Selecting the Appropriate Statistic (Part 1)
Used to test
hypothesis about: Number of Groups Question to be
Answered (Verbal
statement of the
hypothesis)
Statistic to Choose Null and
Alternative
Hypothesis
Means One Is the mean of the
group significantly
different from some
number?
Single Sample t-test H0: µ = number
H1: µ … number
or
H1: µ — number
or
H1: µ ™ number
example
H0: µ = 30
Two independent
(separate) groups Is the mean of one
group significantly
different from the
mean of the other?
Between Subjects
(Independent
Samples) t-test
H0: µ1 = µ2
H1: µ …µ
or
H1: µ—µ
or
H1: µ™µ
Two groups of
matched subjects or
two measures on
the same subjects
(repeated measures)
Is the mean of one
set of scores
significantly
different from the
mean of the other
set of scores?
Within Subjects
(repeated measures
or dependent
groups) t-test
H0: µ1 - µ2 = 0
H1:µ1 - µ2 … 0
Ranks Two independent
(separate) groups Are the ranks
assigned by one
group higher or
lower than the
ranks assigned by
the other group?
Mann-Whitney U
Test HO: The ranks are
no different
H1: The ranks are
higher or lower.
Two groups of
matched subjects or
two measures on
the same subjects
(repeated measures)
Are the ranks
assigned in one
situation higher or
lower than the
ranks assigned in
another situation?
Wilcoxon Matched
Pairs Signed Ranks
Test
HO: The ranks are
no different
H1: The ranks are
higher or lower.
Shape of the
Distribution
Frequency
Distribution
One group/set of
scores Are the scores
normally
distributed?
Or
Are the scores
distributed in some
specified way?
Chi Square
goodness of fit H0: The distribution
is normal
H1: The distribution
differs from normal
Frequency
Distribution/
Counts
Count the number
in each of several
categories on two
dimensions.
Is there a
relationship
between the
dimensions
(variables)?
Chi Square for
Contingency Tables H0: There is no
relationship (the
dimensions/
variables are
independent)
H1: There is a
relationship.
