Random Samples and Estimation

My final warning applies to both interval estimation and point estimation--and this is by far the most important of my end-of-chapter cautionary comments. Simply stated, the entire process of estimation requires that the data used to form the inference come from a random sample. For the techniques of estimation to work properly, therefore, there must be a legitimate connection between the sample and the population such that either (1) the former is actually extracted, randomly, from the latter (with no refusals to participate, mortality, or response rate problems); or (2) the population, if hypothetical, is conceptualized so as to match closely the nature of the sample. Without such a link between sample and population, neither form of estimation can be expected to function very well.

(From Chapter 6 of the 6th edition, pp. 129-130)

Copyright © 2012

Schuyler W. Huck
All rights reserved.

| Book Info | Author Info |

Site URL: www.readingstats.com

Top | Site Map
Site Design: John W. Taylor V