Consumer Sustainability: Is Knowledge Linked to Behavior in Recycling?

Abstract

Sustainable consumer behavior encompasses any action that benefits both the environment and society. Recycling is a prime example of such behavior. However, there is a lack of research examining the connection between knowledge about recycling and recycling behavior. The purpose of this study was to examine factors associated with consumer recycling behavior under the guidance of an extended theory of planned behavior with an emphasis on the role of recycling knowledge. Using data collected from a national online survey in the U.S., results from the structural equation modeling showed that ascription of responsibility to others is negatively associated with recycling behavior, while behavior skill is positively associated with it. Also, both subjective and objective knowledge measures are positively associated with recycling behavior. In addition, subjective recycling knowledge moderates the relationships between attitude and recycling behavior.

1. Introduction

Recycling is a sustainable consumer behavior that benefits both the environment and society. Research on consumer recycling behavior has been conducted for many years, and several review papers have been published on this topic in recent years [1,2,3,4]. However, research studying the role of recycling knowledge in explaining recycling behavior using structural equation modeling (SEM) is limited. This study aimed to fill this gap.

The purpose of this study was to identify factors associated with recycling behavior with an emphasis on the role of recycling knowledge. Using a framework informed by the Theory of Planned Behavior (TPB), we examined factors associated with recycling behavior. Specifically, we examined the direct and indirect effects of recycling knowledge on recycling behavior.

Using data from a national online survey and employing structural equation modeling (SEM), we found that both objective and subjective recycling knowledge measures are positively associated with recycling behavior, while the interaction term of subjective recycling knowledge and attitude toward recycling is negatively associated with recycling behavior. In addition, we found that behavioral skill, a variation of perceived behavior control in the TPB term, is positively with recycling behavior, and ascription to others’ responsibility, a new factor developed on the basis of previous research, is negatively associated with it.

This study contributes to the literature in three ways. First, this study used an extended framework informed by the TPB to directly examine self-reported recycling behavior. In previous research, almost all studies have used behavior intention to measure recycling behavior (see studies reviewed in the hypothesis section). In this study, we used self-reported recycling behavior to measure the behavior directly. Second, we focused on the role of recycling knowledge and explored its direct and indirect effects on recycling behavior, which is lacking in the literature on recycling behavior research using SEM. Third, we used data collected nationwide to study recycling behavior among American consumers. In previous research of this kind, only two studies were found, which collected data from either one state [5] or an unspecified location [6]. To the best of our knowledge, this is the first study using a national dataset in the U.S. on this topic.

2. Literature Review and Hypothesis Development

2.1. Research on Recycling Behavior

Recycling behavior is defined as an individual’s waste collection behavior to allow materials to be reused [2]. Research on recycling behavior has been extensively studied for many years, with a substantial body of literature published on the subject. Several reviews have summarized research on consumer recycling behavior from diverse perspectives [1,2,3,4]. A meta-analysis of studies on individual and household recycling revealed that behavior-specific factors were better predictors of recycling than general factors [2]. A systematic review across multiple disciplines examining factors that influence household recycling behavior among adults in urban areas of high-income OECD countries revealed a comprehensive, multi-level hierarchy of potential determinants [3]. A bibliometric analysis aiming to identify current trends, research networks, and hot topics revealed that 60% of papers on this subject were published between 2015 and 2020, highlighting its global relevance [1]. A systematic review identified seven content clusters, including environmental behaviors and their determinants, household recycling behavior, and behavioral theories, with the TPB being the most commonly employed framework [4]. For the purpose of this study, we reviewed studies on recycling behavior that used the TPB as the theoretical framework and SEM as the analytic approach.

2.2. Conceptual Framework

The conceptual framework used in this study is an extended theory of planned behavior. In the TPB, behavioral intention predicts behavior, while behavioral intention itself is predicted by three factors: attitude toward behavior, subjective norm, and perceived control. In addition, perceived control also directly predicts behavior [7]. The TPB is commonly used in the study of consumer recycling behavior [4]. Based on this theory and the relevant literature, we made two modifications. First, we assumed that the factors identified by the TPB directly predict self-reported behavior. Second, we incorporated three additional predictors: ascription of responsibility to others, personal motivation, and recycling knowledge. The conceptual model and corresponding hypotheses are presented in Figure 1.

