Confidence intervals and regions

Confidence intervals and regions are often preferred to point estimates of unknown parameters because they give an indication of the precision with which a parameter is known. The basic idea of confidence intervals involves statements about both observable random variables and unknown parameters. A statement about the probability distribution of random variables is reinterpreted, after the random variables have been observed, to give a confidence statement about the parameters. There are several schools of thought concerning statistical inference, the most popular of which supports confidence intervals and regions. Other schools of thought suggest that confidence intervals may have undesirable properties. At the theoretical level there is considerable controversy. However, in the most common practical situations, the most widely accepted schools of thought support very similar procedures. Thus people using confidence intervals in simple situations need not be very concerned about the soundness of the theory behind them (a). This report forms a draft contribution to the encyclopedia of statistical sciences to be published by john wiley and sons.

  • Record URL:
  • Corporate Authors:

    ARRB

    Melbourne, Victoria  Australia 
  • Authors:
    • Robinson, G
  • Publication Date: 1980-2

Language

  • English

Media Info

  • Pagination: 23P
  • Serial:
    • Issue Number: AIR 816-1

Subject/Index Terms

Filing Info

  • Accession Number: 01439128
  • Record Type: Publication
  • Source Agency: ARRB
  • Files: ATRI
  • Created Date: Aug 24 2012 10:02PM