A mapping study of scientific merit of papers which subject are Web applications test techniques, considering their validity threats

Abstract

Progress in software engineering requires (1) more empirical studies of quality, (2) increased focus on synthesizing evidence, (3) more theories to be built and tested, and (4) the validity of the experiment is directly related with the level of confidence in the process of experimental investigation. This paper presents the results of a qualitative and quantitative classification of the threats to the validity of software engineering experiments comprising a total of 92 articles published in the period 2001-2015, dealing with software testing of Web applications. Our results show that 29.4% of the analyzed articles do not mention any threats to validity, 44.2% do it briefly, and 14% do it judiciously; that leaves a question: these studies have scientific value?