2.3. Psychological Factors and Recycling Behavior

We identified ten studies on consumer recycling behavior that used the TPB to guide their hypothesis development and SEM for data analyses.

Table 1. Literature of research on recycling behavior using TPB and SEM.

Study	Intention Predictors	Other Links	Sample
[5]	Behavioral skill (PBC) + Ascription of responsibility + Personal norm + Personal motivation + Social motivation + Information +ns Emotion +ns	Intention, habit, personal norm->behavior Personal motivation, Social motivation, Information->behavioral skill Personal motivation, awareness of consequences->ascription of responsibility Ascription of responsibility, social motivation, awareness of consequences information->personal norm	520 New York state residents
[6]	Attitude + Peer influence (SN) + Social media influence (SN) +ns PBC +	Sense of duty->attitude Convenience, cost->perceived control Intention->direct, indirect behavior	467 Americans
[8]	Attitude − Subjective norm + PBC +	Intention->behavior	400 housewives in Turkey
[9]	Attitude +ns Subjective norm + PBC +	Intention->behavior	427 citizens in Croatia
[10]	Attitude + Subjective norm + PBC +	Intention->behavior PBC->behavior	2004 consumers in South Africa
[11]	Attitude + Subjective norm + PBC +	Intention->behavior	180 consumers in Malaysia
[12]	Attitude + Subjective norm + PBC +	Intention->behavior PBC->behavior	205 respondents in Turkey
[13]	Appreciated guilt + Anticipated pride +ns Attitude −ns Subjective norm + Perceived effort (PBC) + Knowledge (PBC) +	Concern, consequences->guilt, pride Intention->behavior	287 consumers in UAE using a two-wave survey
[14]	Attitude + Subjective norm + PBC + Normative social influence + Informational social influence +	Intention->behavior	353 consumers in Karachi, Pakistan
[15]	Attitude + Subjective norm + PBC + Moral norm + Awareness of consequences + Convenience +	Intention, PBC-> behavior	1303 respondents aged 18–35 in China

Notes. + means the variable is positively associated with the variable in column 3; − means the variable is negatively associated with the variable in column 3; ns means “not significantly”.

presents some details of these studies. Among them, five used factors such as attitude toward recycling, subjective norms, and perceived behavior control to predict behavioral intention and then used both behavioral intention and perceived behavioral control to predict behavior [8,9,10,11,12]. The other five used the three factors specified by the TPB and additional factors developed by themselves, in some of which researchers made the structure more complicated than the original TPB [5,6,13,14,15].

The Theory of Planned Behavior assumes that three factors are associated with behavioral intention, which are attitude, subjective norm, and perceived behavioral control. In addition, two factors predict actual behavior: behavioral intention and perceived behavioral control [7]. In this study, we used self-reported behavior to substitute actual behavior, which could be considered a combination of behavioral intention and actual behavior. Also, we used social motivation to refer to subjective norm and behavioral skill to refer to perceived behavioral control, following previous research [5]. Among the surveyed studies, six out of ten showed that attitude toward recycling was positively associated with recycling intention. Eight out of ten demonstrated that subjective norm was positively associated with behavioral intention. All surveyed studies indicated that perceived behavioral control contributes to behavior intention. Also, all surveyed studies showed that behavioral intention was positively associated with self-reported recycling behavior. Some studies used different terms when presenting their factors. For example, one study used behavioral skill to refer to perceived control and social motivation to refer to subjective norms [5]. In this study, we used some terms employed by previous research [5]. Based on the literature, we propose the following hypotheses:

H1: 

Attitude toward recycling is positively associated with recycling behavior.

H2: 

Social motivation is positively associated with recycling behavior.

H3: 

Behavioral skill is positively associated with recycling behavior.

