I'm reading a very informative and fun book about human social psychology, Experiments With People (2nd ed, 2018).

... 28 social psychological experiments that have significantly advanced our understanding of human social thinking and behavior. Each chapter focuses on the details and implications of a single study, while citing related research and real-life examples along the way.

Here I summarize each chapter so that you can save time. Some results are old news to me, but some were quite surprising. I often skip over the experimental details, such as how the psychologists used ingenious tricks to make sure the participants don't guess the true purposes of the experiments. Refer to originals for details.

The experiments start in the 1950s and get up to 2010s, and occasionally literatures from before 1900s are quoted.

Chapters I find especially interesting are:

  • Chap 14. It lists the many failures of introspection, and raises question as to what consciousness can do.
  • Chap 16. It has significant similarity with superrationality and acausal trade.
  • Chap 20. It warns about how credulous humans are.
  • Chap 27. It is about the human fear of death and the psychological defenses against it.
  • Chap 28. It shows how belief in free will can be motivated by a desire to punish immoral behaviors. Understanding why people believe in free will is necessary for a theory of what is the use of the belief in free will.

Chap 1. Conforming to group norms

Asch conformity experiment, from Opinions and Social Pressure (Asch, 1955)

Video demonstration.

Groups of eight participated in a simple "perceptual" task. In reality, all but one of the participants were actors, and the true focus of the study was about how the remaining participant would react.

Each student viewed a card with a line on it, followed by another with three lines labeled A, B, and C (see accompanying figure). One of these lines was the same as that on the first card, and the other two lines were clearly longer or shorter. Each participant was then asked to say aloud which line matched the length of that on the first card... The actors would always unanimously nominate one comparator, but on certain trials they would give the correct response and on others, an incorrect response. The group was seated such that the real participant always responded last.

It was found that

  • When there are over 3 actors giving unanimously the wrong answer, the participant went along 1/3 of time.
  • Increasing the number of actors above 3 did not increase compliance.
  • Even when the difference between the lines was 7 inches, there were still some who complied.
  • If there is at least one actor disagreeing with the majority, the participant decreased compliance.
  • If the fellow dissenter joins the majority, the participant increased compliance to the same level of 1/3.
  • If the fellow dissenter leaves, the participant increased compliance only slightly.

There are two reasons for this compliance. One is heuristic about knowledge: the majority is usually more correct. Another is normative: social acceptance matters more than being correct.

The effect of a dissenting minority is notable.

Research finds that, whereas majorities inspire heuristic judgments and often compliance, minorities provoke a more systematic consideration of arguments, and possibly, an internal acceptance of their position.(Nemeth, 1987) Majorities tend to have a greater impact on public conformity, whereas minorities tend to have more effect on private conformity. (Chaiten & Stangor, 1987)

Chap 2. Forced compliance theory and cognitive dissonance

In When Prophecy Fails, the story of a UFO cult was detailed. When the doomsday prophecy failed, most people left, but some became even firmer believers.

(My own example, not appearing in the book.) In Borges's story A Problem, Borges asks, how would Don Quixote react if he kill a man?

Having killed the man, don Quixote cannot allow himself to think that the terrible act is the work of a delirium; the reality of the effect makes him assume a like reality of cause, and don Quixote never emerges from his madness.

This chapter reviews of Cognitive consequences of forced compliance (Festinger & Carlsmith, 1959)

  1. Participants were asked to do an extremely boring task.
  2. Then the experimenter asked the participant to deceive the next participant that the experiment was fun. Half were paid $1, another half paid $20.
  3. A control group was not asked to lie.
  4. Then they were nudged to take a survey about how they felt about the experiment.

The result is that, those paid $1 thought the experiment was fun, and those paid $20 thought it was boring, and those that didn't get asked to lie thought it was very boring.

Festinger explains this by the theory of cognitive dissonance:

  1. An attitude (thinking the experiment was boring) and a behavior (saying it was fun) clashes, creating an uncomfortable feeling.
  2. The participant then is motivated to remove the discomfort by changing the attitude by rationalization (thinking the experiment was actually fun).
  3. If the participant was paid $20, then there was no dissonance, as there was a ready explanation of the dissonant behavior.
  4. If the participant was paid $1, then there was dissonance, because the participant regarded the lying behavior as mostly voluntary.

An alternative explanation from Self-perception: An alternative interpretation of cognitive dissonance phenomena (Bem, 1967):

  1. We don't form beliefs about ourselves by direct introspection, instead, we infer it through
  2. When we behave against previously self-beliefs, this creates an update on our self-beliefs.

See also Chap 14 for more on the lack of introspection.

Current consensus is that both theories are correct, in different situations. The self-perception effect happens when the behavior is mildly different from self-beliefs, and the cognitive dissonance effect happens when the behavior is grossly different.

There are also many complications, such as in Double forced compliance and cognitive dissonance theory (Girandola, 1997), which reported that even if participants performed a boring task, then told others about how boring it was, afterwards they still felt the task was more interesting afterwards.

There is a lot of ongoing research.

