OUTLINE FOR THE 1^{st} HALF OF CHAPTER
7 IN THE 6^{th} EDITION
Hypothesis Testing
(NOTE: This outline covers pages
131-144 of Ch. 7. A different outline covers pages 144-159 of
this chapter.)
- Introduction
- The goal in hypothesis testing (HT): Make educated guesses about unknown population
parameters
- Overview of the chapter
- The 4 preliminary questions that must be answered before HT starts
- An Ordered List of the 6 Steps
- A Detailed Look at Each of the 6 Steps . . . Looked
At "Out of Order"
- Step #1: The Null Hypothesis
- The definition of a null hypothesis . . . and its symbolic
representation
- Where the null hypothesis comes from
- The notion of Ho positioned
on a "continuum of possible values"
- Examples from real studies
- Sometimes Ho is set up to
be a "no difference"
statement; sometimes it's not
- Ho & the researcher's
hunch; they can be the same, opposites, neither
- Step #6: Deciding What To Do With the Null Hypothesis
- Two options: Reject Ho or
fail-to-reject Ho
- The different ways researchers talk about having rejected/not
rejected Ho
- Step #2: The Alternative Hypothesis
- How it's symbolically represented . . . and its necessary
connection to Ho
- The directional (one-sided) and nondirectional (two-sided)
option for Ha
- The directional/nondirectional option and one-tailed vs. two-tailed
tests
- Step #4: Collection and Analysis of Sample Data
- The basic logic of hypothesis testing: State Ho,
then collect data; Reject Ho if
data are inconsistent
with Ho, otherwise, fail-to-reject
Ho
- Two ways of summarizing the sample data into a single numerical
value:
- Converting the sample data into a standardized number
that's called a "calculated value" (or "test statistic"),
such as "t = 2.91" or "F = 12.73"
- Letting a computer determine p, the probability
of having sample data that deviate as much or more from
Ho
as do the sample data, assuming for the moment that Ho
is true
Note: Steps #5 & #3, along with
2 other facets of HT, are covered on the outline for the 2nd half of Chapter
7 |