Are some animals smarter than others? It’s hard to say, because you can’t sit a chimpanzee or a mouse down at a table for an IQ test. But a new study, in which scientists tested wild robins on a variety of skills, concludes that they do differ in the kind of “general intelligence” that IQ tests are supposed to measure.
General intelligence is usually defined as the ability to do well on multiple cognitive tasks, from math skills to problem solving. For years, researchers have questioned whether measurable differences exist in humans and nonhumans alike. In humans, factors like education and socioeconomic status can affect performance.
When it comes to animals, the problem is compounded for two main reasons: First, it is very difficult to design and administer tests that pick up on overall smarts instead of specific skills, such as the keen memories of food-hoarding birds or the fine motor skills of chimpanzees that make tools for finding insects in trees. Second, differences in animal test scores can depend on how motivated they are to perform. Because most experiments award would-be test-takers with food, an empty (or a full) stomach might be all it takes to skew the results. Thus, even studies that suggest variations in intelligence among mice, birds, and apes all carry the caveat that alternative explanations could be at play.
To get around some of these limitations, a team led by Rachael Shaw, an animal behavior researcher at Victoria University of Wellington, turned to a population of New Zealand North Island robins for a new round of experiments. The robins live at the Zealandia wildlife sanctuary, a 225-hectare nature paradise in Wellington where more than 700 of the birds live wild and protected from predators in the middle of the city.
Shaw and her colleagues briefly captured and banded 20 adult birds, which they trained to carry out six different tests to measure various skills thought to be linked to specific types of animal cognition (defined as the ability to acquire, process, and act on information about the environment). In a test of motor skills, for example, the birds learned to flip plastic lids covering wells in a wooden board, one of which had a tasty mealworm at the bottom. In another—designed to test their ability to distinguish colors—the birds had to figure out whether the worm was hiding under a red or a blue lid. To show they could distinguish symbols, they had to discern whether the worm was under a lid marked with a cross or a square. And because these robins store food for the winter, the scientists also measured their spatial memories. They saw how well the birds could remember in which of eight wells the researchers had hidden a worm.
To be sure that the robins were motivated to cooperate in the experiment, the researchers trained the birds to hop onto a scale and eat a single worm both before and after each test. This demonstrated that they were hungry enough to do their best on the exam.
In a paper published online this week in Animal Behaviour, Shaw and her colleagues report wide variations in how well the birds did on multiple trials of each test. For example, the birds required between 10 and 18 trials to successfully complete the motor skills test, between 12 and 80 trials to figure out the color discrimination challenge, and between 13 and 86 trials to get the symbol test right. On the spatial memory test, the birds flipped the wrong lids between zero and 13 times in their attempts to find the worm 24 to 48 hours after they had seen where the experimenter put it.
The researchers then turned to computers to analyze how the birds did overall. Their program found a weak but positive correlation between how well any particular bird did on one test and its performance on the other five tests. For example, if a robin did better on the motor skills or the color discrimination test, it was more likely to do better on the spatial memory test as well. To tease out whether these correlations were due to a small sample size or chance, the researchers used a more sophisticated statistical technique called principal components analysis, testing the actual results against 10,000 randomly generated simulations. They concluded that their results were statistically significant.
Among humans, researchers have found that general intelligence can account for about 40% of the individual differences in IQ-type test scores. In other words, a battery of different, well-designed tests appears to be measuring overall intelligence rather than simply how well a person does on each individual test. Shaw’s team found that between 34% and 45% of the differences in the robins’ overall performances could be explained by a similar general intelligence factor, suggesting that the birds’ performance is not just due to their talent on individual tests, but to their overall level of smarts.
Shaw says the results suggest that general intelligence “has evolved independently in the avian lineage” and that parallel evolution of general intelligence among animals “could be more widespread than we currently realize.”
The paper is “a big step forward” because running such tests in the wild “is exceptionally challenging,” says Alex Thornton, an animal cognition researcher at the University of Exeter in the United Kingdom. Nevertheless, he cautions that some of the tests could be measuring overlapping cognitive skills rather than separate ones. If so, he says, “the extent to which the study really captures something analogous to general intelligence in humans is somewhat questionable.”