More on Free Will

If you have been reading my blog regularly, you know I have been thinking a lot about free will.  So this weekend I picked up two movies for which free will is part of the background message.  It wasn’t really important to understand the plot or the action or the characters.  But both of them had free will messages.

Minority Report was the first one I saw.  Twice during the movie, the person who the precogs had “seen” commit murder in the future were made aware of this future.  They had the choice to change what they did.  They had the “free will” to use that knowledge to do something different.  Both of them did something different, although not by much (I don’t want to spoil the movie with details.  It’s not relevant to the post).

Then I saw Slaughterhouse Five.  I heard that the Vonnegut book was fantastic so I was really looking forward to the movie.  The free will part of the movie is that time is just an illusion.  Everything that has ever happened, is happening now, and will ever happen, are really all happening at the same time.  So whatever you will do, you have also already done and are doing now.  In that view, the concept of free will doesn't make sense. This is a cool premise, but the movie sucked.  I hope the book was better.

How we really think. Not what you might expect.

Have you heard of the Trolley Problem paradox?  It is the best illustration of how human thinking is not rational like a computer.  That doesn’t mean it’s irrationally silly; it is actually a lot more complicated.  There are many pieces that evolved to do totally different things working in conjunction but not working together.  That’s what evolution does, compared to an engineering lab.

The reason the Trolley Problem is such a great illustration is that it is really simple to understand.  And it is very reliable – it has been tested over and over again with all different kinds of people and dozens of variations.  And there have been fMRI scan studies that clearly show what brain areas are driving each part of the decision.

Here is the basic problem.  A trolley is coming uncontrolled down the track.  If you don’t do anything, it will run into and kill 5 people instantly.  Or you can pull a track switch and move the trolley to a track that only has one person on it.  You will be saving the 5 people, but you will be signing the death sentence of the other one.  What do you do?

Then here is another problem.  Same basic premise – a trolley coming uncontrolled down the track, about to run into 5 people and kill them instantly.  There is another track that with nobody on it.  But you are on top of a bridge and the switch to shift the trolley is on the ground below.  You are not heavy enough to jump off the bridge and land on the switch to shift it.  So your only choice is either do nothing and let the five people one the track die, or to push a heavy guy standing next to you off the bridge, killing him, but saving the 5 people on the track.  What do you do? 

There are no right or wrong answers.  If you are a utilitarian (always choose the option that does the most good for the most people), then you would pull the switch and you would push the heavy guy to his death off the bridge.

If you don’t believe that you (or anyone) have the right to decide who should live and who should die, no matter how many people are involved, then you should not pull the switch and you should not push the heavy guy off the bridge. 

But what makes this a paradox is that most people pull the switch, but don’t push the heavy guy off the bridge (you probably made the same choices).  We seem to have one ethical value in one case and the other ethical value in the other case.  And when asked, we have a lot of trouble explaining the discrepancy.  This inconsistency broke all the models of early psychologists and early philosophers.  There were hypotheses, but nothing could ever be proven to a statistical probability.  It wasn’t until the invention of fMRI that we were able to learn what was happening.  And it is really fascinating !!!

So here is what we have learned.  In the first case, when you are pulling a switch, the cortex does most of the heavy lifting.  This is the rational part of the brain that thinks utilitarian.  The most good is to pull the switch.

But in the second case, the “ick” factor (composed of visceral feelings of guilt, sadness, and disgust) of personally contacting the person who is killed causes the amygdala and gyrus to override the cortex and make the opposite decision.  The amygdala and gyrus are evolutionarily older parts of the brain, so they win in any disagreement.  Not for logical reasons, for emotional reasons.

But people who have injuries in their amygdala/gyrus always make the utilitarian choice (push the heavy guy off the bridge).  They tend to be sociopaths because they don’t care about the person who is hurt as a result.  At least not emotionally.

And if we fatigue the cortex by making a person do a lot of thinking and reasoning, then they are more likely to make the less “icky” choice (they won’t even pull the switch). When rational thinking is tired, emotional thinking is happy to jump into the gap.