Chap 3. Suffering can create liking

Such curious phenomena as hazing has been studied since The effect of severity of initiation on liking for a group (Aronson, 1959)

An experiment was conducted to test the hypothesis that persons who undergo an unpleasant initiation to become members of a group increase their liking for the group; that is, they find the group more attractive than do persons who become members without going through a severe initiation.

The group was a made-up thing by the experimenters. It purports to discuss interesting sexual things, but the participants, after finally "joining", would only hear a very boring group discussion about animal sex.

This hypothesis was derived from Festinger's theory of cognitive dissonance." 3 conditions were employed: reading of "embarrassing material" before a group, mildly embarrassing material to be read, no reading. The results clearly verified the hypothesis.

The "embarrassing material" are lists of obscene words. The "mildly embarrassing material" are lists of mildly sexual words.

Result: the very embarrassing ritual increased liking for the group.

Explanation was by the theory of cognitive dissonance: "I have already invested so much to join the group. I must be a fool if the group turned out to be bad! And I'm not a fool."

Cognitive dissonance has been used for brainwashing, persuasion, education, and many other kinds of things.

One of the authors learned from an investigative journalist about how a dodgy car company... had customers unnecessarily wait or hours while their finance deal was supposedly being negotiated upstairs.

[Commitment and community: Communes and utopias in sociological perspective (Kanter, 1972)] noted that

19th-century utopian cults requiring their member to make significant sacrifices were more successful. For example, cults that had their members surrender all their personal belongings lasted much longer than those that did not.

Some bad investments are continued far after they had become clearly unprofitable, this is the sunk cost fallacy.

Chap 4. Just following orders

The banality of evil is the theory that everyday people can do great evils such as the Holocaust, by simply following orders.

Behavioral study of obedience (Milgram, 1965) reported the famous Milgram experiment. A video recreation is here.

This is a very famous experiment with many followups. There is sufficient material freely online, such as the Wikipedia page. So I won't recount it here.

I was most surprised to learn that personality had very little effect. That is, obedience exhibited by the participants in this experiment was mostly situational, instead of stemming from the personality of the participants.

Chap 5. Bystander apathy effect

The murder of Kitty Genovese stimulated research into the "bystander effect". On March 13, 1964 Genovese was murdered... 38 witnesses watched the stabbings but did not intervene or even call the police until after the attacker fled and Genovese had died...

In Bystander intervention in emergencies: Diffusion of responsibility (Latané & Darley, 1968) attributed the lack of help by witnesses to diffusion of responsibility: because each saw others witnessing the same event, they assumed that the others would take responsibility.

This phenomenon has a big literature, and is very popularly known, possibly due to the dramatic stories.

Concerning the original experiment by Latane and Darley, I was again surprised that personality factors had little effect, except one: growing up in a big community is correlated with a lower probability of helping.

Chap 6. The effect of an audience

When people perform a task in the presence of others, they perform better if the task is easy, and worse if the task is hard. One theory is that presence of others increases physiological arousal, which then enhances performance of simple tasks and decreases performance of hard tasks. Other theories

In Social enhancement and impairment of performance in the cockroach (Zajonc, 1969), it is found that this is true even for cockroaches. In the experiment, Zajonc gave cockroaches two possible tasks: going through a straight maze, or a more complex maze. They either did the task alone, or while being watched by others outside (the maze was transparent).

While being watched, cockroaches solved faster on the straight maze but slower on the complex maze. This demonstrates that the physiological arousal theory is correct in cockroaches: the effect of an audience can happen without any complex cognitive ability.

However, complex cognitive ability sometimes does occur in humans. As reported in Social facilitation of dominant responses by the presence of an audience and the mere presence of others (Cottrell et al, 1968), blindfolded audience does not exert an effect on the performer.

Chap 7. Group conflicts from trivial groups

This chapter begins with the Robbers Cave experiment, which was a study that investigates the realistic conflict theory, which sounds very common-sense:

  1. group conflicts and feelings of resentment for other groups arise from conflicting goals and competition over limited resources
  2. length and severity of the conflict is based upon the perceived value and shortage of the given resource
  3. positive relations can only be restored with goals that require cooperation between groups

Then it recounts the blue eyes-brown eyes experiment. The problem, then, is, what is the least amount of group-difference in order to make a difference? Enter the minimal group paradigm of Experiments in intergroup discrimination (Tajfel, 1970). Participants first took a test on estimating dot numbers, then divided into "overestimators" and "underestimators", while in truth they were random. Then, they were given points (convertible to cash) to divide among the groups. Participants favored their own groups significantly more.

In fact, the most favored strategy was to maximize (own group) - (other group), even though it did not maximize (own group). Thus, even the most minimal social groups induced ingroup-outgroup conflict.

The minimal group paradigm has been studied in many ways. It was also found that outgroup homogeneity effect, that is, "they are all the same; we are diverse", could also arise from minimal groups.

One theoretical explanation is Tajfel and Turner's social identity theory (Tajfel & Turner, 1979), which states that: 0. A person's self-esteem depends on having a good identity.

