For a project in his statistics class, Mark conducted a survey in which he asked participants to assign status ratings to three photographs of cars (Corvette, Toyota RAV4, Honda Civic) on a scale of 1 to 10. The average ratings were: (a) Corvette 8.3; (b) RAV4 6.7; (c) Civic 3.9). On the basis of t-tests, he concluded that all of the differences among the three mean ratings were statistically significant. He was disappointed to learn that he received a low grade on the project due to the fact that he had used the wrong statistical test. What was wrong with Mark's approach to analyzing his data, and which test should he have used.