The Man Who Lied to His Laptop - What Machines Teach Us about Human Relationships

Clifford Nass and Corina Yen

Created on Thursday, June 27, 2013.
Filed under
 

An engrossing overview of how computers have been used as “the perfect research confederates”, and what these human-machine studies have revealed about human emotions and biases.

 

I searched through the social science literature to find simple tactics that unpopular people use to make friends. The most powerful strategy I found was to create a scapegoat. I therefore designed a new version of Clippy. After Clippy made a suggestion or answered a question, he would ask, Was that helpful? and then present buttons for yes and no. If the user clicked no, Clippy would say, That gets me really angry! Lets tell Microsoft how bad their help system is. He would then pop up an e-mail to be sent to Manager, Microsoft Support, with the subject, Your help system needs work! After giving the user a couple of minutes to type a complaint, Clippy would say, C'mon! You can be tougher than that. Let em have it! We showed this system to twenty-five computer users, and the results were unanimous: people fell in love with the new Clippy!

 

Although every aspect of the interaction was identical except for the voice, participants who heard the female voice reported that the computer taught love and relationships more effectively, while participants with the male-voiced computer reported that it more effectively taught technical subjects. Male and female participants alike stereotyped the gendered computers. When we asked participants afterward whether the apparent gender of the voice made a difference, they uniformly said that it would be ludicrous to assign a gender to a computer. Furthermore, every participant denied harboring any gender stereotypes at all!

 

As established in the classic paper on social facilitation by Robert Zajonc and much subsequent research, the effect of other students depends on how confident the student is. When you feel confident, having other people present improves how well you learn and perform. However, when you feel insecure, having other people around makes you nervous and pressured so you don’t learn as well. As a result, we decided to have the teaching environment be a virtual classroom but with a variable number of students. When users were doing well on the practice tests, more students would appear at the desks, but when their practice test scores were low, there would be fewer students and more empty desks.

 

Participants reported that they liked the flatterer computer (which gave random and generic feedback) as much as they liked the accurate computer. Why did people like the flatterer even though it was a brown-noser? Because participants happily accepted the flatterers praise: the questionnaires showed that positive feedback boosted users perceptions of their own performance regardless of whether the feedback was (seemingly) sincere or random. Participants even considered the flatterer computer as smart as the accurate computer, even though we told them that the former didn’t have any evaluation algorithms at all!

[…]

The results of this study suggest the following social rule: don’t hesitate to praise, even if you’re not sure the praise is accurate. Receivers of the praise will feel great and you will seem thoughtful and intelligent for noticing their marvellous qualities — whether they exist or not.

 

When the thalamus cannot identify the valence of the evaluation — for example, if someone speaks with a straight face — it sends the information to the more sophisticated parts of the brain before signalling your body to react. Those higher-thinking processes interpret what is being said in order to make the positive-versus-negative judgement (e.g., recognizing praise versus criticism); they then send this information back to the thalamus, which subsequently guides your reactions. Thus, if your manager walks up to your desk and unsmilingly tells you in a monotone voice that you are getting a very large bonus, you will react with happiness more slowly than if your manager had bounded up with a big grin and shared the news in an enthused voice.

 

Linguists Robert Schrauf and Julia Sanchez have shown that only 20 percent of typical words in English or Spanish have a completely neutral connotation. Another 50 percent of words have negative orientations, and the remaining 30 percent have positive orientations.

 

One fascinating side effect of the power of negativity is that you remember less of what is said before receiving criticism because negative remarks demand so much cognitive power that the brain cannot move the prior information into long-term memory. Known as retroactive interference, this explains why it is often difficult to give a detailed answer when asked, What made that person yell at you? People frequently cannot remember what they were doing just before their computer started behaving strangely — a common problem for technical support professionals attempting to troubleshoot. (Because praise and positive events do not require significant cognitive resources, they do not cause retroactive interference.) Immediately after a negative evaluation, however, the brain and body go into full alert. People have a number of consequential choices after receiving a negative remark: walk away, defend, argue, escalate, physically threaten, or plead for a solution. We immediately seek information that will help guide us in our decision. So, after a negative event, our memory is actually improved, an effect known as proactive enhancement. This is why you should present information you want remembered immediately after a negative remark.

 

When you want to give a mix of positive and negative feedback, the order is critical. Tradition states that one should give praise first to soften up the person before giving her or him bad news. However, this is a poor idea: although the immediate reaction to the negative remark will be softened, in a short time retroactive interference will come into play and all that will be remembered is the negative remark. It is better to present the negative feedback first and then the positive evaluation. The criticism will bring people to attention in time to listen to your praise.

