Factors That Influence Clothing Upcycling Behavior in Japan: Extending the Theory of Planned Behavior

: In recent times, clothing upcycling has emerged as a sustainable solution to tackling textile waste. Despite its popularity, there remains a limited understanding of the factors driving individuals’ upcycling behavior. This study delves into the determinants of clothing upcycling behavior within the Japanese population. Administered via an online survey involving 433 participants in Japan, this research utilized structural equation modeling to assess a theoretical model that integrates personal norms, attitudes, social norms, perceived behavioral control, and intention within the theory of planned behavior. The outcomes reveal that both personal norms and perceived behavioral control exert a substantial influence on individuals’ intentions to participate in upcycling activities. Interestingly, the impact of attitude on upcycling intentions is moderate, while subjective norms surprisingly showed no significant influence. These findings contribute to the pro-environmental behavior literature and can guide the development of focused interventions that promote sustainable fashion consumption through upcycling.


Introduction
The world is currently grappling with the severe issue of climate change and pollution caused by excessive CO 2 emissions and the substantial volume of waste generated annually.Within the realm of polluting industries, the fashion sector stands out as a significant contributor, responsible for an estimated 8% of the world's annual carbon emissions [1,2].Additionally, wastewater resulting from textile production contains various chemicals and dyes, posing a threat to aquatic life and the environment.The environmental consequences of this pollution, coupled with the depletion of natural resources and greenhouse gas emissions, have sparked concerns about the fashion industry's sustainability [1,3,4].
Even with all these problems, the growth of fast fashion shows no sign of slowing down, with global textile production nearly doubling from 2000 to 2015 [2,5], and the demand for clothing and footwear is projected to rise by 63% by 2030, increasing from the current 62 million tons to 102 million tons [6].Yet, the average frequency of wearing a garment before its disposal has decreased [5].Fast fashion reflects our fast-paced lives, offering quick, trendy, and affordable clothing.While it lets us stay stylish, it comes at a cost.It encourages impulsive buying, adding to the problem of overconsumption that leads to tons of waste, polluting our environment, and exploiting labor in the pursuit of cheap production [7].
Niinimäki et al. highlight that fashion consumption has outpaced population growth since 2010, leading to the overconsumption of textile products [8].Additionally, the proportion of monthly income spent on clothing and fast fashion is influenced by family income, with higher incomes correlating with greater expenditure on clothing [9].In contrast, many low-income countries maintain relatively high clothing utilization rates.
However, in the United States, clothes are worn for only about a quarter of the global average duration, and a similar trend is seen in China, where clothing utilization has dropped by 70% over the past 15 years [5].Globally, consumers miss out on approximately USD 460 billion annually by discarding clothes that could still be worn [5].
In Japan, a survey conducted by the Ministry of the Environment revealed that about 450,000 tons of textile waste were disposed of in landfills or incinerators in 2022 [10].This figure represents 66% of the total textile disposal.Additionally, the survey indicated that, on average, people own 35 pieces of clothing that they have not worn even once in a year, highlighting issues related to overconsumption.
Overconsumption challenges us to rethink our relationship with material things and adopt sustainable behaviors.The emerging discussion on sufficiency highlights the importance of consumers decreasing their present consumption levels [11,12].Clothing upcycling has emerged as a promising approach to promoting a more sustainable and circular fashion system [13][14][15] as it emphasizes resource efficiency, minimizing environmental impact [16][17][18][19], and the reduced consumption of new textiles [20].The term "upcycling" is defined as "the process of creating new furniture, objects, etc., from old or used items or waste material", according to the Cambridge Dictionary [21].Moreover, the literature provides various explanations of upcycling, often describing it as the transformation of discarded materials or used items into valuable or high-quality outputs, either in the form of products or materials [22][23][24].However, within this study, consumer clothing upcycling pertains to the activities of repairing, altering, and repurposing old or discarded garments to create new and appealing items, all performed by the consumers themselves.This activity will lead to the extension of products' lifespan [10,25], a reduction in waste generation [22], enhanced creativity [26], the personalization of clothing [27,28], and product attachment [27,29].
Manufacturers' regulations and local laws play a critical role in driving ecological behavior in the market.However, our study specifically focuses on consumer behavior.Studying consumer pro-environmental behaviors is vital as it can significantly enhance the effectiveness of policies or technological developments [30] and aid interventions aimed at promoting eco-friendly practices [31].This understanding can help tailor more effective strategies to encourage sustainable consumption.Several studies suggest approaches to extending the lifespan of clothing [32][33][34], including practices like recycling, reusing (such as repairing and repurposing), and upcycling [24,[35][36][37][38].However, based on our literature review, there is still a lack of studies that explore the factors that influence consumer engagement in clothing upcycling behavior in Japan.The present paper fills this gap.This study seeks to identify the key psychological and social factors that influence individuals' intentions to engage in upcycling activities.We carried out a survey in Japan, with a final sample comprising 433 participants.Subsequently, we conducted an analysis using structural equation modeling to assess the integrated framework incorporating personal norms and the theory of planned behavior (TPB).
The remaining sections of this paper follow this structure: we commence with an introduction to the theoretical framework.We then detail the methodology applied in this study.Subsequent to this, we present the results of the analysis and delve into a discussion of our findings.Lastly, the paper is concluded.