  1. A person's identity has two parts: personal and social.
  2. Personal identity are about one's own traits and outcomes.
  3. Social identity are derived from social groups and comparison between groups.
  4. A person is motivated to improve self-esteem, and thus social identity.
  5. Thus, one is motivated to improve the standings of one's ingroups and decrease the standings of one's outgroups.

One supporting evidence is that when a person has more self-esteem, they are less discriminating against outgroups (Crocker et al, 1987).

Chap 8. The Good Samaritan Experiment

In the parable of the Good Samaritan,

a traveller is stripped of clothing, beaten, and left half dead alongside the road. First a priest and then a Levite comes by, but both avoid the man. Finally, a Samaritan happens upon the traveller. Samaritans and Jews despised each other, but the Samaritan helps the injured man.

This inspired an experiment reported in "From Jerusalem to Jericho": A study of situational and dispositional variables in helping behavior. (Darley & Batson, 1973), participants were theology students asked to give a short talk in another building.

People going between two buildings encountered a shabbily dressed person slumped by the side of the road. Subjects in a hurry to reach their destination were more likely to pass by without stopping.

The experiment was 2 x 3: the participant was asked to either give a short talk on the parable of the Good Samaritan, or on an irrelevant topic. They were either very hurried, hurried, or not hurried by the experimenter.

Hurrying made significant difference in the likelihood of their giving the victim help. The topic of the talk had some influence, according to a reanalysis by (Greenwald, 1975), despite the original paper's claim of no influence. Self-reported personality and religiosity made no difference.

The lesson from this as well as many other social psychology experiments is that seemingly trivial situational variables have a greater impact than personality variables, even though people tend to explain behaviors using personality. See The Person and the Situation: Perspectives of Social Psychology (Lee Ross, Richard E. Nisbett, 2011)

Chap 9. External motivation harms internal motivation

Extrinsic motivations are motivations that "come from the outside", such as money, praise, food. Intrinsic motivations are from the inside, such as self-esteem, happiness. Both can motivate behaviors. However, it's interesting that sometimes extrinsic motivations can harm internal motivation.

In Undermining children's intrinsic interest with extrinsic reward: A test of the" overjustification" hypothesis (Lepper et al, 1973), children are given markers to draw with. Some were told that they would be rewarded with a prize for playing, others got a prize unexpectedly, others were left alone as control group.

After some days, the amount of time children spent playing the markers were: got expected prize < control group < got unexpected prize

The book didn't talk much about why the unexpected prize created higher motivation, but I think it is similar to how gambling addiction comes from variable reward.

There are some explanations for why extrinsic reward lowered subsequent motivation. One is that extrinsic reward provides overjustification effect, where external rewards "crowd out" internal rewards,

Once rewards are no longer offered, interest in the activity is lost; prior intrinsic motivation does not return, and extrinsic rewards must be continuously offered as motivation to sustain the activity.

Another explanation is that humans heuristically view means to an end as undesirable. In (Sagotsky et al, 1982), children were given two activities, playing with crayons and markers. They were equally fun at the beginning, but one group was told that in order to play with crayons, they had to play with markers first. After a while, they became less interested in playing with markers. The other group, the reverse.

I think this is the psychological basis of some ethical intuitions in the style of Kant:

we should never act in such a way that we treat humanity as a means only but always as an end in itself.

A third explanation is that people consider extrinsic rewards a threat to their freedom and autonomy, and thus tend to rebel against it. I saw a news today about Amazon's program to gamify work. Some complained that it was threatening the workers' autonomy, which is a strange complaint: if gamification actually increases intrinsic motivation for work, doesn't it increase autonomy? Autonomy is freedom to follow one's intrinsic motivation, and thus, if a worker acquires an intrinsic motivation to do a good job, they would have more autonomy.

I think this complaint can be explained as a different kind of autonomy: freedom from prediction. Humans are evolved to want to be unpredictable (at least by others), because to be predictable is to be threatened by manipulation, which often decreases fitness.

Chap 10. Actor-observer asymmetry

Other people did what they did because of who they are. We did what we did because of outside events.

In 1975, parts of the Watergate scandal was recreated in a very dramatic psychology experiment, reported in Ubiquitous Watergate: An attributional analysis (West, 1975).

80 criminology students were asked to meet the experimenter privately for a mysterious reason. There, they were asked to join a burglary team for secret documents in an ad agency. There were four versions presented:

  1. The burglary plan was sponsored by a government agency, for secret investigation purposes. Government would provide immunity if caught.
  2. Same, but without immunity.
  3. The plan was sponsored by a rival ad agency, with $2000 reward.
  4. The student was asked to only join a test run of the plan, without stealing anything.

Afterwards, they were debriefed and asked to explain their decision to join/not join.

Separately, 238 psychology students were presented the above situation, and asked to guess what percentage would agree to the plan.

Then, half of the participants were asked, "Suppose John agreed to participate, explain why John agreed."


  • About 45% of participants agreed to join the burglary in the government-with-immunity situation. Otherwise, about 10%.
  • Most students in the second part thought they would not agree to the burglary plan.
  • Students in the first part who agreed to join the burglary explained their behavior as due to the circumstances.
  • Students in the second part explained the hypothetical John's behavior as due to John's personality.

