Professional paper

What is the percentage of »false discoveries«?

B. Sorić ; Dom zdravlja »Pešćenica«, Zagreb i Odsjek za psihologiju, Filozofski fakultet Sveučilišta u Zagrebu, Zagreb, Hrvatska
B. Petz ; Dom zdravlja »Pešćenica«, Zagreb i Odsjek za psihologiju, Filozofski fakultet Sveučilišta u Zagrebu, Zagreb, Hrvatska

Abstract

The risk of discarding a true mull hypothesis is assessed by means of statistical verification of the results of scientific investigations. The way of reasoning about this risk is mistakenly applied to the risk (percentage) of obtaining »false discoveries«, e. g. to the percentage of ineffective drugs among the accepted medicaments. The percentage of mistakenly accepted results in a set of statistically significant results (»discoveries«) depends not only on the level of statistical significance but also upon the unknown situation in the set of investigated populations, or, likewise, upon the ratio of significant and insignificant results, which could be estimated, but which is neglected today.

Keywords

Hrčak ID:

153078

URI

https://hrcak.srce.hr/153078

Publication date:

19.1.1988.

Article data in other languages: croatian

Visits: 1.098 *