Theoretical Framework
Consumer clothing upcycling behavior can be counted as a pro-environmental behaviors, because this behavior can minimize negative effects on the environment [39,40] by extending the material lifespan and reducing the waste being thrown to landfill.Consequently, the most popular psychological theories that are applied to the study of pro-environmental behavior are the theory of planned behavior and the norm activation model [41].

The Theory of Planned Behavior
The theory of planned behavior [42] serves as a vital theoretical framework for understanding environmentally conscious behavior.The TPB has been extensively utilized to examine environmentally friendly behaviors, including the recycling of electronic waste [43], energy conservation [44,45], saving water [46], sustainable consumption [47][48][49], transportation [50], and others.A comprehensive review of the TPB model used for studying environmental behavior is covered by Yuriev et al. [51].
The theory of planned behavior is recognized as one of the most potent models for designing interventions aimed at influencing behavior [52].Moreover, the flexibility of the framework is also recognized in the TPB.While the initial model revolves around three main predictors, scholars often expand the theory by incorporating additional variables like moral norms [53] and various new factors.Given the demonstrated applicability and versatility of the TPB in behaviors related to the environment, along with its adaptability, we contend that the theory of planned behavior aligns well with our study's objectives.
The TPB posits that behavioral intention and perceived behavioral control are the key determinants of behavior [42].In a sequential manner, behavioral intentions are believed to be shaped by attitudes, subjective norms, and perceived behavioral control (PBC).An individuals' attitude toward a behavior denotes their favorable or unfavorable assessment of a particular behavior, influenced by their beliefs about that behavior.Subjective norm represents an individual's perception of social pressure related to engaging in a particular behavior.Perceived behavioral control involves an individual's perception of the ease or difficulty associated with performing a behavior [54].TPB is designed with singledirectional relationships to reflect the causal pathways through which attitudes, subjective norms, and perceived behavioral control influence behavioral intentions and subsequent behavior [42].

The Integrated Model
While the TPB often garners robust empirical support, critiques have been raised regarding its perceived inadequacy in accounting for the influence of morality on environmental behavior [55,56].Thøgersen contended that environmental behavior is situated within the moral realm, suggesting that it is not solely determined by the cost-benefit calculations outlined in the TPB but also by moral beliefs concerning right and wrong [56].Meta-analyses conducted by Klöckner [41] and Morren and Grinstein [57] revealed that personal norms can introduce a fourth dimension to intentions, filling a gap in the TPB, aligning with findings from Bamberg and Möser [58] and Harland et al. [59].These studies have demonstrated that personal norms significantly enhance the predictive power of the TPB concerning pro-environmental intentions and behaviors.
Researchers have integrated personal norms into the TPB to investigate diverse proenvironmental behaviors.For instance, Hu et al. [60] and Wu et al. [61] found that integrating personal norms into the TPB model significantly improved the prediction of recycling behaviors, underscoring the role of moral obligations in driving sustainable practices.Similarly, studies by Ali et al. [62] and Bamberg et al. [50] on public transportation utilization, and by Abrahamse and Steg [63] and Ru et al. [64] on energy conservation, further validated the importance of personal norms in enhancing the TPB's explanatory power.
Based on this theoretical framework, we propose six hypotheses for our study, as outlined below.The structural model is shown in Figure 1.

