Statistics is a discipline of modelling, estimation and hypothesis testing. For today’s digital world, the black-and-white thinking is supported by the part of the concept of p-values thought to be well understood by public.
The p stands for probability, ranging from 0 to 1, or from 0% to 100%, but in decision-making based on empirical data reduced by comparing with a threshold value of statistical significance, often 5%. It is telling how likely a null hypothesis is given the observed or any even more extreme result derived from the data.
Last year, the American Statistical Association (ASA) thought it is worth and time to publish principles around the old concept of p-values that are recommended to be followed by every scientist. The paper triggered further discussions about the topic.
Estimondo put together a list of references, old and new ones, we are happy to share with you upon request via email. Also, Estimondo is currently preparing materials for a webinar to demystify the concept of statistical hypothesis testing and p-values and recommending alternatives to a broad audience free of charge. Please check out our news regularly to notice when the webinar is going to be scheduled to register for one of the spots offered.