Let's say that we have the following scenario:
A survey of local residents reveals that 60% of all residents get no regular exercise, 25% exercise sporadically, and 15% exercise regularly. After doing some fancy modeling and putting some community services in place, the survey was repeated with the same questions. The follow up survey was completed by 500 residents with the following results:
No Regular Exercise: 260
Sporadic Exercise: 135
Regular Exercise: 105
Total: 500
Now, we want to determine if there is evidence for a statistically significant shift in the responses of the residents. Our null and alternate hypotheses are as follows:
- H0: The deviations from the previously observed percentages are due to pure chance ...