Analysis of students’ critical thinking skills through science, technology, engineering and mathematics (STEM) approach

Abstract

This research aims to find out the process of implementing science learning with STEM approaches and to find out students' critical thinking skills with STEM approaches. The methods used in this research are qualitative methods. Instruments used in the form of observations, tests, questionnaires and documentation. Data analysis in this study includes the data reduction stage, the display data stage, and the conclusion drawing stage. The results of this study concluded that the process of implementing science learning with a STEM approach consists of four steps: basic competency analysis and identification of appropriate topics, formulating competency achievement indicators (critical thinking indicators used: FRISCO indicators presented by Robert Ennis (1996)), designed lesson plan and instruments, and the implementation of learning with a STEM approach to knowing students' critical thinking skills. The critical thinking ability of students with STEM approaches, it is found that the critical thinking skills of students in very high categories as much as 31.25% with FRISCO indicators are mostly all visible. 53.125% are in the category of high critical thinking skills, with prominent critical thinking indicators being Focus and Reason, while inference, situation, clarity and overview indicators are sufficient. Students with moderate critical thinking skills as much as 9.375% critical thinking indicators that stand out Focus only while for reason, inference, situation, clarity and overview indicators are still lacking. Students with a low critical thinking ability category of 6.25%, all FRISCO indicators are so lacking that they are still less prominent and yet to be seen. This research contributes that STEM learning can be used to develop students' critical thinking skills, and the development of this research can be continued by developing of evaluation sheet, teaching materials, learning media, and learning models.