Efectividad de técnicas de prueba de software aplicadas por sujetos novicios de pregrado
It highlights a methodological issue in software engineering experiments by showing that using inexperienced student subjects yields poor results, suggesting the need for more skilled participants.
This study examined the effectiveness of black-box and white-box software testing techniques when applied by novice undergraduate students, finding both techniques had low observed defect effectiveness (around 4%) and black-box testing was slightly more effective for observable defects (21% vs. 16%).
The main objective of this work is to examine possible effects of using freshman student subjects in software engineering experiments. Particularly in this work we report the effectiveness measured as percentage of observed and observable defects of two software testing techniques: Black-box and white-box. Regarding observed defects, both techniques show an effectiveness around of 4%. With respect of observable defects by test cases, black-box testing is slightly more effective (21%) than white-box testing (16%), although this difference is not significant. We observe a considerable lack of technical skills of subjects for applying both software testing techniques. Due to observed findings, we suggest to employ students with more technical skills for carrying out software engineering experiments. ----- El objetivo de este trabajo se centra en investigar los efectos que conlleva realizar experimentos en ingeniería de software (IS) empleando como sujetos experimentales a estudiantes de pregrado cursando su primer año de estudios de la carrera en ingeniería de software. De manera particular en este trabajo se investiga la efectividad medida en porcentaje de defectos observados y observables de las técnicas de prueba de software funcional (caja negra) y estructural (caja blanca). Con respecto a los defectos observados por los sujetos, ambas técnicas obtuvieron una efectividad del 4%. Con respecto a los defectos observables por los casos de prueba, la técnica funcional es ligeramente superior (21%) que la técnica estructural (16%), aunque esta diferencia no es significativa. Se observa un nivel de inexperiencia considerable en los sujetos para aplicar las técnicas. Dado los hallazgos encontrados, se sugiere emplear sujetos de pregrado con un nivel mayor de experiencia.