diet coke, mentos, and science fair projects

At a family gathering a while back, one of my cousins was trying to get some ideas for a high school science fair project. Of course, I thought back to my old experiment on edible mushroom cultivation and what kind of substrates they best grow on. I clearly remember losing the competition to another project that effectively answered the question, "Why do shellfish turn from blue to red when cooked?" The technical aspects of both experiments were comparable, but the shellfish question was just a lot more interesting than my mushroom question.

Three computer scientists (my siblings and I) brainstormed to find a flashy but viable idea for my cousin to think about and we came up with a question, "What causes the Diet Coke and Mentos reaction?" Yes, we were thinking of the youtube video.

I tested this idea at a get together with some friends. I'm not sure if it was the beer talking, but it was a hit. Within a minute of mentioning it, everyone was coming up with hypotheses, possible experiments, and just some plain old hilarious ideas.

The problem about high school science fair projects is how limiting they are. Essentially my cousin had to title his experiment:

The Effect of (Independent Variable) on (Dependent Variable)

Phrasing the science project in the question above almost implies that one has to know that there is a relationship between the independent variable and the dependent variable, otherwise the project will be a dud (i.e. the conclusion = there is no effect.) Well, it's never really a dud, because knowing that a relationship between two variables does not exist is just as important as when a relationship does exist. Unfortunately, no one puts any reward or glory into finding the non-relationships.

As another aside, I never liked those terms independent variable and dependent variable. Independent of what? Dependent on what? My preference is for the engineering or statistical terminology:

The Effect of (Predictor Variable) on (Response Variable)

Back to my main point though -- one and only one variable?! Boring! No wonder kids don't get excited about science. The world is multivariate, non-linear, non-continuous, and full of complex relationships. Students shouldn't be focusing on data collection and plotting an x-y line graph. Students should be focusing on problem solving and experimental design. That is, how does one go about answering the question at hand (or proving their hypothesis). Not that data collection and data analysis aren't important. I just think that they should be activities taught as tools to reach an end goal, not as the end goal themselves.

In the case of the Diet Coke and Mentos idea, the science project could have been to run experiments to isolate which two elements in the Diet Coke and Mentos react together to produce gas. More or less, running experiments to eliminate variables that don't matter and to confirm the ones that do matter.

In the Diet Coke and Mentos idea, think of what questions need to be answered:
- Can the reaction occur with another carbonated drink?
(Ginger ale is fizzier than diet coke)
- Can the reaction occur with another candy?
- Is the reaction physical or chemical?
- If the reaction is physical, do pebbles work instead?
- If the reaction is chemical, what are the compounds at work?

I think this would be fun for kids to work on and help them work on their design and logical thinking skills, kind of like a detective mystery, but fizzy. Besides, we all want to know the answer to the question.