SOME HELP WITH SPSS ASSIGNMENTS 8, 4, AND BEYOND
Some of you have trouble with this
symbol, <, and confuse it
with this one, >.
A <
B
means that A is LESS THAN B.
A >
B
means that A is GREATER THAN
B.
So, p < .03 means that the probability that
a result would occur by chance is LESS THAN 3%. You would NOT write
p > .03 to communicate that
idea.
Remember that the standard we
typically use is that when p <
.05, we reject the null hypothesis and accept the alternative (research)
hypothesis. If p is NOT
< .05 (i.e., p > .05), we
RETAIN the null hypothesis, and do NOT accept the alternative (research)
hypothesis.
Example relevant to Assignment
8
Suppose I measure the height and
weight of 30 people, and calculate a correlation coefficient. Here are
some possible results, and their interpretations.
|
r(28) =
.276, p <
.13 |
Retain Null – Do not accept
Hypothesis; |
|
r(28) =
.389, p <
.07 |
Retain Null – Do not accept
Hypothesis; |
|
r(28) =
.532, p <
.02 |
Reject Null – Accept
Hypothesis; |
|
r(28) =
.651, p <
.002 |
Reject Null – Accept
Hypothesis; |
NOTE that as the correlation gets
HIGHER, the probability that the correlation is occurring by chance gets
LOWER. When that probability dips below .05 (5%), we reject the NULL
hypothesis and accept the HYPOTHESIS.
Example relevant to Assignment
4
Suppose I do a true experiment in
which I measure the heart rate of 20 people who have just finished WALKING up 10
flights of stairs, and 20 people who have just finished RUNNING up 10 flights of
stairs, to test the hypothesis that RUNNING will result in a higher heart rate
than WALKING. The results show that the mean heart rate for the RUN
group is higher than for the WALK group, but I want to know whether the
difference in the two means is great enough to warrant REJECTING the Null
Hypothesis, and thus, accepting the hypothesis. Here are some possible
results, and interpretations.
|
t(38) =
1.24, p <
.16 |
Retain Null – Do not accept
Hypothesis; |
|
t(38) =
1.57, p <
.07 |
Retain Null – Do not accept
Hypothesis; |
|
t(38) =
1.93, p <
.03 |
Reject Null – Accept
Hypothesis; |
|
t(38) =
2.41, p <
.001 |
Reject Null – Accept
Hypothesis; |
NOTE that as the t value gets HIGHER, the probability that
the difference between means found in the experiment is due to CHANCE gets
lower. When that probability dips below .05 (5%), we reject the NULL
hypothesis and accept the HYPOTHESIS.
What about degrees of freedom
(df). Degrees of freedom is a fairly complicated concept, related to the
number of scores that are contributing to a statistic. Here I will give
you the information you need for determining the df for a correlation
coefficient (Assignment 8) and a Between (or Independent) Groups t-test.
CORRELATION COEFFICIENT: df =
number of pairs of scores minus 2. This is the same as saying that df =
the number of subjects (who contribute pairs of scores) minus 2. So, if
the N is 26, the df = 24. If the N is 3,478, and df =
3,476.
BETWEEN GROUPS T-TEST: df =
total number of scores minus 2. This is the same as saying that the df =
the (number of scores in the first group minus 1) plus (the number of scores in
the second group minus 1). So, if there are 32 subjects taking blue tests,
and 25 subjects taking white tests, then the degrees of freedom for the t-test will be 55 (32+25-2) or
([32-1]+[25-1]).
Hope this
helps!
SJG
|
Steven
J. Gilbert, Ph.D.
Professor of Psychology & Department
Chair
127A Fitzelle Hall,
Phone: 607-436-2557 FAX: 607-436-3753
E-MAIL: gilbersj@oneonta.edu
Office: 127A Fitzelle Hall
Office
Hours: M 1:00; T 10:00; W 1:00; Th 11:00
(other times by
appointment)
Web Address: http://www.oneonta.edu/faculty/gilbersj/Stevepage.htm
Psychology Department Web Site: |