The criminology students were "actors", and the psychology students were "observers". An asymmetry was that actors attributed their behavior to situations, while the observers attributed to personalities. This is the actor-observer asymmetry.

Complications in this asymmetry are noted in The actor-observer asymmetry in attribution: A (surprising) meta-analysis (Malle, 2006). Malle found that there are two kinds of biases: when the behavior is negative, the actor blames the situation and the observer blames the person. When the behavior is positive, the reverse happens. As such, this can be explained as a self-serving bias.

The authors conclude with a funny note:

it's interesting to how athletes often publically thank the Lord for a personal victory, but do not publically blame the Lord for a defeat!

Chap 11. We are number 1

They never shout, "They are number 1."

People like to think good about themselves. Even in collectivistic societies, people regard themselves as above average in collectivistic traits, according to Pancultural self-enhancement (Sedikides et al, 2003)

Americans... self-enhanced on individualistic attributes, whereas Japanese... self-enhanced on collectivistic attributes

An experiment is reported in Basking in reflected glory: Three (football) field studies (Ciadini et al, 1976), where students are asked to describe a recent university sports team's victory/defeat. Before that, half received criticisms that decreased to their self-esteem, and others received praises that increased their self-esteem.

The result was that among those who had higher self-esteem, they described the sports outcome using "we won" or "we lost" 1/4 of the times. For those who had lower self-esteem, they used "we won" 40% of the times when the team won, but used "we lost" only 14% of the times when the team lost.

The explanation is that people in need of boosts to self-esteem try to BIRG (Basking in reflected glory) and CORF (Cut off from reflected failures). Reflected glory also improves their social standing.

Methods of increasing one's social standing are called impression management, and include:

  • BIRG and CORF, as noted above;
  • ingratiation: we praise and agree with others, so as to be liked;
  • self-handicapping: a student gets drunk before a big test, so that if they fail, they could blame on the drunkenness instead of their study ability;
  • exemplification: behave virtuously and make sure others saw it.

A lot of these techniques are listed in (Jones and Pittman, 1982).

Chap 12. Deindividuation

Effects of deindividuation variables on stealing among Halloween trick-or-treaters (Diener et al, 1976) reported an experiment in real life.

The experiment was run in a Halloween. An experimenter place a bowl of candy in her living room for trick-or-treaters. A hidden recorder observes. In one condition, the woman asked the children identification questions such as their names. In the other condition, children were completely anonymous. Some children came individually, others in a group.

In each condition, the woman invited the children in, claimed she had something in the kitchen she had to tend to, and told each child to take only one candy.

Result: being in a group and being anonymous both increased frequency of transgression (taking more than one candy). If the first child to take candies in a group transgressed, other children were also more likely to transgress.

The authors then defined deindividuation as when private self-awareness is reduced.

The truly deindividuated person is alleged to pay scant atetntion to personal values and moral codes... to be inordinately sensitive to cues in the immediate environment.

One study, The baiting crowd in episodes of threatened suicide (Mann, 1981), examined 21 cases from newspapers, in which crowds were present when a person threatened to jump off a high place.

Baiting or jeering occurred in 10 of the cases. Analysis of newspaper accounts of the episodes suggested several deindividuation factors that might contribute to the baiting phenomenon: membership in a large crowd, the cover of nighttime, and physical distance between crowd and victim (all factors associated with anonymity).

Two theories of why deindividuation were given. One is that anonymity makes people feel safe to transgress. Another is that (Reicher & Postmes, 1995) people in a crowd would categorize themselves mainly by their social identity, and their behaviors would reflect the group norm than their personal norms.

I was disappointed that the authors did not give evolutionary psychological explanations for deindividuation. Humans are the only animals that wage wars. A deindividuation effect can be an evolutionary adaptation to prepare humans to fight more effectively in a crowd.

Chap 13. Mere exposure effect

People prefer familiar things. Really, that's quite a banal observation. What's delightful about this chapter is the ingenuity of the experiment design.

Think about your own face. You see them in a mirror image (unless you take a selfie), but others see it directly. This means that you are familiar with your face in the mirror image, but others in the direct image.

This is exploited in Reversed facial images and the mere-exposure hypothesis (Mita et al, 1977). Couples were separately shown photos of the female one's face, some mirrored, others not. They were asked to pick the one they prefer. The female one preferred the mirrored photo, and the male one preferred the direct photo.

Mere exposure effect is robust in real life and across species. (Grush et al, 1978) found that

previous or media exposure alone successfully predicted 83% of the [US congress election] primary winners

And (Cross et al, 1967) found rats who heard Mozart music in infancy preferred Mozart over Schoenberg as adults, and vice versa.

One possible evolutionary psychological explanation were given: preference familiarity is safer, and thus more adaptive. The authors warned however that it's not so simple, as people also have a preference for mild novelty.