There are also personality differences.  People with more trait anger (not just temporarily angry, but more angry in general) make more utilitarian judgments (push the guy off the bridge).  On the other hand, trait empathy and trait disgust increase the “ick” feeling and make people less utilitarian.

Being temporarily happy (or really any positive emotion) decreases your ick feeling (guilt, sadness, and disgust) so you are more likely to be utilitarian. 

These results have been reproduced in a wide variety of decisions, not just trolleys.  The scenarios have gone with administering the death penalty, throwing a hand grenade, hitting someone by bowling a ball down the lane, and more.  Researchers have looked at old people, young people, ethical experts, ethical ignoramuses, males, females, different cultures, religions, etc.  There are many individual differences that change the exact percentages, but the main effect is generally the same.  More of each group pull the switch than don’t and more people don’t push the heavy guy off the bridge than do.  Except extreme groups like sociopaths.

Eyewitnesses

This originally started as an investigation into the various ways that error can be introduced into eyewitness (as opposed to expert witness) testimony.  With over two decades of human factors experience under my belt, I am well aware of how memory works, the challenges it faces, and how counterintuitive much of it is.  But for this study, I wanted to limit the scope of my review to studies specifically about eyewitnesses.  For Daubert hearings (where the judge decides on the scientific validity and relevance of the expert for the case), it is always helpful to be able to cite specific, focused research because judges don’t always grasp the generalizability of cognitive research.

Even with the cynicism that my experience has forced upon me, I was shocked by just how much evidence there is that eyewitness testimony is perhaps the least reliable kind of evidence available to the court (since hearsay and total bullshit are usually left out).  I plan to write this up in more detail and complete with references for the Annual HFES Conference in the fall.  But I wanted to get something more informal out now while the emotion is still fresh and I can give you all a laugh.

Let’s start with the context in which the event is witnessed.  First of all, the witness is rarely looking directly at the location where the incident occured.  Until their attention is refocused, if it is at all, then it is the notoriously inaccurate peripheral vision that starts the memory out.  And even then, the contents of the witnesses’ existing vision, thoughts, beliefs, daydreams, stereotypes, and past memories are stronger contributors to what gets encoded.  If the witness has been in similar situations before, these are all blended together into the memory of this event.  This causes the witnesses’ memory to condense, smooth out, fill in gaps, omit inconsistent details, and reorder the sequence to fit their personal experience and expectations.

After the witness does notice that something interesting is going on and refocuses, their attention is captured by the most salient and central attributes rather than what is important to be an effective witness.  We know this as the “weapon effect.”  They remember the gun in incredible detail but couldn’t tell you how tall the perp was, the color of his hair, or about the big stain on his pant leg.  This is hilariously captured in the famous Invisible Gorilla test by Daniel Simons.

If it is a stressful event for the witness, the fight or flight instinct starts the release of hormones that degrades memory (who cares what the tiger looks like, just get the hell out of there).  Attention narrows to just the most salient attributes and anything more subtle is missed. 

Then what happens after the event introduces even more inconsistency into the witnesses’ memory.  We all have a tendency to exaggerate the event when we first tell the story, especially if this first telling is to ourselves, friends, or especially the media.  By the time the on-scene investigator starts his or her interview it is way too late.  And unless the investigator is both well trained and well intentioned, it is easy to implant whatever memories will get the case solved as fast and easy as possible.  Consider the difference between these questions:

·         What happened . . . . . ?

·         Was there a weapon?

·         Was there a gun?

·         What color was the gun?

In this last case, witnesses who never saw a gun all of sudden remember one clearly.  And then later on, when they are recounting the story, the gun becomes a permanent part of their story.  By the time trial rolls around, the witness is absolutely sure there was a gun.  The prosecuting attorney can make it even easier by asking the same question, “What color was the gun?”  Just the natural desire for internal consistency would cause even the most well meaning witness to give the same answer as they did the first time. 

What makes this even worse is that juries believe eyewitnesses more than they believe most other kinds of evidence.  As Groucho Marx famously said, "Who are you going to believe, me or your lyin' eyes?"

