Influence by Dr. Robert Cialdini covers some useful and intriguing psychological constructs that shape how we act in society. As I mentioned in my previous post, one must take care in applying ethics to the situation and analyzing which subset of tactics is acceptable.

Below I walk through the book by chapter highlighting topics and examples that I found interesting.

Chapter 1: Weapons of Influence

One theme this book exposes is the subconscious things we would be surprised influence us but actually do. There are empirical studies throughout the book supporting the various points. Take these three requests in the context of waiting in line for a copier and asking to skip to the font.

  • “Excuse me, I have five pages. May I use the Xerox machine because I’m in a rush?” 94% success
  • “Excuse me, I have five pages. May I use the Xerox machine?” 60% success
  • “Excuse me, I have five pages. May I use the Xerox machine because I have to make some copies?” 93% success

One would assume that being in a rush would lead to a greater chance of skipping in line, but it really was the trigger word because.

Contrast principle: much more likely to purchase a $95 sweater after purchasing a $495 suit than the other way around. The same thing applies to realtors showing overpriced, dilapidated properties leading up to the “good” deal they want to sell their clients on. The contrast fresh in the client’s mind makes him or her see the value of the final house after the setup sequence.

Another theme of the book is that humans use pattern matching and take shortcuts to make sense of the world. For example, expensive=good. If we analyzed everything from first principles every time, we would function at the speed required to make decisions in normal life. The obvious observation this raises is we generally think of stereotyping as bad, but this emotionless observation says it is required to function in life. I have some questions. Is there something else we can categorize certain learning patterns as? Pattern-matching? Inference? Or is it all just plain old stereotyping? Inference conjures the thought of machine learning models and generative AI. Generating Barbies from various countries exposed the stereotypes and racism unchecked inference can lead to.

Chapter 2: Reciprocation

Reciprocation is a phenomenon that is common in American culture. When someone helps us out, we want to return the favor and help them in return. This can be a solid foundation of good working relationships and even friendships, or it can be manipulated out of balance. This is because the initiator chooses the both first favor and the requested return favor. The indebtedness can be exploited when the return favor is much larger than the original.

In Luke 14:12-14, Jesus says:

He said also to the man who had invited him, “When you give a dinner or a banquet, do not invite your friends or your brothers or your relatives or rich neighbors, lest they also invite you in return and you be repaid.

But when you give a feast, invite the poor, the crippled, the lame, the blind, and you will be blessed, because they cannot repay you. For you will be repaid at the resurrection of the just.”

A variation on the reciprocation of favors is reciprocal concession. In this scenario, the initiator does not give anything to start the process. He simply makes a large request, knowing the answer will probably be “no.” The reciprocation comes into play with the follow-up request. The initiator’s concession down to the smaller request will make the subject feel pressured to return the “favor” and concede by agreeing to the second proposition.

My kids have figured out this principle already! They will sometimes ask to bring a toy into whatever destination we are driving to. When I say no, they then ask to bring it but leave it in the van. I’m more likely to say yes due to the reciprocal concession principle. If they had led with their second request, I would have more likely said a flat no.

Here is the defense angle: don’t eschew all favors as some will be genuine. Pay attention to if the requested reciprocation is substantially larger than the first act or if negotiation ensues “walking down” to smaller and smaller concessions. Then you know it was a ploy from the beginning and do not need to feel indebted to repay what was given with ulterior motives.

Chapter 3: Commitment and Consistency

Continuing down a sub-optimal path due to resources already invested is the definition of sunk cost fallacy. This can be a risk in software engineering. One quote says that good engineers are the ones willing to throw away their code when the opportunity arises. Sunk cost is a commitment to previous actions, but the book broadens the concept to also include consistency with previous statements.

One case study was how to keep toy sales up after big sale seasons like Black Friday and Christmas. Toy companies sometimes use this sequence to make post-Christmas sales.

  1. Launch strong ad campaigns for a specific toy during the Christmas shopping season
  2. Parents will make commitments to kids to get that specific toy
  3. Undersupply that toy to stores so it is not available during holiday shopping
  4. Run more ads in January to remind kids of the previous commitments
  5. Parents, wanting to stay true to their previous commitments, will buy said toy even though they already bought something else for the kids’ presents.

