# Comparing Survival over Time

The subscribers data has three years of complete starts. The analyses so far have mixed all the data together, calculating “average” hazards over the entire period. *Have the hazards have changed over time?*

This section presents three ways to approach this problem. The first looks at whether a particular hazard has changed over time. The second looks at customers by the year in which they started, answering the question: *What is the survival of customers who started in a given year?* The third takes snapshots of the hazards at the end of each year, answering the question: *What did the hazards look like at the end of each year?* All these ways of approaching this question use the same data. They simply require cleverness to calculate the hazards.

The next chapter presents yet another way to look at this problem. It answers the question: *What did the hazards look like based on the stops in each year?* Answering this question requires a different approach...