Society-wide decision making

In my policy demogogue blog, I just wrote about a new company called Blueseed that is using the seasteading model to alleviate the Silicon Valley dearth of qualified tech workers.  Please read it because it is an interesting way to game the system.  But I wanted to arouse a discussion of the HF implications.  Read a couple of the posts on Econlog for a good review of the pros and cons. 

The basic idea of seasteading is that you can create your own society by putting your floating city in international waters and creating a system of laws according to your own values.  The logistics might be tough, but imagine all the different legal systems we could try out to see how they work.  Different kinds of democracies, republics, dictatorships, religion-based governments, pure communism. 

Why is this relevant to HF? Imagine the experimental environments you could create to study decision making, organizational culture, macroergonomics, etc.  I would love to design a series of seasteaded cities with different decision making procedures, run them for a couple of years, and see how the society evolves.

Thoughts??

The effects of self-esteem on willingness to apologize

I was kind of interested in reading about this study because I had never really thought about it much before.  Andrew Howell at Grant MacEwan University (Canada) studied people's willingness to apologize (in general) and then compared them to a whole battery of personality assessments. 

The obvious connections were that people with more compassion and more agreeableness apologize more.  But what were more surprising were the effects of self-esteem.  People with really low self-esteem are hesitant to apologize, even if they feel bad about what they did, because they direct the shame internally.  They feel bad, but may also feel sorry for themselves, and therefore not apologize to others.

Those with really high self-esteem (i.e. egocentrics, narcissists, etc) were also less likely to apologize.  Perhaps they felt entitled to do whatever they did and therefore didn't need to apologize for it.

So how can we use this information?  First, think about how you feel when you do something that most people generally apologize for.  If you apologize more than others, then your self-esteem is right in the sweet spot and you should feel good.  But if you apologize less, think about which end of the tail you fall on.  You may need to work on your self-esteem, or you may need to come off your pedestal.

Consistency v flexibility in rule enforcement

Kristen Laurin from the University of Waterloo has just published some research findings that I think have some important implications for Human Factors, especially in shop floor management (any behavior related to safety, productivity, quality etc.).  Even though her work was actually focused on emigration, it is amazing how it relates to our work.  Hey, no thanks necessary.  Linking disparate areas of research is what I am here for.

What she found is that when rules are absolute (no chance for either changing them or getting away with violating them) we tend to accept them as a fact.  By being attached to our "fact" schema, our subconscious rationalizes some justifications for why we wanted to comply anyway, to maintain our feelings of internal consistency.  We don’t like to feel 100% forced into things, so our subconscious plays tricks on us.  It tells us that it must be a good rule.  That the leaders must be smarter than we thought.

On the other hand, when there is leeway, we associate the rule with a fuzzy likelihood schema. If it's not absolute then we can find an exception.  We don't have 95%, 99%, 50% schema.  Either it's absolute or it's not.  So if it's not, our brain is always looking for loopholes.  Can we get away with violating the rule?  Will there eventually be new leaders who change the rule?  Can we convince the leaders that we deserve an exception to the rule.  If we think there is a chance, then our brain keeps working on solutions.  And because of that, the rationalization never happens.

Dr. Laurin’s research focuses on emigration policies.  She found that if you have no chance of leaving a country because of a dictator, then people’s subconscious rationalizes that the dictatorship isn’t as bad as it seems.  We have stability don’t we?  Security?  This is kind of similar to the Stockholm syndrome experienced by kidnap victims.  But if there is some leeway.  A weak dictator, someone to bribe, etc. this never happens.

So the general conclusion is that if you are going to implement unpopular rules, you need to make them absolute.  You will get that subconscious rationalization and buy-in.  But if there is any leeway, people will keep trying to find loopholes.  And no buy-in.

So why do I think that research on emigration in dictatorships is related to shop floor management?  Well, we always have some rules and policies that are unpopular.  Hopefully not many, but there are always a few.  The key to having them accepted is that they have to be absolute.  100% enforcement on 100% of the relevant employees.  No turning a blind eye when the rule would delay shipment to a good customer.  No exceptions for pet workers.  You might think that this is being the corporate asshole Robert Sutton writes about.  But if you are really 100%, then you get that subconscious rationalization effect.  Workers can’t help but think the rule must be better than they thought.  You must be smarter than they thought for having it.  But it only works if you enforce it 100%.

