Modeling the reuse intention and practices of secondhand clothing: evidence from a developing nation

The study addresses escalating environmental impacts from the textile and garment industry and the rapid growth of secondhand markets, especially in China. While prior research has focused mainly on purchasing/using secondhand clothing in developed contexts and relied on the Theory of Planned Behavior (TPB), this paper investigates selling or swapping practices in a developing context (China) and incorporates affective elements often overlooked by TPB. Grounded in the Theory of Interpersonal Behavior (TIB), the study aims to evaluate Chinese consumers' intention to reuse secondhand clothing by integrating environment-related constructs into TIB. The research proposes and tests hypotheses that perceived values on sustainability (PVS), problem awareness (PRA), ascription of responsibility (ASR), sense of community (SOC), and anticipated guilt (ATG) positively influence reuse intention (RUI), and that RUI positively influences reuse practices (RUB).

The paper grounds its framework in the Theory of Interpersonal Behavior (TIB), which posits that intentions are shaped by cognition/attitudes, social factors, and affect, offering a broader lens than the more rationalist, static TPB. Cognition is represented by perceived values on sustainability (PVS) and problem awareness (PRA), social factors by ascription of responsibility (ASR) and sense of community (SOC), and affect by anticipated guilt (ATG). The review synthesizes evidence that: (a) PVS (closely related to green perceived value) is associated with pro-environmental intentions; (b) PRA can heighten concern and encourage pro-environmental behaviors, though results vary by context; (c) ASR fosters moral obligation and pro-environmental actions; (d) SOC aligns individuals with community norms favoring environmental protection; and (e) ATG, as a negative anticipatory emotion, can motivate sustainable intentions. Hypotheses H1–H5 posit positive effects of these five constructs on RUI; H6 posits that RUI positively predicts RUB.

Design: Cross-sectional, quantitative survey analyzed using partial least squares structural equation modeling (PLS-SEM) in SmartPLS. Sampling and data collection: Target population was Chinese adults (≥18 years). Convenience sampling was used due to the absence of a complete sampling frame. A priori power analysis (G*Power 3.1) indicated a minimum sample of 146 (six predictors, power 0.95, effect size 0.15). Data were collected via WJX (https://www.wjx.cn/) and distributed through WeChat/QQ between June 14 and July 29, 2022. Of 888 responses, after screening (excluding those without selling/swapping experience, too-short completion time, and uniform responses), 514 valid questionnaires remained. Measurement: The questionnaire had demographics and construct measures. Items were adapted from prior studies: PVS (Kim et al., 2020; Han et al., 2019); PRA, ATG, RUB (Attiq et al., 2021a); ASR (Kim et al., 2021); SOC (Li-Chun, 2018; Huang et al., 2020); RUI (Borusiak et al., 2020a; Kim et al., 2021). Seven-point Likert scales were used. The English survey was translated and back-checked by two specialists for the Chinese context; a pilot test (n=25) preceded final data collection. Common method bias: Full collinearity assessment showed VIFs <5 for all constructs (PVS 1.774; PRA 2.373; ASR 2.11; SOC 2.243; ATG 1.927; RUI 1.792; RUB 2.851), indicating acceptable CMB. Distributional assumptions: Mardia’s multivariate skewness and kurtosis tests (Web Power tool) had p-values <0.05, indicating non-normality; thus PLS-SEM was appropriate. Analysis: Measurement model assessed reliability and validity (Cronbach’s alpha, rho_A, composite reliability, AVE, VIF, HTMT, Fornell-Larcker, loadings). Structural model used bootstrapping (one-tailed, 5000 iterations) to test paths. Multi-group analysis (MGA) with MICOM invariance testing was conducted across gender (male vs female), age (≤30 vs >30), education (undergraduate or below vs master’s and above), and income (≤RMB4500 vs >RMB4500).

Measurement model: All constructs exhibited high reliability and convergent validity (alphas ≥0.951; composite reliabilities ≥0.962; AVEs ≥0.835). Discriminant validity was supported (Fornell-Larcker and HTMT). Structural model: R² for RUI = 0.401; R² for RUB = 0.365. Effects on RUI: PVS → RUI β=0.106, t=2.081, p=0.019, f²=0.011 (small); PRA → RUI β=0.090, t=1.571, p=0.058, f²=0.007 (ns); ASR → RUI β=0.126, t=2.132, p=0.017, f²=0.013 (small); SOC → RUI β=0.206, t=3.281, p=0.001, f²=0.035 (moderate); ATG → RUI β=0.248, t=4.431, p<0.001, f²=0.057 (moderate). Effect on behavior: RUI → RUB β=0.604, t=12.802, p<0.001, f²=0.575 (substantial). Common method bias and multicollinearity were not problematic (VIF range ≈1–2.851). Multi-group analysis: Significant group differences observed for: gender—PRA→RUI and ASR→RUI (p<0.05); education—PVS→RUI (p=0.016); income—PRA→RUI (p=0.033). No significant differences found for age groups or for other paths across groups.

Findings support TIB’s inclusion of cognitive, social, and affective determinants of intention in the pro-environmental context of secondhand clothing. Perceived sustainable value (PVS) increases intention to sell/swap secondhand clothing, highlighting the role of perceived environmental utility. Ascription of responsibility (ASR) fosters a moral obligation that translates into higher reuse intention. Sense of community (SOC) emphasizes social belonging and normative influence in encouraging reuse. Anticipated guilt (ATG) emerged as the strongest predictor among antecedents, underscoring the importance of affective, self-conscious emotions in motivating sustainable intentions. Contrary to several prior studies, problem awareness (PRA) did not significantly predict intention—possibly due to message fatigue and desensitization regarding environmental problems—suggesting that information provision alone may be insufficient. Intention robustly predicted actual reuse practices (RUB), aligning with intention-behavior links. Group differences indicate that demographic factors moderate some antecedent-intention relationships, implying the need for targeted interventions.

The study advances understanding of secondhand clothing circularity by focusing on selling and swapping behaviors in China and by extending TIB with environmental constructs. Results show that PVS, ASR, SOC, and ATG significantly enhance reuse intention, which in turn strongly promotes reuse practices; PRA was not a significant predictor. The theoretical contribution lies in integrating cognitive, social, and affective elements into a TIB-based model tailored to pro-environmental consumer behavior, addressing TPB’s limitations regarding affect. Practically, the findings inform policymakers and platforms to design interventions that elevate perceived sustainable value, foster responsibility, leverage community dynamics, and ethically engage anticipatory guilt to increase reuse. Future research should examine additional circular fashion behaviors (e.g., renting, redesign), adopt longitudinal and probability-based designs for generalizability, and test moderating roles of demographics to clarify directionality in subgroup effects.

Cross-sectional, non-probability sampling limits causal inference and generalizability. Focus centered on selling and swapping, potentially omitting other circular options (e.g., renting, redesign). While MGA indicated some demographic influences (e.g., gender, education, income) on specific paths, the precise directions require further testing; these characteristics could be modeled as moderators in future research.