The impact of tinkering and planning behaviors on learning introductory programming concepts

• Ano de defesa: 2020
• Autor(a) principal: Silva, Fernando Henrique Carvalho
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Agrupamento; Ciência da computação; Clustering; Computer science; Computer science in education; Data mining; Informática na educação; Mineração de dados; Problem solving; Resolução de problemas; Tinkering
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/55/55134/tde-26082020-164043/

Computer programming skills are becoming essential across different fields and disciplines, and as integrating part of computer science basic formation curriculum, these skills are considered problem-solving activities. Problem-solving is an internal cognitive process of a high order that searches for a solution for a problem or a path to a given goal. As an internal process, problemsolving strategies cannot be observed directly and must be inferred. That necessity led several researchers to identify and classify those strategies based on students behaviors. Among those researches, the classification between tinkerers and planners proposed by Turkle and Papert was used in this work. In this classification, planners are students who present a structured, top-down strategy to solve problems, which is commonly reinforced in computer science environments. On the other hand, tinkerers employ an alternative, less structured, bottom-up problem-solving strategy that uses the systems feedback to achieve their goals. To Investigate the students tinkerer and planner behaviors in online environments and its impact on the students grades performance, firstly an educational data mining approach using clustering was employed to identify the students behaviors and the features related to tinkering, followed by an assessment of the students grades and the time that they required to finish the assignments. As a result, four behaviors were identified related to Turkle and Papert classification. Statistical tests showed that students that employed a higher amount of tinkering presented a better performance in final tests and required more time to finish their online assignments. These results indicate that tinkering is a valid problem-solving strategy with potential that worth being explored by educators.