H1. Favorable attitudes among respondents amplify their intention to upcycle clothing.
H2. Elevated subjective norms in respondents correlate with increased intentions to engage in clothing upcycling.

H3.
Strengthened perceived behavioral control in respondents increases their intentions to participate in clothing upcycling.
H4. Enhanced personal norms in respondents correspond to increased intentions to participate in clothing upcycling.
H5.The perceived behavioral control in respondents has a direct positive impact on clothing upcycling behavior.
H6. Respondents' intentions to partake in clothing upcycling positively impacts the frequency of their actual behavior.H3.Strengthened perceived behavioral control in respondents increases their intentions to participate in clothing upcycling.
H4. Enhanced personal norms in respondents correspond to increased intentions to participate in clothing upcycling.
H5.The perceived behavioral control in respondents has a direct positive impact on clothing upcycling behavior.
H6. Respondents' intentions to partake in clothing upcycling positively impacts the frequency of their actual behavior.

Measurement
The dependent variables in this study were clothing upcycling intentions and behavior frequency.Clothing upcycling intentions, which aimed to measure participants' willingness or readiness to perform clothing upcycling behavior, were assessed using a 7point Likert scale (1-strongly disagree, 7-strongly agree).Additionally, behavior frequency aimed to measure the extent of individuals' engagement in clothing upcycling activities.It was evaluated utilizing a 9-point frequency scale, ranging from "never" to "about once every six months" to "at least once a week."Information regarding these dependent variables, namely clothing upcycling intentions and behavior frequency, was gathered using distinct measures.In the survey, intention was assessed through three indicators, while behavior frequency was measured according to the highest frequency of at least one of 13 activities (including upcycling, reuse, and repair).
The independent variables consisted of four constructs derived from the extended TPB model; these were a itude toward the behavior, the subjective norm, PBC, and personal norm.Each of these four constructs was measured with three indicators using the same scale as intention.The questionnaire used for this research was designed according to previous studies that have tested and verified its validity and reliability [65][66][67][68][69].The questionnaire is provided in the Supplementary Materials.

Measurement
The dependent variables in this study were clothing upcycling intentions and behavior frequency.Clothing upcycling intentions, which aimed to measure participants' willingness or readiness to perform clothing upcycling behavior, were assessed using a 7-point Likert scale (1-strongly disagree, 7-strongly agree).Additionally, behavior frequency aimed to measure the extent of individuals' engagement in clothing upcycling activities.It was evaluated utilizing a 9-point frequency scale, ranging from "never" to "about once every six months" to "at least once a week."Information regarding these dependent variables, namely clothing upcycling intentions and behavior frequency, was gathered using distinct measures.In the survey, intention was assessed through three indicators, while behavior frequency was measured according to the highest frequency of at least one of 13 activities (including upcycling, reuse, and repair).
The independent variables consisted of four constructs derived from the extended TPB model; these were attitude toward the behavior, the subjective norm, PBC, and personal norm.Each of these four constructs was measured with three indicators using the same scale as intention.The questionnaire used for this research was designed according to previous studies that have tested and verified its validity and reliability [65][66][67][68][69].The questionnaire is provided in the Supplementary Materials.

Translation
The second author, a native Japanese speaker, performed the translation, aligning with the questionnaire's target language [66].Moreover, he is also experienced in questionnaire design in the Japanese language.Additionally, the translated questionnaire was pre-tested with nine laboratory members who are native speakers of the Japanese language, and all necessary modifications were made before the questionnaire was sent to the survey company.

