Even though they weren’t instructed to restrain themselves from considering such a solution, they were unable to “see” the white space beyond the square’s boundaries.
Only 20 percent managed to break out of the illusory confinement and continue their lines in the white space surrounding the dots.
It's an interesting experiment, but the author's conclusion cannot possibly follow from the results of it.
your conclusion: that the second experiment disproves the theory that thinking outside the box is useful in solving problems, is itself a fallacy.
In other words, the “trick” was revealed in advance.
Would you like to guess the percentage of the participants in the second group who solved the puzzle correctly?
Most people assume that 60 percent to 90 percent of the group given the clue would solve the puzzle easily. What’s more, in statistical terms, this 5 percent improvement over the subjects of Guilford’s original study is insignificant.