Chap 14. Shortcomings of introspection

This chapter reviews a study that shows a particular instance of introspection failure:

people's ideas about how their minds work stem not from private insights but from public knowledge. Unfortunately, however, this public knowledge is often not accurate. It is based on intuitive theories, widely shared throughout society, that are often mistaken.

The book referenced Verbal reports about causal influences on social judgments: Private access versus public theories (Nisbett, 1977), although I find Telling More Than We Can Know: Verbal Reports on Mental Processes (Nisbett & Wilson, 1977) to be better.

Subjects are sometimes (a) unaware of the existence of a stimulus that importantly influenced a response, (b) unaware of the existence of the response, and (c) unaware that the stimulus has affected the response. It is proposed that when people attempt to report on their cognitive processes... they do not do so on the basis of any true introspection. Instead, their reports are based on a priori, implicit causal theories, or judgments about the extent to which a particular stimulus is a plausible cause of a given response.

In the experiment, a subject is given a fictitious application from Jill for the job of staff at crisis center. These applications are the same except on a few attributes of the applicant: attractiveness, intelligence, etc. Then, the subject is asked how much each attribute is correlated with the decision to accept.

The situation is then described to some observers (who didn't do the job application review), who are asked how much each attribute is correlated with the decision of the subject to accept.

Subjects who read that Jill had once been involved in a serious car accident claimed that the event had made them view her as a more sympathetic person. However, according to the ratings they later gave, this event had exerted no impact... the only exception pertained to ratings of Jill's intelligence. Here, an almost perfect correlation emerged between how subjects' judgments had actually shifted and how much they believed they had shifted. Why so? The researchers argued that there are explicit rules, widely known throughout a culture, for ascribing intelligence to people. Because subjects could readily recognize whether a given factor was relevant to intelligence, they could reliably guess whether they would have taken it into consideration.

The determinations of subjects and observers coincided almost exactly.

There are other introspection failures demonstrated by social psychology. People are unaware of the halo effect at work in their own judgments of others (Nisbett &Wilson, 1977). People are unaware of the source of their own arousal. People are unaware of their bias even if they know of such bias (Pronin et al, 2002).

In a further twist, introspection can degrade judgment. In (Wilson & Kraft, 1993), participants reported how they felt about their romantic partners. Their expressed feelings correlated well with the duration of relationship. However, if they introspected on the reason of their feelings, before reporting their feelings, the correlation disappeared.

The authors conclude by suggesting that traveling, by putting oneself into novel situations, would be particularly helpful for one to know oneself.

Chap 15. Self-fulfilling prophecies

Again, a very well-known subject with a lot already written. This chapter reviews Social perception and interpersonal behavior: On the self-fulfilling nature of social stereotypes (Snyder et al, 1977)

Male "perceivers" interacted with female "targets" whom they believed to be physically attractive/unattractive. Tape recordings of each participant's conversational behavior were analyzed by naive observer judges for evidence of behavioral confirmation... targets who were perceived to be physically attractive came to behave in a friendly, likeable, and sociable manner in comparison with targets whose perceivers regarded them as unattractive. It is suggested that theories in cognitive social psychology attend to the ways in which perceivers create the information that they process in addition to the ways that they process that information.

Philosophically, a self-fulfilling prophecy is a prediction about a future that is true iff the act of prediction is done. Usually, predictions themselves are supposed to be independent of the future that they talk about. Of course, all useful predictions must affect the future -- the predictor would try to profit from the prediction. However, such effects on the future are on the predictor, not on the predicted.

Social psychologists have found that human behaviors are more influenced by the situation than the personality (as noted in The Person and the Situation book). Snyder et al suggested that, in fact, personality traits are one of those self-fulfilling prophecies.

our believing that others possess certain traits may cause us to behave in certain consistent ways toward them. This may cause them to behave in consistent ways in our presence.

In other words, a lot of the persistence of personality could arise from the fundamental attribution error.

Chap 16. How to live like a predeterminist

So then, God has mercy on whom he chooses to have mercy, and he hardens whom he chooses to harden. -- Romans 9:18, which Calvinists quote a lot.

Suppose an urge to smoke and a propensity to lung cancer are both genetically determined, and smoking does not cause lung cancer, why not smoke? If you feel the urge to smoke, it's already too late.

Believers of Calvinism think that God has chosen some people to be saved, and others are damned. Those who are favored by God would both be naturally free from the urge to sin in this world, and enjoy paradise after death. Those who are not, would feel the urge to sin in this world, and go to hell after death.

So if a Calvinist feels an urge to sin, it's already too late. Why not sin? Instead, Calvinists keep resisting the urge to sin, and moreover, deny that they are resisting such urges, and insisting that they are effortlessly virtuous, evidence of God's favor.

In Causal versus diagnostic contingencies: On self-deception and on the voter's illusion (Quattrone & Tversky, 1984) two experiments are reported.

In the first one, participants exercised, then were asked to put their hands in ice water until the pain makes them withdraw. Then they were told a version of the lung cancer puzzle: There are two kinds of hearts, type 1 and type 2, caused by unchangeable genetics. Type 1 heart is associated both with health and with a higher tolerance to the ice water after exercise. Type 2 heart is associated with early death and a lower tolerance. They then did the ice water test again, and they exhibited longer tolerance to the ice water, even though many of them denied that they were trying to do so.