Based on the literature review, this study incorporates two additional psychological factors into the conceptual model: ascription of responsibility to others and personal motivation. Ascription of responsibility originally referred to consumer responsibility for sustainable behavior [16] and was later applied in predicting consumer recycling behavior [5]. In this study, we hypothesize that if consumers believe recycling is the responsibility of others rather than their own, they may be less likely to engage in recycling. We use the term “others’ responsibility” to represent this new factor. Therefore, we propose the following hypothesis:

H4: 

Ascription of responsibility to others is negatively associated with recycling behavior.

In a previous study [17], consumers perceived external barriers were labeled personal cost, which was assumed to be linked to recycling behavior, but the link was not statistically significant. Later, similar items were used in another study, labeled personal motivation [5], in which personal motivation predicted three factors: behavioral skill, ascription to responsibility, and recycling intention. Thus, we propose the following hypothesis:

H5: 

Personal motivation is positively associated with recycling behavior.

2.4. Recycling Knowledge and Recycling Behavior

In consumer behavior research, knowledge is related to behavior in specific domains. For example, financial knowledge is linked to financial behavior [18]. In recycling behavior research, a review concluded that while recycling knowledge is associated with recycling behavior, the relationship is not particularly strong [2]. In addition, recycling knowledge may play a stronger role in predicting recycling behavior when interacting with other factors [2]. Among ten studies using SEM to study recycling behavior, only one mentioned recycling knowledge [13]. However, upon examining the items used to measure recycling knowledge, we found that they primarily reflected the perceived ability to recycle, which aligns more closely with the concept of perceived behavioral control. To our knowledge, no studies on recycling behavior using SEM have explicitly considered recycling knowledge as a distinct factor. Based on the above discussions, we propose the following hypothesis:

H6: 

Recycling knowledge is positively associated with recycling behavior.

H7: 

Recycling knowledge moderates the relationships between other factors and recycling behavior.

3. Methods

3.1. Data

This study utilized data from a national survey examining recycling behavior, approved by the institutional review board (IRB) of the researchers’ university. The questionnaire is available from the authors upon request. The sample was collected through an online platform, Cloud, between January and March of 2024, and included 1343 participants across the United States. Cloud is a crowdsourcing platform that connects researchers with participants who are willing to take part in online surveys. The platform has a large and diverse pool of participants, which makes it ideal for collecting data nationwide.

Demographically, the sample was predominantly female (57.48%) and identified as White (80.15%). The ages of respondents ranged from 18 to 91 years, with an average age of 52.79 years (SD = 17.9). In terms of recycling behavior, 80% of participants reported engaging in recycling activities. After removing observations with missing data on relevant variables, the final sample size used in the analyses was 1248.

3.2. Variables

Recycling behavior. Recycling behavior was measured by the question “In this study, recycling refers to the action of sorting household wastes into different bins before sending them to the garbage collection place. Do you recycle?” with two options, yes or no, in which yes was coded to 1 and no was coded to 0.

Psychological factors. Items for five psychological factors were adapted from previous research, such as personal motivation, social motivation, behavioral skill, ascription of responsibility, and attitude [5]. The descriptions of these factors and their items are shown in

Table 2. Items of consumer perceptions on recycling.

Personal motivation (α = 0.743)

1.   Finding room to store recyclable materials is a problem

2.   The problem with recycling is finding time to do it

3.   Storing recycling materials at home is unsanitary

Social motivation (α = 0.904)

4.   Most people who are important to me think I should recycle

5.   My household/family members think I should recycle

6.   My friends/colleagues think I should recycle

Behavioral skills (α = 0.913)

7.   I can recycle easily

8.   I have plenty of opportunities to recycle

9.   I have been provided satisfactory resources to recycle properly

10. I know which materials/products are recyclable

11. I know when and where I can recycle materials/products

Ascription of responsibility (α = 0.789)

12. I am responsible for recycling properly (moved to behavioral skill)

13. Government should be responsible for recycling properly

14. Producers should be responsible for recycling properly

Attitude (α = 0.753)

15. Recycling is a desirable behavior (removed)

16. Recycling is not necessary

17. Recycling is benefiting society (removed)

18. Recycling has little benefit for individuals

Note: All variables were measured using a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). The α values were calculated with finalized scales.

