Mechanism of attitude, subjective norms, and perceived behavioral control influence the green development behavior of construction enterprises

The study addresses the challenge that construction enterprises—due to high pollution, emissions, and energy consumption and the inherent mobility/long cycles of projects—face in achieving green development. It applies the Theory of Planned Behavior (TPB) to understand how cognition (attitudes, subjective norms, perceived behavioral control) drives intention and, in turn, green development behavior. The research questions are whether TPB explains the antecedents of green development behavioral intention and behavior in construction enterprises and how regional green development level and enterprise size moderate these relationships. The purpose is to construct and empirically test a TPB-based model for green development behavior among construction firms in China, offering guidance for policy and managerial practice. The study is important because the construction sector significantly impacts decarbonization targets and requires effective behavioral mechanisms to support balanced economic growth and environmental protection.

The paper grounds its framework in Ajzen’s TPB, where attitude, subjective norms, and perceived behavioral control (PBC) predict intention, which predicts behavior. Prior research in environmental contexts has validated TPB across domains such as construction waste management, green supply chains, and green technology innovation. Studies show attitude often has a strong positive effect on environmental behavioral intention; subjective norms (e.g., regulatory pressure, media, consumers) motivate environmental intentions; and PBC (resource/capability perceptions) influences intentions and sometimes behaviors. However, existing work on construction enterprises has not fully examined intention as a mediator or moderators shaping green development behavior. This study advances the literature by: (1) applying TPB to organizational green development behavior in construction; (2) specifying and validating measurement scales (attitude, subjective norms, PBC, intention, behavior); and (3) testing moderation by regional green development level and enterprise size using multigroup analysis.

Design: A structured questionnaire based on TPB and prior scales was developed and refined via expert review (five professors/PhDs) and a pilot test (n=50; Cronbach’s α=0.949, KMO=0.738). The final instrument measured five latent constructs—attitude (6 items), subjective norms (5), perceived behavioral control (5), behavioral intention (7), and behavior (13)—with a 5-point Likert scale (1=strongly disagree to 5=strongly agree). Sampling and data collection: Random sampling targeted construction enterprises (e.g., contracting, installation, decoration, mechanized construction, engineering, and other specialized firms) across China. Data were collected via electronic questionnaires from January to March 2022. Inclusion criteria spanned firm types, sizes (small/micro, medium, large based on revenue/assets thresholds), and positions (senior, middle, grassroots). Of 419 returns, 306 valid responses remained (effective rate 73.03%) from 28 provinces/cities (excluding Hong Kong, Macao, Taiwan, Hainan, Qinghai, Tibet). Sample characteristics (selected): 56.54% female; age <30 (55.23%), 30–39 (42.48%); education mainly specialized/undergraduate (89.87%); positions: grassroots (61.76%), mid-level (33.01%), senior (5.23%); regional green development level: high (64.38%), low (35.62%); enterprise size: large (36.60%), medium (38.89%), small/micro (24.51%). Measures and analysis: PLS-SEM (SmartPLS 3.0) was employed to estimate measurement and structural models, suitable for complex models and non-normal data. Measurement model: reliability indicated by Cronbach’s α (0.779–0.921 per construct; overall 0.945) and composite reliability (CR=0.850–0.932). Validity: content validity ensured through expert review; structural validity supported with KMO (overall 0.928; constructs 0.823–0.926), AVE=0.514–0.543 (>0.5), and discriminant validity via HTMT=0.495–0.834 (<0.85). Common method variance tested via Harman’s single-factor test (largest factor=35.503%<40%), indicating CMV not severe. Structural modeling: The TPB-based model tested paths from attitude, subjective norms, and PBC to intention (H1–H3), PBC to behavior (H4), intention to behavior (H5), and intention’s mediating role (H5a–H5c). Moderation by regional green development level (high vs. low; H6a–H6e) and enterprise size (small, medium, large; H7a–H7e) was examined using PLS-MGA with 5,000 bootstrap iterations. Model fit assessed via R² for intention and behavior.

