The chief ingredients for this experiment are cans or 2-liter bottles of Pepsi and Coke -- enough to fill a plastic cup with a gulps-worth for each student taking part in the experiment. Add a generic brand of cola to make the results even more interesting. You'll also need extra plastic cups to fill with water to clean the palate between tastings, as well as several slips of paper to record each student's guess for each tasting.
Divide the students into three smaller groups: the tasters, who believe they can tell the difference between the sodas; the pourers, who will be the only ones to know which sodas are in each batch; and the recorders, who will mark down the tasters' guesses as they sample the different drinks. Have everyone leave the room except the pourers, who will flip a coin to determine which soda they'll use for the first round of samples -- a random selection that mirrors ordinary scientific inquiry.
The recorders should make a list of the tasters, with three columns next to each name. When the tasters comes back into the room, they should taste the first round of sodas, then silently write their guess on a piece of paper, which they pass to the recorder before leaving the room again. This will ensure that each guess is based on each person's opinion and untainted by their classmates. Have students continue through the sodas until finished, drinking some water between each soda.
Some people will be astute at identifying the differences between the sodas, especially those who drink them regularly. Others may miss the mark completely, which could cause them to reconsider their advertising-based opinions. You can arrange the results on a poster board to show how many students guessed correctly and at which points, or you can get more complicated by introducing a computer-based analysis program, such as a simulation program, which can arrange your data in graphs or pie charts.