Data Collection
For this research, an online survey was employed, chosen for its convenience, efficiency, and ability to reach a broad and diverse participant pool in a relatively short period and at a reduced cost [70].The primary data were collected through an online survey conducted in Japan by a private Japanese company in April 2023.To ensure representative sampling, we employed a random stratified cluster design, selecting observations according to interlocking quotas of gender and age groups.The target respondents were individuals aged 20 to 69 years old, with no other restrictions.Participating respondents were given modest compensation as a gesture of gratitude upon finishing the questionnaire, which could be utilized during shopping.Before participating, all respondents consented to the terms and conditions outlined by the survey company.Ethical considerations and privacy matters were diligently overseen by the company.
At the beginning of the questionnaire, we included a statement emphasizing that there are no correct or incorrect responses, and we asked respondents to give honest responses.Moreover, the definition of clothing upcycling was also provided at the top and middle of the questionnaire.
The sample size used in this study was calculated by using the following Equation ( 1): where n denotes the required sample size, Z represents the value from the standard normal distribution corresponding to the confidence level, p represents the standard deviation, and e represents the desired margin of error.
For the determination of the sample size, we opted for a 95% confidence level (Z = 1.96), a standard deviation of 0.5, and a margin of error of ±5%.However, to mitigate the potential of unusable data, the sample size was increased to approximately 500 samples.The data received from the survey company consisted of 532 samples, which is more than our requirement.Nevertheless, during the data screening process, 99 samples were removed as they lacked engagement, primarily providing the same answer to most of the questions.Fortunately, there were no missing data in the survey results, which led to a final dataset with 433 samples.

Data Analysis
In this paper, structural equation modeling (SEM) with bootstrapping was utilized to examine the interaction among the constructs and test the proposed model by using AMOS 23.0.Alternatively, other analyses such as descriptive analysis and non-parametric tests were performed using IBM SPSS Statistic 23.Moreover, information regarding the activities used for observing the frequency of clothing upcycling behavior, the skewness and kurtosis statistic, and the exploratory factor analysis (EFA) are provided in the Supplementary Materials.