. After exploratory factor analyses, items for the five factors were identified (

Table 3. Results of exploratory factor analyses.

Rotated Component Matrix a
Item #	Component
	1	2	3	4	5
8	0.824	0.184	0.135	−0.114	−0.028
7	0.805	0.198	0.090	−0.191	−0.050
11	0.801	0.247	0.104	−0.090	−0.043
9	0.799	0.244	0.054	−0.130	0.091
12	0.732	0.202	0.236	−0.053	−0.082
10	0.634	0.161	0.245	−0.094	−0.129
6	0.318	0.859	0.156	−0.037	−0.020
4	0.290	0.842	0.077	0.024	−0.031
5	0.359	0.824	0.150	−0.045	−0.090
13	0.121	0.036	0.855	0.030	0.077
14	0.176	0.151	0.848	−0.026	0.037
17	0.434	0.166	0.524	0.012	−0.419
15	0.446	0.221	0.513	−0.047	−0.234
1	−0.189	0.014	0.035	0.804	0.003
3	−0.090	−0.071	−0.011	0.785	0.155
2	−0.108	0.012	−0.037	0.773	0.212
16	−0.055	−0.041	−0.061	0.185	0.857
18	−0.027	−0.031	0.066	0.167	0.854

^a. Rotation converged in 6 iterations. Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.

). Most items were loaded to the pre-assigned factors, but two adjustments were made. First, two items, item 15 and item 17, were loaded to two factors and then dropped from the following analyses. Second, item 12 was originally assigned to the factor “Ascription to responsibility” but it was loaded to “Behavioral skill”. Then, the item was used to construct the latent variable for behavioral skill. These measures’ reliability scores—Cronbach’s alpha (α) values—are adequate (Table 2) and their validity is supported by previous research [5].

Knowledge variables. Two recycling knowledge variables were used. One was subjective knowledge with a question asking, “On a scale from 1–7, with 1 meaning very low and 7 meaning very high, how would you assess your recycling knowledge?”, which was adapted from the National Financial Capability Survey [19] by changing the wording from financial knowledge to recycling knowledge. The other was a set of nine true/false questions asking about recycling-related practices. The questions were originally used by a state nonprofit organization for a recycling behavior survey [20]. Each of the questions was first coded as 1 = correct and 0 = incorrect, and then the scores were summed to a total score, ranging from 0 to 9.

3.3. Analyses

3.3.1. Sampling Weight

To adjust for sampling bias and nonresponse, we adopted a weighting approach to make the sample representative of consumer behavior across the USA. A weight variable was calculated using the inverse of selection probabilities combined with a poststratification approach, following established methods [21]. Sampling strata were defined by geographic region (Northeast, Midwest, South, West), sex, and age group (≤35, 35–65, >65), with each stratum containing at least 30 sample units [22]. Selection probabilities were determined by the ratio of respondents within each stratum to the total population in that stratum according to census data. The weight variable was then normalized so that the sum of the weights equaled the total number of respondents.

3.3.2. SEM Analysis

The analyses aimed to identify factors associated with recycling behavior and explore the potential moderation effects of knowledge variables on the relationship between various latent predictors and recycling behavior using structural equation modeling (SEM) [23]. The predictors included attitude, personal motivation, social motivation, others’ responsibility, and behavioral skill as latent variables, with subject and objective knowledge serving as observed variables.

The SEM analysis was conducted using the lavaan package in R, a widely used tool for SEM that provides a robust framework for exploring complex relationships between latent and observed variables [24,25]. The model estimation employed the Diagonally Weighted Least Squares (DWLS) estimator, which is particularly suited for handling ordinal data, such as Likert-scale measures, ensuring accurate parameter estimates [24].

