Development and validation of students’ competency instrument on science process skills

Abstract

Science process skills are fundamental to science learning and consequently promote scientific literacy among students. Scientists use these skills to construct knowledge during investigation. Thus, science process skills should be acquired not only by scientists, but by all individuals. This study reviews the validity and reliability of an instrument development for science process skills usinginterval rating scale as answer choice ranging from poor to excellent. This instrument was developed based on the perception of 300 respondents who were the primary school leavers, aged thirteen. As the instrument was newly developed, a pilot study was conducted to determine the empirical proof on the validity and reliability of the test. A total of 68 items were constructed in the test using verified indicators based on the experts and literature reviews. Rasch Measurement Model was employed to analyse the data. Overall, the person reliability is found at 0.97 while the item reliability is 0.99. The range of Point Measure Correlation (PMC) is positive between 0.33 to 0.67 for all items. All items are accepted as the outfit mean square (MNSQ) has the range between 0.59 and 1.55 while the infit MNSQ is between 0.74 to 1.56 indicating a good measure of latent variables for item fit. Based on the item map, the findings suggest that, for experimenting skill, only 11% of students could design scientific steps independently. Meanwhile, for using space and time relationship skill, respondents reported that the item ‘determination for the object position with time’ was the most excellent item. The scale used in this instrument was also found to adequately measure the latent constructs of science process skills.