A framework for experimental studies on the integration of software testing into programming education

• Ano de defesa: 2019
• Autor(a) principal: Scatalon, Lilian Passos
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Ensino de programação; Estudos experimentais; Experimental framework; Experimental studies; Framework experimental; Programming education; Software testing; Teste de software
• Link de acesso: http://www.teses.usp.br/teses/disponiveis/55/55134/tde-29082019-095834/

Introductory programming courses compose the core of several undergraduate degree programs, since programming is a crucial technical skill for professionals in many STEM areas. Aiming to address students learning difficulties in these courses, instructors can adopt different teaching approaches, since there are several varying aspects in programming education (e.g. programming languages and paradigms, development practices, platforms, supporting tools etc). In particular, the teaching approach that consists of integrating software testing into this context has been prominent in the area, since it may lead students to think more critically while working on programming assignments. Even so, this teaching approach can also present significant challenges, such as the students reluctance to conduct testing practices. In this sense, experimental studies have the role to provide evidence about learning outcomes, considering different teaching approaches and contexts. However, studies in the area of programming education often present a lack of theoretical basis, i.e. are not built upon established theories, models and frameworks. In other words, the varying aspects (or variables) used to investigate teaching approaches are not properly characterized in the studies, what leads to difficulties to interpret the obtained results and build knowledge in the area. As a consequence, instructors are prevented from having reliable evidence to make informed choices of teaching approaches used in the classroom. Considering this scenario, we propose in this PhD thesis the use of domain-specific models to support researchers in scoping and designing experiments on programming education. More specifically, we explored the domain of the software testing integration teaching approach, by creating a framework for experimental studies on the integration of software testing into programming education. The framework provides a basic structure of experimental studies in this domain, composed by models of variables related to this teaching approach. We also conducted experiments on the same domain and demonstrated their instantiation into the framework. We intend to support researchers and instructors in the scoping and planning of experimental studies in the educational scenario, specially those aimed at evaluating the integration of software testing into programming courses.