 

It is also important to consider that receiving an equal number of positive and negative remarks feels negative overall because of hedonic asymmetry and the self-serving bias. It is far better to briefly present a few negative remarks and then provide a long list of positive remarks. This can take significant effort — its much easier to remember negative impressions — but generating lists of positive remarks is time well spent. You should also provide as much detail as possible within the positive comments, even more than feels natural, because positive feedback is less memorable.

 

Real constructive criticism is not simply a valid assessment of a persons work — no one initially perceives even the most accurate criticism as constructive — it must also guide the recipient on how to act on the evaluation.

 

A second approach for enhancing positive evaluations is surprise because it gets people to pay attention and think harder about what you just said. For example, if you compliment someone on something that he or she thinks you are unaware of, it will have a bigger effect than if you keep dishing out the same obvious compliments. Slipping these surprising references into a list of more obvious positive remarks is even more effective.

 

Previously, I suggested that praise never hurts. This does not mean, however, that all types of praise are beneficial. Telling people that they are destined to succeed before they attempt a new activity can make any failures crushing. Thus, fixed-mindset praise, meant to make people feel better, can actually make people feel much worse about their work and more negative about the person who praised them if it turns out to be inaccurate.

 

In sum, give growth-minded feedback to motivate people to choose challenging tasks and to confront their mistakes. Praise for taking initiative, completing a difficult task, learning new skills, and acting on criticism all encourage a growth mindset. Managers especially play a key role in creating an environment that encourages a growth mindset by giving feedback and support that praises learning and perseverance rather than inborn talent.

 

People also thought that the computer that criticized was more intelligent than the computer that praised, even though the two versions of the evaluators comments were equally complex in terms of grammatical structure and semantics. This is consistent with a classic study by the great social psychologist Solomon Asch, where he found that a person described as intelligent and polite was viewed as wise only 30 percent of the time, but a person described as intelligent and blunt was viewed as wise 50 percent of the time. In sum, you view someone who criticizes others more negatively than someone who praises, but you also view that person as more intelligent.

 

The foregoing suggests that if you evaluate a person, it will change others perceptions of that person, even when they are qualified to judge her or him for themselves. For example, if you praise your friends performance, people will think that your friend did well, even though they know that your evaluation might be biased. Conversely, you can undermine virtually anyones success by highlighting even a single deficiency. And saying, It’s just my opinion doesn’t obviate this effect. If you truly want people to make judgments for themselves (and to not judge you), keep your opinions to yourself.

 

The presumptive intelligence of negative evaluators also occurs in the case of movie critics: critics who dislike most movies are seen as much smarter than critics who like most movies. Surprisingly, people link criticism and intelligence instinctively. Amabile showed that when asked to present in front of an audience that is described as having a higher intellectual status than the presenter, presenters became more negative. In sum, critics, and all other evaluators, must decide whether they want to seem clever and contemptible or kind and clueless. Thus, criticize only when it is urgent to do so or when you’re trying to look smart.

 

How do people feel about those who praise themselves versus those who praise others? In this case, no trade-off exists. Participants liked the computer that praised itself much less than the computer that praised a different computer. Participants also felt that the tutoring computer that praised itself was less competent than the tutoring computer that was praised by another, evaluator computer. This makes the choice between self-praise and other-praise clear: never praise yourself when you can have someone else do it for you. Thus, the best strategy in the workplace is a mutual-admiration society with another colleague: person A praises person B, and person B praises person A. This will lead to both people seeming smarter and more likeable than if they praised themselves.

 

  • Praise others (but not yourself ) freely, frequently, and at any time, regardless of accuracy. Emphasize effort over innate abilities. When possible, establish a mutual-praise agreement in which you and a partner praise each other.
  • Criticize others with caution, keeping it brief and specific, and always with clear follow-up actions. Present ways to improve and resolve the criticism, and emphasize the importance of effort for success. Afterward, give people time to process and to respond when they are ready.
  • When mixing praise and criticism, offer broad praise, brief criticism focused on specific steps toward improvement, and then lengthy and detailed positive remarks.
  • Modesty might win you friends but will also be believed, so only criticize yourself when it is accurate and constructive to do so.
  • If you want to seem competent, then reverse the previous advice: praise yourself, criticize others, and don’t criticize yourself.

 

Personality fundamentally affects how you see the world, and so people with different personalities see it very differently. This makes it difficult for opposite personalities to understand each others point of view. On the other hand, when you encounter people who think and feel just like you, you naturally understand and feel affinity with them and what they say.