We specified the models in lavaan syntax, defining the latent constructs and their indicators on the basis of theoretical underpinnings and prior empirical findings. Moderation effects were incorporated by creating interaction terms between knowledge variables and latent predictors, allowing us to test how knowledge influences the strength of these relationships. Additionally, to account for population-level representativeness, we applied sample weights provided in the dataset using the sampling.weights argument. This modeling approach allowed for a nuanced exploration of the interactions between knowledge and other factors in predicting recycling behavior, providing valuable insights into the mechanisms underlying sustainable consumer practices.

Model fit was evaluated using established fit indices, including the Comparative Fit Index (CFI), Tucker–Lewis Index (TLI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR). Among all the models considered, the model in

Table 4. SEM results.

Predictor	Estimate	SE	z-Value	p-Value	Std. All
Attitude	0.144	0.083	1.737	0.082	0.115
Personal Motivation	0.106	0.097	1.095	0.273	0.064
Social Motivation	0.115	0.074	1.549	0.121	0.080
Others’ Responsibility	−0.226	0.088	−2.552	0.011	−0.146
Behavioral Skill	0.771	0.078	9.943	<0.001	0.556
Objective Knowledge	0.054	0.027	2.046	0.041	0.099
Subjective Knowledge	0.358	0.033	10.691	<0.001	0.499
Attitude × Knowledge	−0.165	0.048	−3.415	0.001	−0.228
Personal Motivation × Subj Knowledge	−0.097	0.053	−1.817	0.069	−0.118
Others’ Responsibility × Subj Knowledge	0.000	0.043	0.004	0.997	0.000
Social Motivation × Subj Knowledge	−0.031	0.037	−0.832	0.405	−0.040
Behavioral Skill × Subj Knowledge	−0.002	0.051	−0.044	0.965	−0.003

with only the subjective variable’s moderation effect had the best overall fit.

4. Results

4.1. Measurement Model

The measurement model was assessed using Confirmatory Factor Analysis (CFA) with five latent constructs: attitude, personal motivation, social motivation, others’ responsibility, and behavioral skill. Each construct was represented by multiple observed indicators. The model was fit to the data using the WLSMV estimator, which is appropriate for ordinal variables.

The fit indices indicated an acceptable model fit, with a Comparative Fit Index (CFI) of 0.918 and a Tucker–Lewis Index (TLI) of 0.914. The Root Mean Square Error of Approximation (RMSEA) was 0.094, with a 90% confidence interval of [0.092, 0.96]. Despite the RMSEA being slightly above the ideal value, the model was deemed acceptable due to the relatively strong CFI and TLI values. Additionally, the Standardized Root Mean Square Residual (SRMR) was 0.046, well within the recommended cutoff of ≤0.08, further supporting the model’s adequacy.

All factor loadings were significant (p < 0.001), ranging from 0.609 to 0.974, demonstrating strong relationships between the latent constructs and their respective indicators. Covariances between latent factors were also significant, with notable correlations such as the positive relationship between social motivation and behavioral skill (r = 0.624) and the negative correlation between attitude and social motivation (r = −0.268). These results support the construct validity of the measurement model.

4.2. Analysis Model

The SEM analysis provided insightful results regarding the predictors of recycling behavior and the moderating roles of knowledge variables and only subjective knowledge showed moderation effects. The findings (Table 4 and Figure 2) indicated that behavioral skill was the strongest predictor of recycling behavior, with a significant positive effect (β = 0.556, p < 0.001). This underscores the importance of practical skills in encouraging consistent recycling activities among individuals.

For interpretation and visualization, we used standardized regression coefficients (β), represented in the Std. All column. Standardized coefficients were chosen over unstandardized ones to facilitate direct comparison of effect sizes across predictors, as they are scale-independent and emphasize the relative importance of variables.

Subjective knowledge, which reflects individuals’ perceived understanding of recycling, also had a strong and significant positive impact on recycling behavior (β = 0.499, p < 0.001). This suggests that individuals who feel more knowledgeable about recycling are more likely to engage in recycling practices. Additionally, objective knowledge, which measures factual understanding of recycling processes, also showed a significant positive but weaker association with recycling behavior (β = 0.099, p = 0.041). This indicates that while accurate knowledge about recycling contributes to recycling practices, the perceived confidence in one’s knowledge plays a more dominant role in driving recycling behavior.

