It is important to teach students programming at an early age so that they do not have difficultyin understanding the programming logic when they reach the age of undergraduate education.Scratch, a visual two-dimensional programming tool, has been developed to accomplish thispurpose. There are studies in the literature related to Scratch on problem-solving skills, but theyare not directed at the steps of problem solving. In order to find a proper solution to a problem, itis necessary to carry out each problem-solving step in an appropriate way. Based on this need,the aim of this study is to investigate how teaching programming through Scratch affects theproblem-solving skills of 5th and 6th grade middle school students and to identify the steps thatstudents go through to solve ill-structured problems. It is also intended that these steps guideteachers in terms of the pedagogies for teaching coding. This study utilized an explanatorydesign, one of the mixed method research designs. In the quantitative stage of the study, a one-group pretest-posttest model, one of the simple experimental models, was used. In thequalitative stage, a case study model was used as the research design. A total of 226 5th and 6thgrade students, including 109 males and 117 females, constituted the sample of the study. Thedata were collected using a problem-solving skills scale and interview questions. A statisticallysignificant difference was found between the mean scores of the first and second measurementsof the problem-solving skills scale according to the Wilcoxon signed-rank test result. According tothe results, when teaching programming through Scratch, it is recommended to have studentsdesign games, as an alternative method, for them to develop problem-solving skills. The solutionof a problem is facilitated when students construct a game in their minds in accordance with real-life conditions and design and detail it in accordance with computational expressions.
