Statistical Analysis
What is it you want
to know?
Statistical test
Values to report
What does P mean?
Do we have enough
evidence from our
samples to infer a
difference between the
two populations?
Do we have enough
evidence from our
samples to infer a
difference between the
two populations?
t-test
t
df
P
*The data collected are
discrete, meaning
there are categories of
values (example: yes/no
(binary) responses)
Chi-squared
Chi-squared
df
P
P is the probability that
the difference between
our samples occurred
only due to chance.
P is the probability that
the difference between
our samples occurred
only due to chance.
If P < 0.05, then there is
convincing evidence
from the samples to
suggest that there is a
difference between the
two populations.
If P < 0.05, then there is
convincing evidence
from the samples to
suggest that there is a
difference between the
two populations.
If P > or = 0.05, then
there is no convincing
evidence, based on the
data collected in our
samples, to conclude
that there is a difference
between the two
populations. (It is too
likely the difference
observed between the
samples was due only to
chance)
If P > or = 0.05, then
there is no convincing
evidence, based on the
data collected in our
samples, to conclude
that there is a difference
between the two
populations. (It is too
likely the difference
observed between the
samples was due only to
chance)
*The data collected are
continuous, meaning
there is a range of
possible values
Is there enough evidence
from our sample to
infer that there is a
relationship between
two continuous variables
in the population?
Linear regression
R-squared
df
P
P is the probability that
the relationship
observed in the sample
was due only to chance.
If P < 0.05, then there is
convincing evidence
from the sample to
suggest that there is a
relationship between the
two variables in the
population.
If P > or = 0.05, then
there is no convincing
evidence based on our
sample to conclude
that there is a
relationship between the
two variables in the
population. (It is too
likely the relationship
between the variables
observed in the sample
was due only to chance.)