Review: Statistical Power
Hypothesis testing with the t-test
A new type of t
http://www.stat.sc.edu/%7eogden/javahtml/
power/power.html
z
=
M
±
±
²
M
t
=
M
±
±
s
M
z
=
M
±
±
²
2
n
t
=
M
±
±
s
2
n
Sample variance
Population variance
testStatistic
=
Difference between data and hypothesis
standard distance expected by chance

This
** preview**
has intentionally

**sections.**

*blurred***to view the full version.**

*Sign up*
±
M
=
±
n
=
±
2
n
sample variance
=
s
2
=
SS
n
±
1
=
SS
df
M
s
=
s
2
n
Estimated standard error
t-test
population variance not
known
z-test
population variance known
1.
State the hypothesis (null and alternative)
2.
Set the decision criterion
Locate the critical t for the speci
ed alpha using the appropriate df
and table B.2
3.
Collect sample & compute statistic
compute mean, sample variance, and t-statistic
4.
Compare statistic to criterion and make a
decision
apply same logic as with a z-test (if obtained t falls in critical region,
then reject the null)
36 workers are switched to a compressed schedule (3
consecutive 13-hour shifts). They rate their preference on a
scale from 1-10 (0=prefer standard, 5=no preference, 10=
prefer compressed).

This is the end of the preview.
Sign up
to
access the rest of the document.