Statistics Quantum Satis

Dinis Pestana

Abstract


Statistics is a privileged tool in building knowledge from information, since the purpose is to extract from a sample limited information conclusions to the whole population. The pervasive use of statistical software (that always provides an answer, the question being adequate or not), and the absence of statistics to confer a scientific flavour to so much bad science, has had a pernicious effect on some disbelief on statistical research. Would Lord Rutherford be alive today, it is almost certain that he would not condemn the use of statistics in research, as he did in the dawn of the 20th century. But he would indeed urge everyone to use statistics quantum satis,
since to use bad data, too many data, and statistics to enquire on irrelevant questions, is a source of bad science, namely because with too many data we can establish statistical significance of irrelevant results. This is an important point that addicts of evidence based medicine should be aware of, since the meta analysis of two many data will inevitably establish senseless results.

Full Text:

PDF (Português)


Refbacks

  • There are currently no refbacks.