The effects of argumentation and authentic learning-based STEM activities on academic achievement were examined in the study. Prior to the implementation, a 120-minute teacher training was conducted with two science teachers every day for nine days. The study was conducted with 120 fifth-grade students in two different public schools for 17 weeks. Purposive sampling was used in the study, and convergent parallel design from mixed method research was utilized. In the quantitative dimension of the study, a quasi-experimental design with a pretest-posttest control group was used. One control group and two experimental groups (argumentation and authentic) were determined in both schools. In the qualitative dimension, a case study design was used. Quantitative data collection tools included academic achievement tests on the propagation of light, human and environment, and electrical circuit elements; qualitative data collection tools included video and audio recordings, semi-structured interviews, teacher and researcher observation forms, and student diaries. As a result of the study, significant differences were found between the pre-test and post-test scores of the three academic achievement tests, especially in favor of the authentic group. In the qualitative findings, expressions were also identified in which students were seen to be affected by the approach towards the units. In the last part of the study, various suggestions regarding the findings are given.