This study compares two methods for assessing discriminant validity: the Fornell and Larcker criterion and the HTMT criterion. The research used data from a previous study involving 429 respondents to evaluate the discriminant validity of a value-based excellence model in Malaysian higher education institutions (HEIs). The model included six latent variables: leadership, culture, productivity, employee, stakeholder, and university performance. The Fornell and Larcker criterion showed that the constructs had good convergent validity, with composite reliability (CR) above 0.70 and average variance extracted (AVE) values between 0.729 and 0.839. The discriminant validity was generally acceptable, with the square root of AVE values exceeding correlations between constructs. However, minor discrepancies were found between some constructs, such as productivity-employee and productivity-stakeholder. In contrast, the HTMT criterion revealed significant discriminant validity issues. HTMT values close to 1 indicate poor discriminant validity, suggesting multicollinearity among the latent variables. Multiple construct pairs, including culture-productivity, culture-employee, and productivity-university performance, showed problems, indicating overlapping items that measure similar concepts. The study concludes that while the Fornell and Larcker criterion is widely used, the HTMT criterion is more sensitive in detecting discriminant validity issues. The HTMT criterion is recommended for more accurate assessment, as it can identify potential multicollinearity problems. The instrument used in the study needs revision to improve discriminant validity. The study highlights the importance of using the HTMT criterion for more reliable discriminant validity assessments in structural equation modeling.This study compares two methods for assessing discriminant validity: the Fornell and Larcker criterion and the HTMT criterion. The research used data from a previous study involving 429 respondents to evaluate the discriminant validity of a value-based excellence model in Malaysian higher education institutions (HEIs). The model included six latent variables: leadership, culture, productivity, employee, stakeholder, and university performance. The Fornell and Larcker criterion showed that the constructs had good convergent validity, with composite reliability (CR) above 0.70 and average variance extracted (AVE) values between 0.729 and 0.839. The discriminant validity was generally acceptable, with the square root of AVE values exceeding correlations between constructs. However, minor discrepancies were found between some constructs, such as productivity-employee and productivity-stakeholder. In contrast, the HTMT criterion revealed significant discriminant validity issues. HTMT values close to 1 indicate poor discriminant validity, suggesting multicollinearity among the latent variables. Multiple construct pairs, including culture-productivity, culture-employee, and productivity-university performance, showed problems, indicating overlapping items that measure similar concepts. The study concludes that while the Fornell and Larcker criterion is widely used, the HTMT criterion is more sensitive in detecting discriminant validity issues. The HTMT criterion is recommended for more accurate assessment, as it can identify potential multicollinearity problems. The instrument used in the study needs revision to improve discriminant validity. The study highlights the importance of using the HTMT criterion for more reliable discriminant validity assessments in structural equation modeling.