In the second experiment, subjects encountered one of two theories about the sort of voters who determine the margin of victory in an election. Only one of the theories would enable voting subjects to imagine that they could "induce" other like-minded persons to vote. As predicted, more subjects indicated that they would vote given that theory than given a theory in which the subject's vote would not be diagnostic of the electoral outcome, although the causal impact of the subject's vote is the same under both theories

One explanation is that the unconsciousness deceived the consciousness, but the authors find this unreasonable, for it still does not explain what motivates the unconsciousness to deceive. They instead favored Greenwald's theory that people avoid analyzing in detail threatening information, just like how we throw away junk mail without looking in detail.

In conclusion, self-deception is not the result of one center of intelligence hoodwinking the other. Rather, it is the result of a low-level screening process that banishes suspicious cognitions before they have the opportunity to be fully entertained by the conscious mind.

Similarity to superrationality and acausal trade.

The behavior of Calvinists is similar to superrationality and acausal trade, in which agents behave in a way that is diagnostic of good outcomes, even if it does not cause good outcomes.

Assuming the superrational player has access to their opponents' source codes/simulations, the superrationality strategy can be justified, but then it would just be usual rationality.

I think normative decision theories are incompatible with sufficiently good prediction. Normative decisions are only defined for agents with apparent free will. An agent apparently has free will only to someone who cannot predict the agent's behavior well. Superrationality and acausal trade both attempt to make a decision theory for agents that are aware that they are too predictable (to themselves or to someone they play with). This is similar to the situation where someone sees the future and then "decides" to rebel against the future. Either they saw the true future and did not rebel, or they did not see the true future at all. It's illogical to say they both saw the future and rebelled against it.

Similar problems happen with Scott Aaronson's solution to Newcomb's paradox (I'm a "Wittengenstein"). A determinist who is self-aware of their determinism would, instead of offering a decision theory ("I should take one box because..."), offer a prediction theory ("I probably would take one box because...").

Chap 17. Partisan perceptions of media bias

People often complain of media biases. People report differently about the same event. Why?

In The hostile media phenomenon: biased perception and perceptions of media bias in coverage of the Beirut massacre (Vallone et al, 1985), the researchers studied how people perceived news about the Bairut massacre,

killing of civilians, mostly Palestinians and Lebanese Shiites... carried out by the militia under the eyes of their Israeli allies.

The researchers took some neutral reports on the event, and as expected, pro-Israel people thought they are biased to be anti-Israel, while anti-Israel people thought they are biased to be pro-Israel.

In a study on biases (Lord et al, 1984), participants avoided bias by this command:

"Ask yourself at each step whether you would have made the same evaluations had exactly the same study produced results on the other side of the issue.

Chap 18. Empathy-altruism hypothesis

Several theorized psychological mechanisms of human altruistic actions are studied in More evidence that empathy is a source of altruistic motivation (Batson, 1982) reported an experiment on whether people would help a person in need.

It was found that: If (empathy OR guilt), then (helping). That is, people can be motivated to act altruistically by empathy without expectation of gain, or to gain relief from guilt. This argues against the theory of psychological hedonism.

Other potential sources of altruism are collectivism (act for the benefit of a group) and principlism (uphold a principle for its own sake). Effective altruism is one example of principlism based on utilitarianism.

Chap 19. Expanding the self to include the other

A psychological phenomenon of love (close personal relationships, such as lover, best friend) is to include that person in one's self. This involves perceiving, and allocating resources to, that person, in a similar way as to one's self.

Three experiments are described, from Close relationships as including other in the self (Aron, 1991).

  1. When allocating money, they allocate about the same to themself as to their friend.

  2. They were asked to imagine nouns paired with their selves, mothers, or strangers. They recalled fewer nouns imagined with self or mother than nouns imagined with a stranger, suggesting that mother was processed more like self than a stranger.
    They explained the reason why it was recalled less by that we usually look at strangers directly, but only ourselves upon reflection (literal or not), and so it's harder to imagine ourselves than strangers.

  3. When faced with a task to sort a list of adjectives into 4 piles: "true/false about me, and true/false about my spouse", they reacted slower on adjectives that were true about one but false about the other. This was explained by that differences between one's own and a close other's properties caused dissonance in the same way that holding opposite attitudes within oneself can cause dissonance.

Chap 20. Believing precedes disbelieving

Descartes divided the mind up into intellect and will. The intellect writes up potential beliefs about the world; the will then chooses which to endorse. Spinoza said that we believe everything that we happen to understand, and then disbelieve only if we find it necessary. You Can't Not Believe Everything You Read (Gilbert, 1993) presented three experiments that supports Spinoza's theory, and discussed its sociological effect.

... we asked subjects in Experiment 1 to play the role of a trial judge and to make sentencing decisions about an ostensibly real criminal defendant. Subjects were given some information about the defendant that was known to be false and were occasionally interrupted [by a distraction task]... We predicted that interruption would cause subjects to continue to believe the false information they accepted on comprehension and that these beliefs would exert a profound influence on their sentencing of the defendant...

