Abstract:
|
Null hypothesis significance testing (NHST) and the P-value have become the cornerstone of research. However, they are often misunderstood and misused. In particular, the critical level of significance of 0.05 has become sacrosanct. A P-value smaller than 0.05 is often interpreted incorrectly as proof the null hypothesis is false. Furthermore, statistical significance is seen as essential for the success of research studies and subsequent publication, ensuring career development. It is without doubt that such misunderstanding and misuse contributed to the recent ban of NHST and reporting of P-values by the editors of a psychology journal. It was claimed that NHST was "invalid" whilst the "P< 0.05 bar is too easy to pass". The ban reignited debate in the statistical community about the value of the P-value. What are the consequences of the ban? Was it justified and does the ban signify the death of the P-value? Or were the editors of the journal also misinformed? The presentation will consider how NHST and the P-value have become to be misused, plus arguments for and against their use. The ethical responsibilities of the statistical community will also be considered.
|