Development of Multigroup Structural Equation Modeling on Structural and Measurement Models For Waste Management Behavior Patterns

This research aims to develop multigroup Structural Equation Modeling (SEM) on structural and measurement models to analyze waste management behavior patterns in Batu City. Secondary data were used from 120 respondents who were grouped into two: Group 1 (away from tourism) and Group 2 (near tourism). Latent variables include environmental quality, waste bank utilization, awareness of the use of 3R, and economic benefits from waste. The analysis was carried out by validity, reliability, linearity (Ramsey RESET), and multigroup SEM. The validity and reliability results showed that all indicators met the criteria (Corrected Item Total Correlation 0.3; Cronbach's Alpha 0.6). The linearity test proves that the relationship between variables is linear. Measurement models using formative indicators showed significant contributions, such as environmental maintenance (Group 1 coefficient: 0.369; Group 2: 0.518) and reuse effectiveness (Group 1 coefficient: 0.555; Group 2: 0.590). In the structural model, environmental quality had a stronger direct effect on 3R awareness in Group 2 (near tourism; coefficient: 0.432), while the use of waste banks had a more effect on Group 1 (away from tourism; coefficient: 0.414). The indirect effects through 3R awareness were also significant, with a total determination coefficient of 0.732, suggesting the model was able to explain 73.2% of the data variance. This study highlights the importance of a location-based approach in waste management policies, particularly the optimization of waste banks in areas far from tourism (Group 1) and the increase of 3R awareness in areas near tourism (Group 2).