EXPERIMENTAL INVESTIGATION AND AGILE PRACTICES: THREATS TO THE VALIDITY OF EXPERIMENTS INVOLVING THE PAIR PROGRAMMING AGILE PRACTICE

Abstract

Even though the process of investigation by experimentation, in the context of software engineering, is not something new, it still lacks new quality studies. In studies that adopt the experiment as a method of research, a complete and detailed evaluation of the threats to the internal, external, construct and conclusion validity is an extremely important quality factor. Based on a detailed analysis of two empirical studies, this work presents a list of 21 possible threats to the validity of an experiment involving the pair programming agile practice, giving an exemplification of 17 possible threats.