SUMMARY
Objective: The research objective was to identify and analyze the relationship between student ability and project-based learning (PBL) with increased learning, the relationship between critical thinking ability and project-based learning with increased learning, and the relationship between project-based learning and learning improvement. project-based learning mediates the relationship between student ability and learning improvement and mediates the relationship between critical thinking ability and learning improvement.Methodology: This research is a type of quantitative research where it examines the relationship between exogenous and endogenous variables. This quantitative study uses the SEM-SmartPLS 3.0 model and also uses a quasi-experimental design method because it observes two groups, namely the control group and the experimental group.Findings: Based on the results of hypothesis testing using Smart PLS it shows that, H1, H2, & H3, are accepted because the t table value is greater than the calculated t value or has an alpha (a) value less than 5%. Whereas H4, H5, H6, and H7 cannot be accepted (rejected) because their value is greater than 5%. The experimental class showed an increase with the previous average value of 54.31 while the average result after treatment was 78.79 so there was an increase of 24.48. The homogeneity of the data was tested with the homogeneity test of variance with the significance value of 0.693 > 0.05.Conclusion: Student Ability has a significant effect on PBL and learning improvement. critical thinking ability has a significant effect on PBL and learning improvement. PBL has no significant effect on learning improvement. PBL does not mediate the relationship between student ability and learning improvement, and the relationship between critical thinking ability and learning improvement. The ability to think critically in learning to write explanatory texts with causal patterns using the Problem-Based Learning model shows good results. This can be seen from the students' scores in the experimental class that has increased after being given treatment (controlled).