Presenting a hypothesis suggested by inspected results as if it had been specified beforehand. The misrepresentation, rather than discovery itself, is the problem. Transparent exploratory analysis, preregistration and independent testing preserve the distinction between generating and evaluating a hypothesis.[ref]
Registered Reports by Center for Open Science. Details: HARKing; Scientific Creativity and Exploration
Articles by others on the same topic
HARKing stands for "Hypothesizing After the Results are Known." It refers to the practice of formulating hypotheses based on the outcomes of data analysis rather than developing those hypotheses beforehand. This can lead to misleading conclusions because it undermines the integrity of the scientific method, whereby hypotheses should ideally be established prior to data collection and analysis.