Summary notes for statistics for Masters of Public Health

Types of data and statistical tests:

Comparison of means – looks for the difference between the means of variables:


Statistical test: Data type: Use:
Pearson’s correlation Continuous Look for an association
between two variables
Spearman’s correlation Continuous Tests for associations
between two ordinal
variables (doesn’t
assume normal
distribution)
Chi-square Categorical or binary Tests for strength of
association between two
categorical or binary
variables
One sample t-test Independent var: Tests for differences
categorical or nominal between two related
(compare to fixed value) variables – does the
population mean differ
from the chosen fixed
value? Comparing
population mean with a
fixed value
Paired t-test Dependent var: Does the population
continuous mean differ between two
Independent var: time related groups?
variable H0 = mean
change/difference is 0.
H1 = mean
change/difference is not
1.
Independent t-test Dependent var: Whether population
continuous mean differs between
Independent var: two groups.
categorical or nominal H0 = population means
of group 1 and 2 are the
same.
H1 = population means
of group 1 and 2 are
different.


Regression – assess if change in one variable predicts change in another variable:

Name: Data type: Use:
Linear regression Dependent variable: Gives a numerical
continuous (outcome) explanation of how