In a study conducted in Pennsylvania, servers in various restaurants wrote "Thank you" on randomly selected bills before presenting the bills to their customers. Tips on these bills were an average of three percentage points higher than tips on bills without the message. Therefore, if servers in Pennsylvania regularly wrote "Thank you" on restaurant bills, their average income from tips would be significantly higher than it otherwise would have been.

Which of the following is an assumption on which the argument relies?

1. The "Thank you" messages would have the same impact on regular patrons of a restaurant as they would on occasional patrons of the same restaurant.
2. Regularly seeing "Thank you" written on their bills would not lead restaurant patrons to revert to their earlier tipping habits.
3. The written "Thank you" reminds restaurant patrons that tips constitute a significant part of the income of many food servers.
4. The rate at which people tip food servers in Pennsylvania does not vary with how expensive a restaurant is.
5. Virtually all patrons of the Pennsylvania restaurants in the study who were given a bill with "Thank you" written on it left a larger tip than they otherwise would have.

I am torn between 2 and 4 here.
Premise: "Thank you" on randomly selected bills - average of three percentage points higher
Conclusion: regularly wrote "Thank you" on restaurant bills - average income from tips would be significantly higher than it otherwise would have been.

4. The rate at which people tip food servers in Pennsylvania does not vary with how expensive a restaurant is. - Negation: If it did vary depending on how expensive a restaurant was then the conclusion falls apart.

2. Regularly seeing "Thank you" written on their bills would not lead restaurant patrons to revert to their earlier tipping habits. - Negation: If it did lead patrons to revert to earlier habits then argument again falls apart.

Please can you identify whats wrong in my approach ?

Negation of four does not cause the conclusion to fall apart.

Even if the rate at which people tip varies with the level of restaurant, the study showed that tips were 3% higher - percent is a relative value, regardless of the base. So whether they used to tip 10% or 15% before the "Thank You" experiment, they still tip 3% higher, on average, after.

The conclusion specifically ties the act of seeing "Thank You" on the bill to tipping more. Only choice 2 causes this link to break down - seeing "Thank You" would no longer cause customers to tip more.

Thanks for the clarity.
It really helped.
This Forum is very professional and one of the best available.
I only wish more people know and can benefit from this amazing place.