(1) It is safe to think that "there are differences in eating habits between southern students and northern students in choosing desserts"; (2)

Test analysis: (1) Substitute the data in the contingency table into the formula to get the value, and then compare it with the table. If it is less than, there is a 95% confidence that "there are differences in the eating habits of southern students and northern students in choosing desserts"; (2) Three students are randomly selected from five students, and there are 11 kinds of results, which constitute the basic event space. Among them, the event of "at most 1 people like desserts" contains seven basic events, which can be substituted into the probability calculation formula of classical probability. < P > (1) The data in the contingency table is substituted into the formula for calculation, so it is < P > safe to think that "southern students and northern students are choosing dessert diets." Among them, students who like desserts, students who don't like desserts,

are composed of 11 basic events, so it is equally possible to cut these basic events. If A is used to indicate the event that "at most 1 out of 3 people like desserts", event A is composed of 7 basic events. Therefore,