Model fit: R² for endogenous constructs were 0.614 and 0.522, indicating good explanatory power. Direct effects: H1–H3 and H5 supported; H4 not supported. - Attitude → Intention (H1): significant positive effect; attitude is the strongest predictor of intention. - Subjective norms → Intention (H2): significant positive effect. - PBC → Intention (H3): significant positive effect. - PBC → Behavior (H4): negative, not significant (rejected). - Intention → Behavior (H5): significant positive effect. Mediation (bootstrapping, 5,000 resamples): - Attitude → Intention → Behavior (H5a): full mediation; VAF=99.34%; total effect=0.305***; direct=0.002 (ns); indirect=0.303***. - Subjective norms → Intention → Behavior (H5b): partial mediation; VAF=71.81%; total=0.227***; direct=0.064; indirect=0.163***. - PBC → Intention → Behavior (H5c): full mediation with suppression; VAF=116.13%; total=0.155***; direct=−0.025 (ns); indirect=0.180***. Moderation (PLS-MGA, path coefficients by group): - H1 (Attitude → Intention): Small 0.445***; Medium 0.506***; Large 0.360***; Low-region 0.429***; High-region 0.429***. Supports H7a and H6a. - H2 (Subjective norms → Intention): Small 0.135 (ns); Medium 0.176*; Large 0.385***; Low-region 0.197*; High-region 0.266***. Supports H7b and H6b. - H3 (PBC → Intention): Small 0.391***; Medium 0.159 (ns); Large 0.260***; Low-region 0.361***; High-region 0.204**. Supports H7c and H6c. - H5 (Intention → Behavior): Small 0.802***; Medium 0.668***; Large 0.769***; Low-region 0.648***; High-region 0.752***. Supports H7e and H6e. Notes: Because H4 (PBC → Behavior) was not significant, related moderation hypotheses H6d and H7d were not supported. Overall, attitude is the strongest determinant of intention; intention is the primary driver of behavior; PBC influences behavior indirectly via intention; and both regional green development level and enterprise size positively moderate most TPB links.

The findings validate TPB in the organizational context of construction enterprises: attitudes, subjective norms, and PBC shape behavioral intentions, which then translate into green development behaviors. Attitude emerged as the strongest predictor, highlighting the centrality of internal, intangible resources (e.g., green culture, managerial commitment) in mobilizing action. PBC did not directly affect behavior, suggesting that resources and capabilities alone do not trigger green actions without the motivational bridge of intention—consistent with the profit-driven nature of firms where intention aligns resources with strategic priorities. The mediating role of intention—full for attitude and PBC, partial for subjective norms—clarifies how internal drivers and external pressures convert into behavior. Moderation analyses show stronger TPB links in regions with higher green development levels and notable differences by enterprise size: large firms show stronger effects from subjective norms (likely due to heightened scrutiny), while small firms’ intentions more strongly predict behavior (perhaps reflecting agility). These results inform both theory (extension of TPB to firm-level green behavior with moderated relationships) and practice (focus on cultivating pro-green attitudes and capabilities while leveraging regional contexts and tailoring strategies by firm size).

This study develops and empirically tests a TPB-based model of green development behavior in construction enterprises using PLS-SEM and survey data from 306 firms across China. It shows that attitude, subjective norms, and perceived behavioral control significantly enhance green behavioral intention, with attitude being the strongest predictor; intention significantly drives behavior; and intention mediates the effects of the three antecedents on behavior (fully for attitude and PBC, partially for subjective norms). The direct effect of PBC on behavior is not supported. Multi-group analyses reveal that both regional green development level and enterprise size positively moderate most TPB relationships, underscoring contextual contingencies. Contributions include extending TPB to an organizational setting in the construction industry, validating comprehensive measurement scales, and evidencing moderated mechanisms. Practically, the study suggests managers should build a strong green culture, align resources to support intentions, and leverage regional ecosystems while tailoring approaches by firm size. Future research could employ longitudinal designs, broaden to cross-industry and cross-country contexts, and further unpack multilevel mechanisms linking cognition, intention, and behavior.

- The mediating mechanism of intention, while tested, was not explored in depth across dynamic or longitudinal settings; cross-sectional data limit causal inference. Future work should use longitudinal data and SEM to examine temporal mediation. - The study focuses on construction enterprises in one country (China), limiting generalizability. Cross-industry and cross-national comparisons are needed to test model robustness and universality. - Measurement relies on self-reported survey data, which, despite CMV checks, may still be subject to bias.