Descriptive Statistics
We obtained 433 valid responses from the survey, with 52.4% of participants identifying as female.The age distribution of the respondents ranged from 20 to 69 years old, and 51.3% of them reported being married.It is noteworthy that a significant portion of the respondents (69.3%) indicated that they did not have any children.This demographic characteristic could have implications for our study, particularly in the context of clothing upcycling behaviors.It is well-documented that children's clothes are often reused and shared among families rather than being discarded or recycled, which could influence the overall clothing upcycling rates.The absence of children in the majority of our respondents may suggest a different pattern of upcycling behavior focused more on adult clothing.Future research could benefit from a deeper exploration of the reasons behind this demographic trend.Additionally, performing an automated analysis of social network groups related to reproductive behavior could provide further insights into how family structures and child-rearing practices impact clothing upcycling activities.In terms of occupation, the majority of the respondents were company employees (38.1%), followed by those engaged in part-time jobs (16.6%).For a detailed overview of the respondents' demographic information, please refer to Table 1. Figure 2 presents statistical data on the frequency with which the participants reported being involved in clothes upcycling.Notably, a significant proportion (26.1%) reported never participating in clothes upcycling.However, a positive trend is observed among those who do engage in clothes upcycling, with varying frequencies.Among the participating individuals, the most common frequency of clothes upcycling is "about once a year" (18.5%).This suggests that a considerable number of respondents incorporate clothes upcycling into their fashion practices annually, possibly aligning with seasonal wardrobe updates or specific occasions.Furthermore, approximately 16.2% of respondents engage in clothes upcycling approximately once every six months, indicating a moderate level of involvement.The data also reveal that 13.4% of respondents engage in clothes upcycling approximately once every 2-3 months.However, the percentage of respondents engaging in clothes upcycling on a monthly or weekly basis is lower; approximately 6.7% reported engaging in clothes upcycling about once a month, while only 3.7% reported doing so about once a week.
wardrobe updates or specific occasions.Furthermore, approximately 16.2% of respondents engage in clothes upcycling approximately once every six months, indicating a moderate level of involvement.The data also reveal that 13.4% of respondents engage in clothes upcycling approximately once every 2-3 months.However, the percentage of respondents engaging in clothes upcycling on a monthly or weekly basis is lower; approximately 6.7% reported engaging in clothes upcycling about once a month, while only 3.7% reported doing so about once a week.The data on behavior frequency, categorized by gender, are displayed in Figure 3.The findings indicate that 36.9% of male respondents have not engaged in clothing upcycling behavior in the last 5 years, while this number is significantly lower, at 16.3%, for female respondents.Additionally, across the frequency scale from "about once a year" to "about once a week", a higher percentage of female participants is observed in all categories compared to male participants.The data on behavior frequency, categorized by gender, are displayed in Figure 3.The findings indicate that 36.9% of male respondents have not engaged in clothing upcycling behavior in the last 5 years, while this number is significantly lower, at 16.3%, for female respondents.Additionally, across the frequency scale from "about once a year" to "about once a week", a higher percentage of female participants is observed in all categories compared to male participants.After observing the data above, we performed a Mann-Whitney test and median test (non-parametric test) to determine if there is a significant difference in clothing upcycling behavior between male and female groups.The results indicate a statistically significant distinction in behavior frequency between the two groups, with p < 0.001 and an effect size of 0.250.Furthermore, the data reveal that the median for females (5.0) is higher than that of males (3.5), suggesting that females are more engaged in clothing upcycling behavior than males.This outcome could be a ributed to traditional gender roles, where historically, women were primarily responsible for sewing tasks within households, while men engaged in community or labor-intensive work that demanded physical strength, such as tree cu ing and construction.However, we did not consider the role of family dynamics in decision-making processes related to clothing upcycling behavior.Family members of- After observing the data above, we performed a Mann-Whitney test and median test (non-parametric test) to determine if there is a significant difference in clothing upcycling behavior between male and female groups.The results indicate a statistically significant distinction in behavior frequency between the two groups, with p < 0.001 and an effect size of 0.250.Furthermore, the data reveal that the median for females (5.0) is higher than that of males (3.5), suggesting that females are more engaged in clothing upcycling behavior than males.This outcome could be attributed to traditional gender roles, where historically, women were primarily responsible for sewing tasks within households, while men engaged in community or labor-intensive work that demanded physical strength, such as tree cutting and construction.However, we did not consider the role of family dynamics in decisionmaking processes related to clothing upcycling behavior.Family members often make collective decisions about clothing disposal and upcycling practices, which can significantly influence individual behaviors.Future research should incorporate a more comprehensive analysis of family influences.This could involve examining how shared responsibilities, familial roles, and collective values shape upcycling behaviors, thereby offering a more holistic understanding of sustainable consumption practices within households.

Structural Equation Modelling (SEM) 4.2.1. Skewness and Kurtosis
We observed fairly normal distributions for our indicators of latent factors and independent variables, with both skewness and kurtosis values within the (−1, +1) range.

Reliability and Validity of the Constructs
The reliability and validity of the constructs were assessed before performing the structural equation modeling.Reliability testing is a crucial step in assessing the consistency and stability of measurements or scales used in research.It helps determine the extent to which a measure produces consistent results over repeated administrations or across different items that are intended to measure the same construct.Similarly, construct validity assesses how effectively a group of indicators represents a concept that cannot be directly measured.Unfortunately, one of the PBC's indicators, "It is mostly up to me whether or not to upcycle clothes", had a very low loading (lower than 0.5) and did not satisfy the discriminant validity test; hence, it was removed from the PBC construct and the rest of the analyses.The cause of this issue might stem from respondents misunderstanding the question.They might have interpreted it as relating to their decision-making rather than their ability to control the behavior.Thus, only two indicators were used for PBC.However, all constructs in the proposed model had a Cronbach's Alpha ranging from 0.854 to 0.944, which is within the acceptable values of 0.70 to 0.95.The outcomes of the reliability assessment for the model constructs are presented in Table 2. Table 3 presents information about the convergent and discriminant validity.The heterotrait-monotrait ratio, or HTMT ratio [71], of correlations was used to verify the discriminant validity of the constructs in this study.Both validities were established based on the recommended value; the average variance extracted (AVE) [72] by each variable must be more than 0.5 [73] and the HTMT ratio must be less than 0.9 [71].

