OUTLINE FOR CHAPTER
5 IN THE 6^{th} EDITION
Foundations of Inferential Statistics
 Introduction
 A brief review of what's done in descriptive statistics
 An overview of the basic goal of inferential statistics
 Statistical Inference
 The basic notions of (1) a sample, (2) a population, and (3) an
inference
 Two kinds of populations: "tangible" and "abstract"
 Picturing the sample, the population, and the direction of the
inference
 The Concepts of "Statistic" and "Parameter"
 4 basic questions a researcher must answer before making
an inferential guess:
 What's the population?
 How will samples be taken?
 What characteristics will be measured?
 What's the "statistical focus"?
 Two important concepts: "statistic" and "parameter"
 Representing a single concept (such as the mean) with different
symbols to signify that concept (1) in the population and (2) in
the sample
 Types of Samples
 Probability samples:
 Simple random samples
 Stratified random samples
 Systematic samples
 Cluster samples
 Nonprobability samples:
 Purposive samples
 Convenience samples
 Quota samples
 Snowball samples
 Three SampleRelated Problems That Sabbotage Statistical
Inferences
 A low "response rate":
 Why it's a problem?
 Preventing low responses rates & dealing with them once
they occur
 Refusals to participate:
 Why they create a problem
 Comparing participants with nonparticipants
 Attrition:
 Causes of this problem (sometimes called "dropout" or "mortality"
 Checking to see if attrition restricts generalization
 A Few Warnings
 There may be a mismatch between the datasuppliers & the intended
population.
 It's the quality of the sample (not its size)
that makes inferential statistics work.
 The term "random" is sometimes used when it really should not
be used.
 Statistical inferences are worthless if there's an inadequate
description of . . .
 the population, if it's a tangible population
 the sample, if it's connected to an abstract population
