The New York Times recently ran a story in which the reporter (a film-maker) took five everyday food items and ran them through a bomb calorimeter. Lo and behold, the calorie counts on the packages don't match the claims, and four out of five were too high, adding up to an extra 500 calories.
I have mixed feelings about this study. Sure, some of the calorie counts are a little higher than claimed -- but the error bars in the bomb calorimeter's findings were not reported. Why are we trusting the instrument so much? This is far more about journalism than it is about science.
And yet, and yet ... it could be about science. What if a chemistry course on calorimetry had each lab group buy different kinds of food and test its calories, then compare to the claimed calories? They'd learn bomb calorimetry but they would really learn whether calorie counts are accurate, which is more interesting because it's what the calorimetry is for. If it's done by students doing it for the first time, their numbers may not be as accurate, but I'm not sure they're so accurate in the first place. It would be like the lab exercises I've read about in which students use DNA barcoding to check whether the fish at the supermarket is indeed the species that the label claims it is. Instant relevant lab exercises -- and if the students like, they can inflate it into a Michael-Moore-style expose more approprite for a communcations class. Feel free.
Or a newspaper as venerable as the Times could talk about error bars and point out that not all differences are truly significant ... but that may be too much to ask. Replication and journalism may be philosophically at odds. So be it.