In contrast, the belief that recycling responsibility lies with others, represented by others’ responsibility, was negatively associated with recycling behavior (β = −0.146, p = 0.011). This suggests that individuals who believe that others should handle recycling are less likely to engage in recycling themselves.

Interestingly, attitude (β = 0.115, p = 0.082), personal motivation (β = 0.064, p = 0.273), and social motivation (β = 0.080, p = 0.121) were not significant predictors of recycling behavior on their own. However, this study found one significant interaction effect for attitude when moderated by subjective knowledge.

The interaction between attitude and subjective knowledge (β = −0.228, p = 0.001) revealed that higher levels of perceived knowledge can reduce the impact of attitude on recycling behavior.

Conversely, the interactions between subjective knowledge and social motivation (β = −0.040, p = 0.405), subjective knowledge and personal motivation (β = −0.118, p = 0.069), and subjective knowledge and behavioral skill (β = −0.003, p = 0.965) were not significant, suggesting that the moderating effect of knowledge does not extend to these factors.

Objective knowledge was not included as a moderator in the analysis due to practical and methodological considerations. Given the sample size (n = 1248), introducing interactions between objective knowledge and other predictors would significantly increase model complexity, potentially leading to estimation challenges. Structural equation modeling (SEM) requires sufficient sample size to estimate complex models reliably, and the addition of multiple interaction terms could compromise the stability and interpretability of the results. Furthermore, the main effects analysis indicated that subjective knowledge (β = 0.499, p < 0.001) has a substantially stronger association with recycling behavior compared with objective knowledge (β = 0.099, p = 0.041). This suggests that subjective knowledge plays a more pivotal role in influencing recycling behavior. By focusing on interactions involving subjective knowledge, the study provides more meaningful insights while maintaining model parsimony and estimation reliability.

Overall, the results underscore the critical importance of both practical skills and subjective knowledge in promoting recycling behavior. The significant moderation effect suggests that enhancing individuals’ perceived knowledge can mitigate the effects of attitudes, implying the limitations of recycling knowledge in promoting recycling behavior.

5. Discussion, Limitations, and Implications

5.1. Discussion

With a national sample and the structural equation modeling approach, we examined factors associated with consumer recycling behavior. We found that four factors are associated with consumer recycling behavior, which are others’ responsibility, behavioral skill, subjective recycling knowledge, and objective recycling knowledge. In other words, if consumers believe recycling is only the government’s or businesses’ responsibility, they are less likely to recycle, while if they believe they are able to recycle (higher level of self-efficacy), they are more likely to recycle. Also, both subjective and objective recycling knowledge measures show significant positive associations with recycling behavior. In addition, subjective knowledge shows a negative moderating effect through attitude. These findings imply that subjective knowledge has the potential to reduce the effect of attitude on recycling behavior. These findings are partially consistent with previous research.

The positive association between behavioral skill and recycling behavior found in this study is consistent with many studies reviewed. In the published studies, many scholars used perceived behavior control to refer to behavioral skill. Based on the Theory of Planned Behavior, perceived behavior control measures consumer willpower to take action and believe that they can do it [7]. In previous research, perceived behavioral control is also referred to as behavior skill, which shows links to consumer recycling behavior [5]. Green efficacy, another term similar to behavioral skill, is also linked to recycling behavior in previous research [26]. The findings of this study support previous research and show that enhancing consumer confidence in recycling behavior is important for encouraging recycling behavior.

