Dissecting Strongly Subsuming Second-Order Mutants

Mutation testing is a fault-based technique commonly used to evaluate the quality of test suites in software systems. It consists of introducing syntactical changes, called mutations, into source code and checking whether the test cases distinguish them. Since there are dozens of distinct mutation types, one of the most challenging problems is the high computational effort required to test the whole test suite against each mutant. Since mutation testing is proposed, researchers have presented techniques aiming at effort reduction in the phases of its process. This study focuses on the potential reduction in the number of mutants provided by a special set of mutants generated by the introduction of two syntactical changes (strongly subsuming second-order mutants). In this work, we exhaustively searched for those second-order mutants Our results show that they (i) are frequently generated by the "expression removal" mutation, (ii) are likely to be killed by the same test cases that kill their constituent mutants, and (iii) have the potential to reduce the number of mutants to be executed by about 22.3%.