ABSTRACT
This chapter focuses on the widespread practice of treating as generalizable insights extracted out of samples and analyses. Moreover, decision-makers’ desire to impose order by finding coherence in barrages of distinct characteristics or events can manifest itself in a yet another form of self-deception known as management lore. From the risk management-related decision-making point of view, granular numeric estimates are ultimately interpreted in a more general context of conclusive categories such as ‘low’, ‘average’, or ‘high’ because risk response choices distinguish among categorically, not merely mathematically different values. When considering data usage from insight generation point of view, a commonly encountered problem stems from the use of statistical significance tests (SST), as a way of validating the truthfulness – in the sense of correctness and generalizability – of sample-based conclusions. SSTs are a hypothesis testing tool, the purpose of which is to identify universally true effects.
