1. Determine and justify if paired or independent data
First, you must determine whether your data are paired or independent. This will determine
which tests are appropriate for your data. Describe your data and provide a justification for why
your data are paired or independent.
2. Null and alternative hypotheses
State both your null and alternate hypotheses. Your alternate statistical hypothesis should be related
to the hypothesis derived from your research question.
3. Test for normality
This will help you determine whether you need a parametric or non-parametric test. We can
determine whether data are normally distributed in Excel by visually examining a frequency
distribution of the data. For instructions, see Hand-out 5.1: “Testing for normality in Excel”
4. Choose and justify statistical test
State whether you will use a paired or independent t-test, a Mann-Whitney U test, or a Wilcoxon
signed rank test, and why.
5. Perform statistical test
Using hand-outs 5.2 – 5.5, perform your statistical test in Excel
Present your data separated into two columns for the two
groups being compared. Provide us with your P-value from the Excel analysis, and the level of
significance used.
6. Interpret results
Based on your obtained P-value, state whether your data support your null or alternate
hypothesis. If you do find a statistical difference between the two groups, examine the means of
each group to determine which group had larger or smaller values.