Structural Model Assessment
The comprehensive evaluation of the overall structural model encompasses several fit indices to determine its appropriateness.A well-fitting model is considered acceptable if the χ 2 /df value is below 5 and if indices like the goodness-of-fit (GFI), the Tucker and Lewis index (TLI), and the Confirmatory Fit Index (CFI) exceed 0.90 [73].Additionally, a fitting model is deemed adequate if the calculated value of the standardized root mean square residual (SRMR) and the root mean square error approximation (RMSEA) is less than 0.08 [73,74].The fit indices of the model consistently fall within the acceptable range: χ 2 /df = 3.430, GFI = 0.919, TLI = 0.948, CFI = 0.963, IFI = 0.964, SRMR = 0.049, and RMSEA = 0.075.
Moving on to the structural equation modeling result, it presents the relationships between individuals' attitude towards the behavior, the subjective norm, personal norm, PBC, intention, and behavior frequency.The age variable was used as a control variable in our structural equation modeling analysis.The hypothesis outcomes are presented in Table 4. Bootstrap technique with a sample size of 5000 was used to indicate the 95% confidence interval of the results [75].The result showed a positive and statistically significant impact of attitude, personal norm, and PBC on behavioral intention, namely (β = 0.072, p = 0.047), (β = 0.374, p = 0.010), and (β = 0.594, p = 0.001), respectively, thereby supporting hypotheses H1, H3, and H4.However, the findings indicated that the impact of subjective norm on intention was not statistically significant (β = 0.016, p = 0.891), hence hypothesis H2 was not supported.Furthermore, the direct effect of perceived behavior control on behavior frequency was found to be statistically insignificant (β = −0.065,p = 0.627), failing to support hypothesis H5.Notably, behavioral intention demonstrated a positive influence on the actual behavior frequency (β = 0.382, p = 0.007), aligning with hypothesis H6.Furthermore, the analysis indicates that the control variable, age, does not significantly affect behavioral intention.This implies that respondents expressed fairly consistent intentions to engage in clothing upcycling, regardless of their age.However, the findings indicate a significant and positive correlation between age and behavior frequency (β = 0.191, p < 0.001).This suggests that older individuals tend to upcycle clothing more frequently than younger individuals.
In relation to the variance explanation, the squared multiple correlation (R 2 ) for intention was determined to be 0.825, signifying that attitude, subjective norm, personal norm, and PBC collectively account for 82.5% of the variance in intention.However, the R 2 for behavior frequency was relatively low, at only 0.133, indicating that the factors of behavioral intention, perceived behavior control, and age account for just 13.3% of the variance in behavior frequency.This suggests a significant gap between intention and actual behavior, highlighting the need for further exploration in this area.The results of the analysis are also displayed in structural form with significant and insignificant paths, as shown in Figure 4.
Additionally, we conducted a multigroup analysis categorized by gender (male and female) to assess the disparities between the two groups in our structural model.According to the chi-square difference analysis, the obtained p-value of 0.063 (df = 15) surpasses the threshold of 0.05.Hence, it can be inferred that there is no noteworthy distinction in our structural model between the male and female groups.R 2 for behavior frequency was relatively low, at only 0.133, indicating that the factors of behavioral intention, perceived behavior control, and age account for just 13.3% of the variance in behavior frequency.This suggests a significant gap between intention and actual behavior, highlighting the need for further exploration in this area.The results of the analysis are also displayed in structural form with significant and insignificant paths, as shown in Figure 4. Additionally, we conducted a multigroup analysis categorized by gender (male and female) to assess the disparities between the two groups in our structural model.According to the chi-square difference analysis, the obtained p-value of 0.063 ( = 15) surpasses the threshold of 0.05.Hence, it can be inferred that there is no noteworthy distinction in our structural model between the male and female groups.

