SPARKing: Sampling planning after the results are known

Abstract

In null hypothesis significance testing, it is important how sample size has been determined (i.e., sample size justification) before data collection. Here we introduce a new questionable research practice related to sample size justification: Sampling planning after the results are known (SPARKing). This paper shows two representative cases of SPARKing and discusses solutions to it; however, we think that effective countermeasures are almost non-existent at this time. The main aim of this paper is to introduce SPARKing to the scientific community and to stimulate discussion.