Journal article Open Access

In defense of P values

Murtaugh, Paul A.


Citation Style Language JSON Export

{
  "DOI": "10.1890/13-0590.1", 
  "author": [
    {
      "family": "Murtaugh, Paul A."
    }
  ], 
  "issued": {
    "date-parts": [
      [
        2014, 
        3, 
        1
      ]
    ]
  }, 
  "abstract": "Statistical hypothesis testing has been widely criticized by ecologists in recent\nyears. I review some of the more persistent criticisms of P values and argue that most stem\nfrom misunderstandings or incorrect interpretations, rather than from intrinsic shortcomings\nof the P value. I show that P values are intimately linked to confidence intervals and to\ndifferences in Akaike's information criterion (DAIC), two metrics that have been advocated as\nreplacements for the P value. The choice of a threshold value of DAIC that breaks ties among\ncompeting models is as arbitrary as the choice of the probability of a Type I error in\nhypothesis testing, and several other criticisms of the P value apply equally to DAIC. Since P\nvalues, confidence intervals, and DAIC are based on the same statistical information, all have\ntheir places in modern statistical practice. The choice of which to use should be stylistic,\ndictated by details of the application rather than by dogmatic, a priori considerations.", 
  "title": "In defense of P values", 
  "type": "article-journal", 
  "id": "894459"
}
719
841
views
downloads
Views 719
Downloads 841
Data volume 201.7 MB
Unique views 699
Unique downloads 783

Share

Cite as