Aha! moments are satisfying in part because they feel so right; all the pieces of a puzzle appear to fall into place with little conscious effort. But can you trust such sudden solutions? Yes, according to new research published in Thinking & Reasoning. The results support the conventional wisdom that this type of insight can provide correct answers to challenging problems.

In four experiments, Carola Salvi, a postdoctoral researcher at Northwestern University, John Kounios, a psychologist at Drexel University, and their colleagues presented college students with mind teasers, such as anagrams and rebus puzzles. At the completion of a timed trial, subjects were asked to report if they had arrived at their answer by thinking the problem through step by step (analytical problem solving) or if the solution had sprung to mind (insight).

In all four experiments, aha! solutions were more often correct than those achieved deliberately. For instance, in one experiment, in which 38 participants had to think of a single word that could form a compound phrase with three previously presented words (such as “apple” for the trio “crab,” “pine” and “sauce”), aha! solutions were correct 94 percent of the time compared with 78 percent accuracy for analytical solutions.

This outcome may result from the way the brain generates insights. Because such processing occurs largely outside a person's awareness, it is all or nothing—a fully formed answer either comes to mind or it doesn't. This hypothesis is supported by EEG and functional MRI scans, which revealed in previous studies that just before insight takes place, the occipital cortex, which is responsible for visual processing, momentarily shuts down, or “blinks,” so that ideas can “bubble into consciousness,” Kounios says. As a result, insights are less likely to be incorrect. Analytical thinking, in contrast, happens consciously and is therefore more subject to rushing and lapses in reasoning.

That is not to say that insight is always the best strategy. The Salvi and Kounios experiments involved puzzles with clear right and wrong answers. So the results may not apply to real-world situations, where problems are typically highly complex and may require days—if not months or years—to solve.

In fact, difficult questions often necessitate several different strategies to arrive at a solution, says Janet Metcalfe, head of the Metacognition and Memory laboratory at Columbia University, who was not involved in the study. She adds: “There may not be a perfect solution to a problem.”