The goal of this study was to determine how the 3 Tier Multiple Choice Diagnostic Test developed in primary students' scientific understanding. This research and development includes fundamental studies, model development, and model testing. 161 fourth-graders served as research subjects. Tests, surveys, and observations are all used to collect data. The Miles and Huberman interactive model was used to analyze qualitative data, and the paired sample t-test and N-Gain Score test were used to analyze quantitative data. The findings indicate that 3 Tier Multiple Choice Diagnostic Test instrument has a V index of 0,80 and a Cronbach's Alpha value of 0,890, indicating that the instrument is valid. Thus, all test items are legitimate and dependable. The instrument was able to identify misconceptions regarding science learning in light material that fit the following profile based on the results of the trials: 28,7% of students do not grasp the idea of light sources; 43,3% have misunderstandings; 17,8% have incomplete comprehension with misconceptions; 20,2% have incomplete understanding without misconceptions; and 14,9% have complete understanding.