Discussion
The results from the structural equation modeling indicate that individuals' a itudes toward clothing upcycling have a positive and statistically significant impact on their behavioral intention.Interestingly, despite the favorable opinions expressed by respondents about clothing upcycling, as indicated by mean values exceeding the median of 4.0 for all three a itude indicators, the loading factor on behavioral intention remains relatively low.This suggests that even individuals with positive a itudes might not always intend to

Discussion
The results from the structural equation modeling indicate that individuals' attitudes toward clothing upcycling have a positive and statistically significant impact on their behavioral intention.Interestingly, despite the favorable opinions expressed by respondents about clothing upcycling, as indicated by mean values exceeding the median of 4.0 for all three attitude indicators, the loading factor on behavioral intention remains relatively low.This suggests that even individuals with positive attitudes might not always intend to engage in upcycling, possibly due to perceived behavioral costs, such as monetary, time, or effort-related factors associated with the process [40].A survey conducted by the Ministry of the Environment in 2022 revealed that 69% of respondents stated that the primary reason for discarding clothes as garbage was convenience, requiring little time or effort.
Contrary to our expectations, social norms, or the perceived social pressure to engage in clothes upcycling behavior do not play a significant role in influencing individuals' intentions to participate in such practices.Similar findings were noted by Davies et al. [76] in their consumer study, specifically within the context of recycling; these authors found that subjective norms are insufficient for predicting behavioral intentions.Moreover, research conducted by Thøgersen contended that the correlations between norm constructs and behavior vary across different behaviors [77].The cause of this result could be due to clothing upcycling typically occurring within the confines of one's home and not usually involving interaction with others, hence remaining unaffected by external influences.Additionally, cultural factors and individual differences within the Japanese population could also be contributing factors, suggesting a need for further investigation in this area.
On the other hand, the study results revealed significant and positive relationships between personal norms and PBC with the intention to engage in clothes upcycling behavior.These results align with findings from prior research studies that have highlighted the importance of personal beliefs and perceptions in shaping the behavioral intentions related to pro-environmental actions [78].For this reason, promoting personal norms related to resource efficiency and environmental protection could be a powerful strategy to encourage clothes upcycling behavior among consumers in Japan, given the moderate influence observed in this study.Moreover, interventions should focus on addressing potential barriers and providing support to enhance individuals' perceived behavior control (e.g., skill, tools, or facility), which may in turn lead to higher intentions and increased behavioral engagement.However, perceived behavior control did not have a direct effect on behavior frequency.This finding implies that while individuals may feel confident in their ability to perform clothes upcycling, this perceived control does not necessarily translate into actual behavioral engagement, which differs from previous research that highlighted the role of PBC as a key determining factor of actual behavior performance [57].The discrepancy between perceived control and the actual behavior frequency observed in this study suggests that people should have the intention to perform the behavior first before they really perform the behavior.
On a positive note, this study has established a significant positive correlation between intention and behavior frequency, emphasizing that greater intentions to participate in clothing upcycling are linked to more frequent engagement in the behavior.This finding offers valuable implications for intervention strategies.Intervention developers could focus on cultivating a positive attitude towards clothing upcycling, encouraging individuals to adopt moral norms associated with the behavior, and equipping them with the necessary tools, skills, and ideas related to upcycling.
One practical intervention involves educational campaigns, which can significantly influence upcycling intentions by emphasizing the environmental and social benefits of upcycling.Leveraging various media channels to raise awareness and foster a sense of moral obligation towards sustainable practices can be highly effective.Additionally, skillbuilding workshops can empower individuals by enhancing their upcycling skills and confidence, given the crucial role of perceived behavioral control in shaping upcycling intentions.These workshops can be organized in community centers, schools, and online platforms to reach a broader audience.Moreover, the social aspect of these gatherings holds significant potential.By bringing people together, these workshops foster a sense of community and can induce social norms among participants.Interactions within these gatherings could reinforce positive behavior, encouraging participants to view clothing upcycling as a social norm within their community.Thus, interventions can leverage both individual and social factors to effectively promote and sustain clothing upcycling behavior.Another intervention is creating collaborative platforms where individuals can share upcycling ideas, resources, and success stories to foster a supportive community.Online forums, social media groups, and local clubs can serve as venues for such interactions, encouraging more people to engage in upcycling.

