Lecture 20
Types, Tales and Power
I)
Analysis
of Errors
A)
Type 1
error-rejection of the null hypothesis when it is true.
Alpha
α
is probability of type 1 error. (divorce)
B)
Type 2 error-
retention when the null is false.
(old maid)
II)
Power of the test
1) Actually Difference (get big)
2) The Standard error of a
difference between means (get small)
a.
Sample size (large)
b.
Sample variability (low)
c.
Alpha (.05 convention)
Practice:
As a result of some of his research in the 18the century, E.L. Thorndike
concluded that animals were incapable of learning by imitation. In 1901, L.L.
Hobhouse reported the results of experiment with cats, dogs, otters, elephants
and monkeys indicating that they could learn
by imitation, Suppose the following study was carried out. One group of
randomly selected hungry cats is shown food being obtained from under a vase.
Another random selected group is not. Shortly afterwards, time (in seconds)
required to upset the vase is recorded for each animal. Results are given below.
Do the results support either of the two theorists?
Shown:
18,15,15,12,11,9
Not Shown
25,22,21,19,16,15