Advertising researchers claim that the power of curiosity can be harnessed to design an effective Internet advertising strategy that results in a better evaluation of the advertised product. They develop six advertising texts with varying amounts of "curiosity" triggers. College students are randomly assigned to one of the six versions of the advertisement text and their evaluation score of the advertised product is recorded.

Which of the following is the most appropriate statistical test to use to test if the six versions of the advertisement text differ in their evaluation score?

1. Matched pairs t-test
2. Two-sample t-test
3. ANOVA
4. Chi-squared test for independence
5. Inference for regression