Assessing students' mathematical problem-solving skills: framework for developing tests on the pythagoras theorem

Abstract

Good mathematical problem-solving skills are acquired through practice and training. This can be achieved by regularly administering tests that measure students' mathematical problem-solving skills. The development of this test aims to obtain a test that is capable of measuring students' mathematical problem-solving skills. The Research & Development (R&D) approach was used to develop a test instrument for mathematical problem-solving skills on the Pythagorean theorem. Seven stages adapted from the Borg & Gall development model were followed to obtain a good test instrument that met the test objectives. The total sample size for the trial was 87 eighth-grade students from a junior high school in Pamekasan Regency. The sample was selected using random sampling techniques. The instrument consisted of an expert validation sheet and a test sheet that was developed. Data were collected through expert validation and pilot testing. The expert validation results were analyzed descriptively to measure the theoretical validity of the developed test. Meanwhile, the pilot test data were analyzed using Classical Test Theory (CTT) parameters, which included empirical validity, reliability, difficulty level, and discrimination power tests. The results showed that each test item was valid, reliable, had different levels of difficulty, and had good discriminating power. Thus, the developed test instrument met the criteria for a good test and could be used to measure mathematical problem-solving skills.