Conclusions
In conclusion, this study has revealed that individuals' attitudes toward clothing upcycling significantly influences their behavioral intentions.Despite having positive attitudes, some individuals might not intend to upcycle due to perceived costs like time and effort.Social norms, contrary to expectations, did not significantly influence upcycling intentions.Personal norms and perceived behavioral control were identified as factors that positively influence intentions regarding upcycling, indicating the importance of individual beliefs and perceptions.However, perceived behavior control did not directly impact the frequency of upcycling behavior.This implies that intentions need to precede actual engagement, as demonstrated by the significant positive effect of intention on behavior.
This study contributes to the expanding body of literature on environmentally friendly behavior, offering valuable insights for policymakers, practitioners, and researchers relating to sustainable fashion practices.The insights gained from this study can also support the development of targeted interventions and initiatives aimed at encouraging greater engagement in clothes upcycling, ultimately contributing to a more environmentally conscious society.While some hypotheses were supported, others were not, highlighting the complexity of pro-environmental behaviors and the need for targeted interventions tailored to the specific context of clothing upcycling.
However, to enhance the promotion of sustainable fashion practices, several recommendations are offered to policymakers, practitioners, and researchers.Policymakers should develop and implement policies that provide incentives for clothing upcycling activities, such as tax benefits, grants for upcycling projects, and support for community-based upcycling programs.Educational campaigns that highlight the environmental benefits of upcycling can also foster a culture of sustainability.Practitioners should focus on creating accessible upcycling workshops and skill-building programs that empower individuals with the knowledge and tools needed for upcycling.These workshops can be conducted in collaboration with local governments, educational institutions, and community organizations to maximize their reach and impact.Researchers should further investigate the nuanced factors that influence upcycling behavior across different demographics and cultural contexts.
Nevertheless, it is imperative to acknowledge the study's limitations when interpreting the outcomes.The specific cultural, economic, and social context of Japan, a well-developed country, may limit the generalizability of our findings to other countries, particularly developing nations.In different regions, the motivation model can vary significantly due to diverse determinants and socio-economic conditions.For instance, in developing countries, motivations and constraints related to clothing upcycling may be influenced by different factors compared to those in Japan.Future research should consider these variations and investigate how different contexts might alter the dynamics of consumer engagement in upcycling behavior.Another limitation lies in the use of a cross-sectional survey model, which limits our ability to observe the relationships between the constructs over time.Longitudinal or experimental designs could provide further insights into the dynamics of clothing upcycling behavior and intention over time.Additionally, the utilization of self-reported measures in the study may introduce potential biases and social desirability effects.Subsequent studies may enhance the credibility of the results by incorporating objective measures or adopting mixed-method approaches.Furthermore, this study is absent of data on the financial status and well-being of the respondents.This information was not collected by the online survey company we used, which restricted our ability to analyze the impact of economic factors on clothing upcycling behavior.Lastly, our study does not consider the influence of time and effort on clothing upcycling behavior.Future research should incorporate this factor to enhance our understanding of clothing upcycling behavior.

Figure 1 .
Figure 1.Extended version of the theory of planned behavior.

Figure 1 .
Figure 1.Extended version of the theory of planned behavior.

Table 2 .
Reliability tests for the constructs.

Table 3 .
Convergent and discriminant validity of the constructs.
Note.The values along the diagonal represent the AVE for each construct.The values below the diagonal indicate the HTMT ratio.

Table 4 .
Structural equation modelling results and confirmation of hypothesis.