This report in The Scientist presents a cautionary tale about the potential effect of the observer on the observed.
A team of scientists from several countries around the world tracked the activities of penguins that had identification bands on their flippers, and compared their behavior and breeding success with penguins that had no such bands. The stunning results--the banded birds arrived up to 3 weeks later at breeding sites and produced 50% fewer chicks. Given the nearly universal reliance on banding as a form of research identification, this study could have a dramatic effect on future wildlife studies. It also calls into question the conclusions drawn in past studies where behavorial or biological data depended on banding or similar identification practices.
"Ecologists are realizing that we can't take the effects of our manipulations for granted--we're studying living organisms in natural systems, not a beaker in a lab," said James Cahill of the University of Alberta, Edmonton. In 2001, Cahill found that by handling the plants in his experiments, he changed the amount that herbivores ate them.
The penguin study "has enormous implications for long-term studies, and for those focused studies where people want to know where every individual is," said Cahill, who was not involved in the study. "This is the type of news that most scientists don't want to hear."
In part, Heisenberg's Uncertainty Principle suggests that "mere" observation of an event can affect the outcome of that event. "Common sense" might object that observation does not require contact, so how can it have an effect? This, of course, excludes the scientific fact that all observation involves some kind of transmission--light, sound, the passing of an x-ray, the handling of a detector, etc. "Common sense" rarely operates scientifically.
(It occurs to me that "affecting the outcome through observation" may have as much to do with changing our perception as it does with "actually" affecting the "out-there" events. To observe something, we must classify, categorize, and reify certain aspects. Philosophically, you might say we "will" an "object" into existence by somewhat arbitrarily discerning an edge or distinguishing pieces of a process. I recall a story about some remote tribe that, when tested, lumped together as "the same color" a large variety of shades English speakers might call "rose, pink, orange, magenta, maroon, etc". As humans, they probably use the same physical mechanism for observing color as English speakers. Would more concerted study by a linguistic pioneer among them "create" more colors? Yes, and no....)
Fortunately, even though "most scientists don't want to hear" evidence about the direct effects of observation on the observed, the nature of scientific research gives some hope that observers will make an effort to adjust to it, rather than rejecting it. Real success in science depends on adapting theory to fact, not the other way round.