The positive association between knowledge and recycling behavior has also been shown in previous research. Interestingly, when comparing both subjective and objective knowledge variables, we found that the coefficient for subjective knowledge was much higher than that for objective knowledge, implying improving consumer confidence in knowledge may be more important than raising real knowledge level in encouraging recycling behavior, echoing similar findings in consumer financial behavior research [27]. Among studies that have applied the Theory of Planned Behavior and SEM to study consumer recycling behavior, no studies have focused on the potential effects of recycling knowledge on recycling behavior. The findings of this study contribute to the literature by showing the potential impacts of recycling knowledge on recycling behavior in the context of the Theory of Planned Behavior with SEM. The findings also show differential effects of two types of recycling knowledge measures, echoing previous research. Previous research indicates that different types of information may affect recycling behavior differently [28]. Another interesting finding is that subjective knowledge shows a negative moderation effect on the association between attitude and recycling behavior. This implies that raising consumer knowledge levels may reduce the impact of attitudes on recycling behavior, which suggests limitations of recycling knowledge.

The negative association between others’ responsibilities and recycling behavior is novel according to our knowledge. In previous research, ascription to responsibility refers to consumer responsibility for recycling. In this study, this factor was changed to others’ responsibility. If consumers believe recycling is mainly the government’s or businesses’ responsibility, they are less likely to recycle. The findings are consistent with common sense, but this is the first time a study has measured this factor. The findings suggest that any government interventions to help improve the sense of consumer responsibility for recycling may help increase recycling behavior among consumers.

Three factors (attitude, social motivation, and personal motivation) show links to recycling behavior, but the links are not statistically significant, which is inconsistent with previous research. These factors are associated with recycling behavior in previous research. Why are their links to recycling behavior not statistically significant? Possible reasons may be that we did not use behavioral intention in this study but used the self-reported behavior directly. When all factors were included, some factors lost explaining power. Even though these factors are not statistically significant, the coefficient estimates’ signs are consistent with the hypotheses.

5.2. Limitations

The limitations of this study need to be acknowledged. First, the cross-sectional nature of the data limits causal inference. Our findings only illustrate associations between psychological/cognitive factors and recycling behavior, as well as the potential moderation effects of subjective recycling knowledge. To test the causal relationship, experimental or longitudinal data are needed. Second, the reliance on self-reported measures may introduce measurement errors. More valid measures for recycling behavior should be observed behavior or administrated records. Third, several factors showing links in other studies did not show significant associations with recycling behavior in this study. Specific reasons need to be explored with other survey data on similar topics. These limitations could be addressed in future research. In future research, to address the causality of influencing factors on consumer recycling behavior, experimental research or longitudinal research should be utilized in data collection. Also, administrative data sources could be explored to link actual consumer behavior with influencing factors, such as the psychological and knowledge factors identified in this study. Researchers using the Theory of Planned Behavior as a theoretical framework may also collect data from other countries to confirm or disconfirm the findings presented in this study.

5.3. Implications

The SEM analysis highlights the importance of both behavioral skills and recycling knowledge in promoting recycling behavior. These findings suggest that interventions to enhance recycling knowledge may be effective in improving recycling behavior. By understanding the roles of recycling knowledge, behavioral skill, and ascription to others’ responsibility, policymakers and educators can design targeted strategies to enhance recycling practices.

Enhancing consumer behavioral skill. Government programs should help consumers enhance their ability to recycle and recycle correctly. Government agencies should provide information to help consumers recycle conveniently and appropriately. Educators should encourage consumer confidence in recycling and convey recycling knowledge to them.

Increasing consumer recycling knowledge. The findings suggest that both objective and subjective recycling knowledge measures may contribute to recycling behavior. Also, comparing the two types of knowledge, subjective knowledge seems more important to help consumers take action. Relevant government agencies should provide sufficient information for consumers who want to learn how to recycle and use resources to train consumers. It is important to train consumers with more recycling knowledge. Equally important, the training may also help increase consumer confidence to recycle, which helps enhance recycling behavior.

Let consumers take responsibility. The findings show that if consumers believe recycling is only others’ responsibility, they are less likely to recycle. Consumers should be encouraged to view recycling as a personal duty for a better future, supported by government policies setting minimum recycling standards. This message should be conveyed through various communication channels. Consumer education should emphasize that recycling is a personal responsibility, not someone else’s. They should also develop hands-on course projects to help consumers understand the concept through practical experience.