Experiments 1 and 2 provide support for the Spinozan hypothesis: When people are prevented from unbelieving the assertions they comprehend... they did not merely recall that such assertions were said to be true, but they actually behaved as though they believed the assertions.

If you want to read more, I have written in detail about this.

Chap 21. Inferred memories

When we recall a memory, that memory is an inference about past based on a number of clues that we have in the present. It is not necessarily accurate.

Experiment from Women's theories of menstruation and biases in recall of menstrual symptoms (McFarland, 1989) found that when women report, day-to-day, their unpleasant emotions, there is no difference between premenstrual, menstrual, and inter-menstrual days (they feel equally unpleasant). But when asked to recall how unpleasant it was, they recall significantly more unpleasant pre-menstrual and menstrual days, and less unpleasant inter-menstrual days.

This is explained by that, when they recall, they used intuitive theories about PMS to infer "how it must have felt" instead of "how it actually felt". This also, as a side effect, casts doubt on whether PMS actually exists.

Memories can be completely made up, as in repressed memory therapies.

The fact that those inferences about the past are felt as genuine recalls, shows how little conscious introspection can give true knowledge about the self.

Chap 22. Ironic process theory

Try to not to think of a polar bear!

The theory of ironic process is that there is a cognitive process called intender who is looking for contents that matches some desired mental state. There is also a monitor who notifies consciousness about errant thoughts.

The intender is a costly process, and the monitor is a cheap process, so when one is under cognitive load, the intender doesn't work well, but the monitor still works well, and ironically, trying to not think of something results in thinking of it.

Ironic Processes of Mental Control (Wegner, 1994) reported an experiment. Participants were asked to consciously improve/deprove their moods with happy/sad thoughts. Half were also asked to do a memory task as cognitive load.

Those not under cognitive load were successful in their mood control, while those under cognitive load achieved the opposite.

This suggests that if you are under some cognitive load (such as busy studying), and you want to improve your mood, you should try consciously to feel worse. Also, if you are in a noisy and distracting environment, and want to sleep, you should try to stay awake.

Another experiment showed that people who try to avoid sexist language become ironically more prone to sexist language when under cognitive load. This is true no matter if they are sexist or not.

Chap 23. Implicit Association Test

In Single-target implicit association tests (ST-IAT) predict voting behavior of decided and undecided voters in swiss referendums (Raccuia, 2016), compared to self-reported political orientation, implicit association was found to be a weaker, but somewhat independent, predictor of voting behavior.

Other similar methods to probe the unconsciousness are studied, and the results are new and mixed.

Chap 24. Prospect theory

People don't behave as expectation-maximizers. Instead they are better modelled by prospect theory:

  1. Gains and losses are measured compared to a changeable default, instead of an absolute zero.
  2. Losses are weighted more than gains, and both have decreasing marginal utilities.
  3. People are more risk-averse with respect to gains, and more risk-loving with respect to losses. prospect-theory-wikipedia

An experiment The systematic influence of gain-and loss-framed messages on interest in and use of different types of health behavior (Rothman et al, 1999). It was found that people used more bacteria-killing mouth wash, if they received positive advertising (about maintaining good health). They used more disclosing mouth wash (which merely detects dental diseases) if they received negative advertising (about the potential disease).

This theory, along with some others, is explained in great detail in Thinking, Fast and Slow (Kahnemann, 2011), which I recommend.

Other mental heuristics include mental accounting (Thaler, 1980), with its own set of irrational effects.

Chap 25. Social isolation increases aggression

If you can't join them, beat them: Effects of social exclusion on aggressive behavior (Twenge, 2001)

Social exclusion was manipulated by telling people that they would end up alone later in life or that other participants had rejected them. These manipulations caused participants to behave more aggressively. Excluded people issued a more negative job evaluation against someone who insulted them, blasted a target with higher levels of aversive noise both when the target had insulted them and when no interaction had occurred. However, excluded people were not more aggressive toward someone who issued praise.

In particular,

These responses were specific to social exclusion and were not mediated by emotion.

This was shown by two experimental facts:

  1. Participants who were told they would end up alone later in life or that other participants had rejected them, did not feel worse than average.

  2. Participants who were told they would end up unlucky later in life, did not act more aggressively than average.

Some psychological theories are given. One is self-determination theory from Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being (Deci and Ryan, 2000), which says that people have three needs:

  • relatedness (to some other people)
  • efficacy (can do important things)
  • autonomy (can control their own future)

Other relevant factors are self-esteem, and stability over time. Stability and level of self-esteem as predictors of anger arousal and hostility (Kernis et al, 1989) found that in feelings of anger and hostility,

unstable high self-esteem > low self-esteem > stable high self-esteem

There is no evolutionary explanation, though. Social exclusion causes fewer offsprings, and aggression only worsens it. An evolutionary psychological explanation would be good. Either it has evolutionary benefit, or it is a side effect of something else.