 

I then turned to the question of whether, if you have a dissimilar personality to a working partner, you should stay true to your own personality or change to be more similar. Does change indicate charming flexibility or despicable ingratiation? The effects were enormous and the conclusion inescapable: people love it when you adopt a similar personality to their own. The computer when it changed its personality to match the participants seemed much more intelligent, helpful, useful, and insightful than when its personality remained mismatched, and participants enjoyed working with it

 

Could being different and then becoming similar to a person be better than consistently being similar? When I analyzed the data, it confirmed my supposition: people felt that the computer that changed to be similar was more intelligent and helpful than the consistently similar computer. People actually found the ingratiator charming: they liked the computer more and enjoyed the interaction more than those who worked with the computer that was similar from the start.

 

Becoming similar generates even more positive feelings than consistent similarity-attraction because people perceive it as unspoken praise. That is, when people change to match you, they are implicitly saying, I realize that you do things the right way and I’ve been doing things the wrong way. I’m so sure that you are right that I am going to change to be like you. In a world of fragile egos, what could feel better than having people adapt themselves to become more like you? Imitation truly is the sincerest form of flattery.

 

Although people had extraordinary amounts of exposure to the faces of the candidates through media coverage, only 3 percent of participants noticed anything unusual about the morphed picture. Even these particularly observant participants thought that the photos were touched up; not a single person realized that her or his own face was blended with the candidates! The effects of the face morphing were remarkable and arguably terrifying for the political process (but a great example of the power of similarity-attraction!). Those people who were morphed with John Kerry supported Kerry 47 percent to 41 percent (12 percent undecided), while participants who were morphed with George Bush supported Bush 53 percent to 38 percent (9 percent undecided). In any race as close as the 2004 election, the use of this type of software in targeted ads to individuals (readily possible through Web sites such as Facebook) could sway an election. A follow-up study by Bailenson and colleagues with unknown candidates found even stronger effects of face morphing.

 

The obsession with similarity applies to even the most trivial matches. For example, the worlds leading authority on persuasion, social psychologist Robert Cialdini, demonstrated that when people find a lost wallet and the owners name is similar to their own, they are more likely to return the wallet to lost-and-found than when the owners name is dissimilar. Researchers have also shown that people are more likely to marry when their first names are similar.

 

social scientists have found ways to reliably create teams out of randomly selected people in a very short time. What are the secret ingredients in the social scientists special sauce? Identification and interdependence.

[…]

While the teams were arbitrarily formed minutes prior to the exercise and designated with only a badge and a sign, these symbols strongly influenced participants judgments. The message is clear: create a shared identifier and you have the makings of a more committed team.

[…]

Along with identification, the other key factor for creating a sense of team is establishing interdependence. To create interdependence, research suggests that team members must share two beliefs. First, they must feel that achieving the groups goals will also serve their own personal goals. Second, team members must believe that their efforts and the efforts of other team members are integral to the success of the team.

[…]

Whether financial, strategic, or emotional, as long as individuals hide their differing motivations (so that they do not threaten identification) and all team members explicitly support the shared goal, interdependence will strengthen the group.

 

Markers of identification sustain teams best when they increase in quantity and exclusivity. Actively seek and create inside jokes, neologisms, and other opportunities to mark your special bonds, and jealously restrict these shared phrases to your own team. A marvelous example of the effects of diluting linguistic shared identity comes from the ingenious strategy of Stetson Kennedy to subvert the Ku Klux Klan in the 1940s (described in Steven Levitt and Stephen Dubner’s book Freakonomics). After infiltrating the group, Kennedy discovered that the Klan had a special set of words and phrases that marked membership. For example, asking for Mr. Ayak was code for Are you a Klansman? their bible was the Kloran, and the local chapter was called a Klavern. Understanding how the Klansmen felt bonded by their own secret code, Kennedy had journalist Drew Pearson disseminate, via radio, these markers of identification to millions of people throughout the country. Once everyone had these symbols of team membership, they no longer were markers of membership. The effects on the KKK were devastating: meeting attendance and membership applications declined dramatically.

 

In the workplace, you probably can’t make team members go through hazing or an initiation process, but whenever possible, remind them that they were selected rather than assigned and that their accomplishments warrant their inclusion in the group. The more you can say about the rigors of selection and of the number of people who want to be in the group but cannot be, the better.