Edgar Schein studied Some Observations on Chinese Methods of Handling Prisoners of War (JSTOR) and also cited in the Defense Technical Information Center. Some of his findings were incorporated in this chapter.

The Chinese communist POW camps tried to indoctrinate prisoners via the “lenient policy.” The program focused on getting soldiers to make concessions about how things were not perfect back home, and maybe Communism wasn’t totally wrong. The prisoners were then coerced to write these statements down. Written evidence helped solidify it in their own minds as well as spread the statements to others who were more resistive (see chapter 4 - social proof.) Another avenue for collecting these statements was political essay contests. They had various ways of encouraging participation - small prizes, middle-of-the-road essays winning just enough to keep people hopeful but still force some concessions.

Traits that make commitments the most effective are being active, public, and effortful (page 92.) A detracting factor is too much external motivation. Minimizing external pressure like large prizes or shaming makes someone own their decision in a stronger way. This is because they cannot blame any other reason for their commitment other than it’s what they wanted to do and “that’s who they are.” This strategy is exemplified in both the POW camp essay choice of prizes as well as modern-day fraternity initiations. The essay contest prizes were small, only a couple of cigarettes or pieces of fruit. Fraternities keep initiation hazing separate from their community service so there is no “mental out” that an applicant can say he did it for a cause other than he wanted to join. On the pressure and shaming side, it has been shown that kids will more deeply internalize parental guidelines if strong threats of punishment are not given. Wariness of punishment only applies if they think they might get caught. Knowing something is inherently the right thing to do is trickier to convince of, but results in stronger adherence.

Chapter 4: Social Proof

Social proof is another shortcut our brain takes when trying to determine the correct action. Learning is picking up patterns in the world and internalizing them. Sometimes that is from first principles, and others it is by observation. The majority of the time, this is a valid learning and reaction mechanism. However, manipulated responses or cascading blind following can make the heuristic incorrect.

Harmless examples of social proof include laugh tracks in comedy TV, claqing in opera houses, and salting tip jars with dollar bills instead of coins. Even when people are aware of those practices, they are still likely to be influenced by them. Unfortunately, very negative examples also exist. Suicide rates peak after publicized suicide stories on the news. Jonestown was a horrific example of how a leader did not have to convince everyone individually, only enough to start a tidal wave of compliance.

A practical phenomenon we will probably run into in our daily lives is pluralistic ignorance. This is where signs exist that something is not right, but bystanders take cues from other bystanders to go on their way and not act. They suppose from the inaction of others that they are misinterpreting the signs of trouble and everything is actually fine. This makes me think of circular dependencies in software. The more public the situation, the less likely the victim is to receive help. When there is only one witness, it is clear that a specific person needs to act. A larger number of witnesses leads to more ambiguity as to who should act. The way to counter this tendency is to assign a specific task to a specific person. The example from the book is, “You, sir, in the blue jacket, I need help. Call an ambulance.”

Defenses against illegitimate use of social proof take two forms. First, try to spot falsified actions of social proof. This includes paid audience members, paid testimonials in an advertisement (even if that person looks like you), and modern-day “influencers.” Sometimes the actions taken by the crowd will not be fraudulent, just misguided. This leads to the second defense, spotting snowballing errors. An example of this is when multiple drivers change lanes at the same time on a highway, especially during rush hour. If visibility is bad up ahead, cars behind them can try to follow suit thinking there is an obstruction in the road up ahead. Ironically, cars merging over to avoid an anticipated wreck up ahead can cause an actual wreck.

A fascinating situation where “following the crowd” is the logical thing to do is during a bank run. SVB Bank fell to this wave of withdrawals as panic was spreading on Twitter. Bank runs are a tragedy of the commons situation. The only way to quell a bank run is for everyone to stop drawing money and let the bank stabilize. But the undeniable self-preserving move is to withdraw one’s own money while it is still accessible. Most times in a snowballing behavior situation the best thing to do is not to get sucked in, but this is one exception. What are others?

This post covers the first half of Influence by Dr. Robert Cialdini. I will cover the remainder in another post coming up! Thanks for reading, and I hope you found some of the tidbits interesting.