Chap 26. Social effects of gossiping

Gossip is found to have a prosocial function. The virtues of gossip: Reputational information sharing as prosocial behavior (Feinberg, 2012)

... prosocial gossip, the sharing of negative evaluative information about a target in a way that protects others from antisocial or exploitative behavior.

In the study, they found experimental support for four hypotheses about the function of gossip:

  • prosocial: gossip is motivated by a desire to protect vulnerable people, without promise of material reward.
  • frustration: seeing antisocial behavior makes people feel bad, which . Prosocial people are more prone to this frustration.
  • relief: gossiping reduces the frustration.
  • deterrence: threat of gossip makes antisocial people behave more prosocially.

Chap 27. Fear of death

Good news: we will be worm food one day!

Good news for worms, I meant.

Terror management theory argues that the terror of death creates such a profound, subconscious, anxiety, that humans spend their lives denying it in various ways, creating culture, religion, and many other social phenomena in the process.

In this chapter are reviewed the first 4 of the 7 experiments from How sweet it is to be loved by you: the role of perceived regard in the terror management of close relationships (CR Cox, J Arndt, 2012). This paper studies

... whether people turn to close relationships to manage the awareness of mortality because they serve as a source of perceived regard.

Perceived regard means "am I a good person as viewed by someone else?" The paper in particular showed that people who have death on their mind exaggerate how much they think they are loved by a partner. Perceived regard from their own selves, and from average strangers, did not change. Having intense physical pain on the mind also did nothing.

They also found that having death on the mind makes people claim to love their partners more. They theorized that this is mediated by increased perceived regard:

death on the mind -> more perceived regard from their partner -> more love for their partner

Study 4 revealed that activating thoughts of perceived regard from a partner in response to MS reduced death-thought accessibility. Studies 5 and 6 demonstrated that MS led high relationship contingent self-esteem individuals to exaggerate perceived regard from a partner, and this heightened regard led to greater commitment to one's partner. Study 7 examined attachment style differences and found that after MS, anxious individuals exaggerated how positively their parents see them, whereas secure individuals exaggerated how positively their romantic partners see them. Together, the present results suggest that perceptions of regard play an important role in why people pursue close relationships in the face of existential concerns.

Personal comment: It has been commented that Transhumanism can be analyzed as a religion. Is there value in analyzing transhumanism through terror management theory? There is at least one paper, Software immortals: Science or faith? (Proudfoot, 2012), that did so. This is important, because if transhumanism is indeed a religion, then the chance is high that it is deluded/unfalsifiable, like most religions have been shown to be.

Also, this would explain why moral nihilism is usually suffered as a mental disease than accepted as a working hypothesis. Despite its theoretical simplicity and moderate empirical support, it just doesn't offer any protection against terror of death.

Chap 28. Motivated belief in free will

Free to punish: A motivated account of free will belief (Clark, 2014)

a key factor promoting belief in free will is a fundamental desire to hold others morally responsible for their wrongful behaviors

Five experiments from the paper are recounted in detail. The authors praised the paper highly for its comprehensiveness.

participants reported greater belief in free will after considering an immoral action than a morally neutral one... due to heightened punitive motivations... reading about others’ immoral behaviors reduced the perceived merit of anti-free-will research... the real-world prevalence of immoral behavior (as measured by crime and homicide rates) predicted free will belief on a country level.

Taken together, these results provide a potential explanation for the strength and prevalence of belief in free will: It is functional for holding others morally responsible and facilitates justifiably punishing harmful members of society.

Personal comment: Instead of philosophically studying whether free will exists, it's more productive to assume it doesn't exist, and see what behaviors can be explained. If everything can be explained without free will, then the problem of free will dissolves. Else, we will have concentrated what free will is for, and made subsequent studies more focused.

It is also useful to study the human intuitive belief in free will, as important phenomena about humans, independent of whether they are right or wrong. This is analogous to the study of folk psychology and naive physics. See From Uncaused Will to Conscious Choice: The Need to Study, Not Speculate About People’s Folk Concept of Free Will (Monroe, 2009)

the core of people’s concept of free will is a choice that fulfills one’s desires and is free from internal or external constraints. No evidence was found for metaphysical assumptions about dualism or indeterminism.

In the "Afterthoughts", the authors considered what a post-free-will society could be like. I think that such a society's theory of crime and punishment would be more like "because this follows the natural order of things", than "because criminals are morally bad".

Think of the joke about "my brain made me commit the crime"

The criminal: "My brain made me commit the crime." The judge: "My brain made me sentence you."

And now, instead of taking it as a joke, imagine both of them saying them very seriously. That's what I think could be true in the future.

New Comment
1 comment, sorted by Click to highlight new comments since: Today at 1:11 AM

The first edition of this book was published in 2003. In 2005, Ioannidis' paper "Why Most Published Research Findings Are False" started the reproducibility avalanche. How well have these experiments replicated? My university library only has the first edition. I can see from the Amazon preview of the second edition (2017) that the authors address this, but I can't see enough pages to see what their response is. I understand from other sources that priming and ego-depletion have not stood up well.