use [81]. Especially in high-density urban areas, the level of IC is often positively correlated with the public’s willingness to engage in green travel.

From the model path, situational variables such as PS and IC do not directly influence behavioral intentions; rather, their effects are mediated through the two variables ATU and PBC. This cognitive mediation mechanism supports the TPB framework proposed by Ajzen [48], which posits that external environmental factors must be filtered and evaluated through individual psychological cognition before they can be translated into actual behavioral tendencies.

Notably, Esmaili et al. [78] emphasized that in the practical application of electric cargo bikes, policy clarity and the extent of infrastructure support strongly determine the public’s willingness to participate and the frequency of use. In the absence of clear institutional guidance or adequate physical support, even individuals who endorse environmental protection goals may still be constrained in their adoption behavior.

5.3. Theoretical Implications

First, this study introduces PR as a negative predictor, verifying its role in influencing the public’s adoption of cargo bikes as delivery carriers. Unlike the traditional TPB model, which focuses on positive psychological variables, this research emphasizes how BI is inhibited when the public faces potential uncertainties. This finding is consistent with the results of Savas-Hall et al. [77] and Arciniega-Rocha et al. [79], indicating that PR can significantly weaken the public’s positive judgments regarding utility and controllability.

Second, this study incorporates PS and IC as external variables, which indirectly affect BI through ATU and PBC, respectively. This design not only expands the explanatory scope of the TPB model but also aligns with the theoretical suggestions of Venkatesh et al. [52] regarding the intervention pathways of contextual variables in the UTAUT2 model. The results indicate that PS and IC play a fundamental supporting role in promoting green crowdsourcing behaviors.

Finally, this study focuses on the public becoming crowdshippers. From the perspective of the public as service providers, the research addresses the criticism that existing crowdsourcing studies have paid insufficient attention to individual participation motivation mechanisms. Compared with traditional research, this analysis enriches the behavioral understanding of green crowdsourced delivery adoption through the extended TPB model.

5.4. Practical Implications

The findings offer important implications for both government agencies and platform operators.

First, policymakers should fully recognize the potential of cargo bikes to reduce distribution-related carbon emissions, decrease traffic noise, and improve urban livability. It is recommended to establish shared cargo bike pilot zones with designated distribution lanes in areas of high logistics demand (e.g., Seoul, Busan), allowing cargo bikes to have priority in core urban areas and ensuring preferential allocation of infrastructure.

Second, based on the positive effect of PS on ATU, the government could implement a green logistics crowdsourcing reward mechanism to provide multiple incentives (e.g., point subsidies, tax exemptions) to crowdshippers who use cargo bikes to complete deliveries. Such institutional guidance can effectively lower the participation threshold and enhance individuals’ positive expectations of behavioral rewards [81].

To alleviate the public’s concerns about risks, particularly in high-risk situations such as rainy or snowy weather and nighttime deliveries, the platform’s responsibility and guarantee mechanisms should be strengthened. The platform should offer comprehensive insurance services (e.g., personal accident insurance, cargo loss insurance) for crowdshippers and develop an intelligent scheduling system to avoid inappropriate delivery periods. Additionally, it should enhance the visualized route safety reminder function to reduce users’ subjective perceptions of uncertainty through technological measures. Furthermore, the platform should proactively collaborate with cargo owners to manage operational risks that may arise under adverse conditions, such as severe weather or natural disasters [14,82,83]. This includes establishing pre-arranged agreements regarding potential delivery delays and developing alternative delivery methods and contingency logistics plans.

Based on the analysis presented in this study, it is specifically recommended that policymakers prioritize promoting insurance and liability protection mechanisms. This initiative can significantly alleviate public risk concerns. Secondly, it is recommended that more investment be made in infrastructure. Practical improvements to operational conditions can often increase public adoption. Finally, in the mid- to long-term promotion phase, it is suggested that managers clarify regulatory oversight and regulations. These phased policies should be gradually implemented to provide access guarantees for the sustainable development of green crowdsourced delivery, increase the crowdshippers’ desire to use the green crowdsourcing tools, and ultimately promote green delivery.

Currently, most crowdsourced delivery platforms in South Korea are operated by private companies. Due to the rapid expansion of the last-mile delivery market, competition among these platforms has intensified, giving rise to several critical issues, including the burden of excessive delivery commissions on participants and concerns related to fair trade practices. These challenges have emerged as prominent social issues in recent years. Therefore, it is recommended to adopt a government–private partnership model to establish a shared distribution platform that ensures fairness, efficiency, and sustainability within the crowdsourced delivery ecosystem.

Although this study focuses on the South Korean market, the model exhibits certain cross-regional applicability. For instance, in developing countries such as Vietnam and Indonesia, where motorcycle penetration is high, traffic regulations are complex, and infrastructure development is uneven, the proposed model demonstrates strong adaptability and can serve as a reference for relevant practices.

6. Concluding Remarks

In summary, this study investigates the public’s willingness to use cargo bikes for participation in crowdsourced delivery. To address this research aim, an extended TPB model was constructed and tested through questionnaire surveys.

Regarding RQ1, the results indicate that most of the public are willing to participate in crowdsourced delivery under certain conditions. Both ATU and PBC exhibit significant positive effects on BI. In other words, the more the public perceives crowdsourced distribution positively, the stronger their willingness to participate. Meanwhile, individuals’ judgments regarding their ability to perform delivery tasks also significantly influence their willingness to participate.

Regarding RQ2, the results indicate that PR has a significant negative effect on both ATU and BI. The public is generally concerned about traffic safety, adverse weather conditions, and unclear delivery responsibilities. These concerns reduce acceptance of cargo bikes and directly diminish willingness to participate in crowdsourced delivery. In other words, in promoting green crowdsourced delivery, mitigating both actual and perceived risks is essential to enhancing public willingness to participate.

Regarding RQ3, the analysis indicates that PS enhances ATU, thereby indirectly increasing BI. Similarly, IC indirectly affects willingness to participate by improving PBC. These findings suggest that institutional guarantees and the physical environment do not directly drive public participation; rather, they influence psychological expectations and behavioral judgments by shaping public cognition.

From a theoretical perspective, this study extends the TPB model within the context of crowdsourced delivery. The results further elaborate on the findings of Esmaili et al. [78] and Kapser and Abdelrahman [81] regarding the effects of PR and the impacts of PS and IC. Practically, the research provides actionable management suggestions, such as establishing shared cargo bike pilot areas and enhancing platform responsibility mechanisms. These recommendations offer valuable policy guidance for South Korea and other cities with similar urban logistics contexts.

Although this study provides both theoretical and practical contributions to understanding public participation in green crowdsourced delivery, it has several limitations. First, the model focuses on PR, PS, and IC; future research could introduce additional incentive factors, such as perceived benefits. Second, the SN variable was excluded from this model. However, it may play a significant role in other cultural contexts. Future studies are encouraged to reintroduce SN and examine its potential interactive effects. Finally, the data used in this study are primarily drawn from the South Korean context. Although the model demonstrates some external applicability, future cross-national research should adjust and localize the constructs to account for cultural and contextual differences.