 

You might think that a desire for uniformity would lead teams to make moderate, if not optimal, decisions. Research initiated by MIT masters student James Stoner in 1961 and later confirmed in hundreds of studies reveals the opposite: teams committed to unanimity tend to polarize rather than moderate their decisions. In most situations, groups become overly optimistic, leading to a risky shift. That is, groups tend to select an option with a large upside, even if the downside is larger and more probable, because they focus on the positive.

 

Fortunately, our research and the social science literature demonstrate that emotions are much simpler than they appear. Specifically, although emotion is intimately tied to everything from the most primitive parts of your brain to your highest-thinking processes, almost everything that you need to know emerges from two fundamental concepts: valence and arousal.

 

people in all cultures face every situation by asking themselves, both consciously and unconsciously, two questions: 1. How well am I meeting my goals? 2. Should I do something about my goals? Emotions help us frame these questions in terms of feelings: 1. How happy am I? 2. How excited am I? The first question, called the valence question, reveals whether you believe that you are meeting your goals — thereby feeling happy — or whether you are failing to meet your goals — thereby feeling sad. The second question, called the arousal question, is like a volume knob on your responses: are you vigorously trying to meet your goals — excitement — or are you letting the situation play itself out — calm? Looked at another way, valence is the judging side of people: where am I in terms of where I want to be? Arousal, conversely, is the doing side: am I ready to act? Thus, emotion is an extremely concise and efficient way to link peoples goals, their current situation, and their attitudes and behaviors.

 

I like to think of a woman’s typical response to the classic wedding proposal as a beautiful example of how the primitive brain and cognitive system work together. You might think that there is nothing more delightful and romantic than having someone go down on one knee and say, Will you marry me? My training as a social scientist makes me see it differently (although I did propose this way myself ). If you watch a woman’s initial reaction to seeing a man start to get down on one knee, you will see extreme negative arousal: a tense body and a face white with fear. This response is due to the woman’s primitive brain saying, He is moving toward the ground. He’s either falling, hiding, or getting ready to attack me! I must prepare for something very bad! So instead of immediately showing joy or surprise, she shows concern about a body moving unusually! It is only after a moment that the higher-level parts of her brain deduce the symbolic meeting of the gesture, and, assuming that she plans to accept the proposal, her body and face transition to the appropriate joyous state: still very aroused but also very happy. While this transition can happen in less than a second, if the man sees the woman’s initial response, he too will exhibit an immediate negative reaction, until the woman’s adjusted response snaps him back to his own happy and excited state.

 

My son, Matthew, experienced the link between physiology and subsequent emotion firsthand when we went to see the movie Batman right after he had used an adrenaline inhaler for his asthma. Even though he normally handles tense movies with aplomb, this time he ran out of the theater and had nightmares for months — his brain mistook medicinal adrenaline for his reaction to the movie! This confusion of bodily feelings with subsequent emotional feelings occurs with calmness as well: after a massage, which triggers the parasympathetic nervous system, even a rocket launch is responded to calmly.

 

In sum, people do not combine emotions when they observe opposing signals; instead, they feel confusion and dislike. Failure to recognize this leads to a number of damaging practices. I have seen managers attempt to make criticism seem less negative by giving it with a smile. Instead of softening the blow, this leaves the evaluated person feeling uncomfortable and confused about what the manager wants changed. Similarly, I have seen managers present exciting news, such as a record-setting quarter, in a calm voice so that people will remain focused for the remainder of the meeting. Unfortunately, the mismatch actually creates a distraction: What is he trying to hide? What is the catch? Is this actually a bad sign? While moderating your emotions can be effective and appropriate (see the discussion below on emotion regulation), you cannot balance levels of valence or arousal by mixing and matching signals.

 

Valence is so contagious that you don’t even have to encounter it directly for it to affect you, according to research done by political scientist James Fowler and internist and social scientist Nicholas Christakis. They studied a social network of people from Framing-ham, Massachusetts, (connected via friendship, family, spousal, neighbor, and coworker ties), and found that one persons happiness can actually increase other peoples happiness out to three degrees of separation. In other words, your happiness is not only related to the happiness of your own friends but that of your friends friends friends. Fowler and Christakis even quantified the effect: a happy friend increases your probability of being happy by about 9 percent, and each additional happy friend adds to this effect. In other words, valence has a collective aspect.

 

When in doubt about how to describe someones valence, err on the side of happy. Even if the person is sad, the pronouncement is effective, albeit inaccurate. This mirrors the flattery research described in the Introduction: people will like you more if you are inaccurate for the sake of positivity. Never tell people they seem sad: they will not feel any better and will feel more negative toward you.

That's all there is, there isn't any more.
© Desi Quintans, 2002 – 2018.