We have been getting a great deal of e-mail asking about the various “experiments” on YouTube (inspired by events on Fast Food Nation) in which a McDonald’s hamburger and fries are left out for weeks or months. To the surprise of some, the food does not rot away to nothing but instead shrivels a little and becomes hard and shiny, but does not get moldy or rotten. The implication is that there is something wrong and unnatural about food that doesn’t rot when left out.

Unfortunately this is what passes for “science” on the interwebs. But it does provide a teaching moment – with lessons about scientific methodology and how the logic of interpreting evidence.

The first thing I need to point out is that none of these “experiments” are actually experiments. They are simply observations. Observation is an important component of experiments, but the raw act of observing is not sufficient to establish an experiment. In fact, in medicine we divide clinical evidence into “observational” and “experimental.”

In order to conduct a scientific experiment you have to control for some variables. There has to be a comparison between different groups so that variables can be isolated. Setting out a hamburger and seeing what happens is an observation – but there is no controlling for variables, so it is not an experiment.

I don’t think I am being pedantic. There is much that passes for science on television and the internet that is a gross distortion of science and teaches the public misinformation about what science actually is. Every ghost hunting show, for example, makes the same mistake as the hamburger videos – they make observations and then draw highly speculative, fanciful, and unwarranted conclusions from those observations. They think they are doing science because sometimes they are making those observations with fancy equipment. But I have never seen a ghost hunter actually do an experiment, and control for variables that might help determine what the nature of the phenomena they are observing is.

Let’s consider what variables might be interesting to control for with the “rotting hamburger” question. The most obvious one would be to set out McDonald’s hamburgers (more than one so that we can at least do some statistics) and compare them to a group of hamburgers that are home cooked with known ingredients vs hamburgers from another restaurant chain. But there are other variables also. How well are the hamburgers cooked (which relates to their moisture content at the beginning of the experiment)? There are many environmental conditions that should also be controlled for, if not varied to see their effect: humidity, temperature, ventilation, and light exposure.

An experiment looking at these variables for comparison would be an actual experiment and would tell us something – unlike the videos on YouTube.

This raises another point – the best experiments are designed to test specific hypotheses, and you should have some idea beforehand how the results will affect each hypothesis. For example, there are two primary hypotheses with the hamburger question: what determines the result of the left-out hamburger – is it primarily characteristics of the hamburger or characteristics of the environment, and which ones.

My hypothesis (we discussed this on the SGU this week, but to extend that discussion a bit) is that environmental factors, specifically temperature and humidity, are the most important. But also the beginning moisture of the hamburger and the thickness of the hamburger are likely important because these variables will determine how long it takes for the burger to dry out. Is that amount of time enough for mold to form? I note that McDonald’s hamburgers are thin and thoroughly cooked, and will therefore dry out quickly (especially in a dry environment) – too quickly for mold to form. Thoroughly cooked meat should also be free of bacteria to cause rotting. So in the end you will have a dried hard patty, but it will not become moldy nor will it rot.

I do not think there is anything inherent to the ingredients of the hamburger that will significantly affect whether or not it molds or rots – which is the exact implication of these YouTube videos. In order to conclude that it is the hamburger ingredients that are to blame, experiments that control for thickness, degree of cooking, and environment need to be done so that the property of the burger itself is isolated as a variable. And then you will need to spend some time scratching your head to see if you can think of any other variables you forgot, while your colleagues specializing in the natural history of left-out hamburgers do the same, until we can be reasonably sure that every plausible variable was accounted for (I say plausible, because there are in unlimited number of implausible variables, like the phase of the moon when the burger was cooked).

If this question were taken seriously, it is probably true that a complex picture will emerge. Perhaps environment is the key factor, along with thickness and juiciness of the burger. But more subtle variables may emerge, such as the leanness of the beef, or the amount of salt added. Genuine controversies may emerge over conflicting or ambiguous evidence, making for some lively scientific meetings.

But in the end I believe the evidence will speak and a consensus of scientific opinion will emerge. Sure, there will be deniers – perhaps some genuine scientific dissenters, but also those with an anti-fast food ideological agenda, for example. They may even form a community of deniers, creating their own websites, and doing their own dubious experiments that science bloggers will then have to spend time criticizing and pointing out all the flaws. The political parties will likely come down on opposite sides of the question, each accusing the other of “junk science.” The news media will always cover the story as a genuine scientific controversy, long after a solid consensus has emerged, and will interview the same few cranks on the denier side for “balance”.

OK – maybe I got a little carried away with this analogy, but you get the idea.