This study developed and implemented a learner-centered assessment named Creative Exercises (CEs) in introductory chemistry courses at a four-year university and a two-year community college. CEs were developed and implemented as an intervention for treatment groups. The control groups only used traditional assessments such as multiple-choice and short-answer questions. A mixed-methods approach was employed for evaluating the effectiveness of CEs in improving student learning and performance. First, quantitative data included student exam scores, DFW rates, and percentages of letter grades were analyzed and compared between treatment and control groups. Second, student responses to CEs were coded as chemistry concepts and then organized into chemistry topics. A series of visual maps were plotted to show students' linking of chemistry topics and progress made throughout the semester. Lastly, student perceptions of the use of CEs were investigated "via" a free-response survey. Quantitative results showed that CEs improved students' academic performance and retention in introductory chemistry courses at both college settings. The implementation at the two settings indicated that the frequency and quality of the use of CEs might impact the effectiveness. The results from qualitative data analyses converged with the positive effects of CEs. Students were able to connect prior and newly-learned topics in response to CEs and made progress on linking more topics over time. Open coding of the free-response survey data identified four themes explaining why the use of CEs helped students: knowledge integration, conceptual understanding, flexibility, and more effective study habits. Best practices for implementation of learner-centered assessments learned in this study and future directions for research are discussed.