May 9, 2002
Vol. 21 No. 15

current issue
archive / search

    Researchers show people overestimate ability to effectively communicate ambiguous information

    By Josh Schonwlad
    News Office

    Most people seriously overestimate their ability to communicate effectively, even when dealing with information they know to be ambiguous, say Chicago psychologists Boaz Keysar and Anne Henly. In a recent study of 40 pairs of listeners and speakers, “When it comes to communication, people overestimate their skill,” said Keysar, Associate Professor in Psychology and the College.

    Keysar and Henly, a graduate student in Psychology, found that speakers believed that their intended meaning was being understood most of the time, but the findings showed that nearly half the time they thought they were understood they actually were not.

    “This really demonstrates how great the potential is for day-to-day miscommunication,” Keysar said of the study, which has been published in the May issue of the journal Psychological Science. Keysar believes the findings suggest a rule of thumb about communication. “Anticipate miscommunication,” said Keysar. “If you want to make sure you’re understood, assume that what you say is much less transparent than what you think it is.”

    The findings are based on an experiment in which speakers read to listeners 12 syntactically ambiguous sentences, such as “Angela shot the man with the gun.” The same speakers also read four lexically ambiguous sentences, such as “The typist tried to read the letter without her glasses.”

    The syntactically ambiguous sentences were problematic because listeners had to decide whether the speaker meant that Angela had the gun or that the man had the gun. With the lexical ambiguity, the listeners had to, for instance, decide whether the speaker was using the word “letter” to refer to a letter in the alphabet or a letter that is mailed.

    Study results showed that when speakers thought the listeners understood their intended meaning, they were misunderstood in nearly half of the cases (46 percent). In contrast, speakers underestimated their effectiveness––thinking they were not understood when they were understood––in only 12 percent of the cases.

    What is perhaps most surprising about these findings, Keysar said, is that the speakers were told in advance that the sentences were ambiguous. “Even then, even though they were warned,” Keysar said, “they still thought they managed to convey their intention with intonation.”

    Keysar believes this overestimation of effectiveness by speakers represents an “illusion of control,” a phenomenon that’s been demonstrated in other activities. For instance, Keysar said, when people buy a lottery ticket and select the numbers, they have a false sense that they have a better chance of winning than when the numbers are selected randomly. This also happens with speakers, Keysar said, because they are the actors, and “they have the illusion that they fully control the outcome.”

    Overestimation of effectiveness was more common with syntactically ambiguous sentences than with the lexically ambiguous sentences.

    “This suggests that speakers appreciated the impossibility of using intonation to resolve the meaning of an ambiguous word,” Henly said, “and thus, were less likely to overestimate their ability to convey the word’s intended meaning.”

    Listeners also were not very accurate in understanding the sentences, misunderstanding 39 percent of the time. They were more successful dealing with syntactic than lexical ambiguity. But what Keysar found interesting was their confidence in comprehension. “Even with very little information, listeners still were confident that they understood,” he said. “In fact, they were just as confident when they misunderstood as when they understood the speakers intended meaning.”

    In a second experiment, the researchers tested the ability of an observer to gauge the comprehension of listeners. These observers knew what the speaker intended and listened to the speaker’s message, but they were not involved in the social exchange. They listened to recordings of the listener-speaker pairs and indicated which meaning they thought the listeners understood.

    In contrast to the speakers, observers did not systematically overestimate the speakers’ effectiveness. Keysar believes these results demonstrate a general point about communication. “When we observe others talking,” he said, “we do not think that what they say is transparent.”

    Speakers, on the other hand, are far more likely to believe that what they say is transparent. Keysar believes this is because the mental effort of speaking makes it hard for people to recognize that what they say is difficult for others to understand.