The results can bleed into the corporate culture.  We can enforce all of our rules and regs 100%.  If a rule needs exceptions, put them in writing or change the rule.  Then we get an entire culture of acceptance and compliance.  If all rules are 100%, you must really be smart and competent.  The rules must be really appropriate and in our best interest.  Consistency in rule enforcement creates this rationalization effect that pervades the entire culture and reinforces itself over time.

But having even just a few rules that are acceptable to violate reduces the compliance culture.  Maybe one or two rules are 100% compliance if they make sense (like wearing a respirator when coal mining).  But the rest . . . . would be hit or miss.

My recommendation – make up the best set of rules you can, write exceptions and exclusions into the official rule.  And then don’t make exceptions on the fly.  The one time that this might be an advantage is well compensated by the extra compliance and buy-in you get from consistency.

More happiness research

Warning, this is a little random - I am having trouble focusing today.

As you know, I am an avid reader of the happiness literature.  Not the psychobabble type, but legitimate research using rigorous and valid methods.  Not that there is anything wrong with that . . . .

Anyway, here are a few recent findings I thought I would share, and then brainstorm a few Human Factors implications. 

One study found that:

• Happiness is composed of 50% genetic predisposition, 10% circumstances, and 40% mental and behavioral strategies to cope.
• Varying the coping strategies is better than using the same one repetitively.

So what is the HF implication?  First, it means that individual differences are the most important thing when it comes to attitude (in this case happiness).  So when we are measuring user satisfaction or branding strength, a lot of what we want to impact through design is already pretty set in stone. 

The impact of coping strategies is also telling.  Many studies find that user performance is as much a factor of things like self-efficacy as they are with design ease of use.  Confidence makes you good at things.  And in a self-fulfilling cycle, being good at things makes you more confident.  This has incredible implications for training and education.  Using techniques like scaffolding, slowly increasing difficulty, quick feedback, user-customized speed, etc all create a virtuous cycle of confidence and performance.

And happiness.

Another finding is that varied thinking increases happiness.  Whether it is varied and fast (like brainstorming) or varied and slow (like daydreaming), varied thinking increases happiness.  On the other hand, focused thinking does the opposite, leading to depression in extreme cases.  Focused and slow (obsession) or focused and repetitive (panic) decrease happiness.

A related finding is that whether you are good at multi-tasking or not (which I have blogged about before), it makes you less happy to do it.  There is something soothing about focusing on one thing at a time (like meditation) that makes you happier.  Of course, keeping in mind what we learned in the previous study, we need to have enough variety in this one focused task to keep it interesting.  Boring isn’t good, but confusion isn’t either. 

So the ideal way to optimize your productivity is the concept of Flow – create a single task that someone can get totally involved in but is challenging enough to be interesting.  Of course, you already knew all of this.  These studies just add some details about how to do it.

Attention blink, emotion, and productivity

When we are multi-tasking, we aren't really doing two things at once.  We are really working on one activity, pausing it while keeping the current status in working memory, and then switching to the second activity.  And then the same thing when we switch back. 


So this depends on two abilities.  You have to have good working memory to keep all of the first activity stored while you are working on the second.  Otherwise, you have to review what you have forgotten every time you switch. This is no longer multi-tasking.  

The second ability you need is attention switching.  The attention blink is the delay that occurs when you switch from one activity to the other.   Your brain can't do it instantly.  There is some amount of time that your brain needs to refocus on the second activity.  This is just lost time.  It makes multi-tasking, even when you have good working memory, inefficient. 


So people with big working memory capacity and quick attention switching ability better at multi-tasking than people who have less.


Interestingly, a new study in the journal Neuropsychology reports that psychopaths have great attention switching ability.  Apparently, their brains are perfectly happy to put aside one activity and move to the next one.  I'm not sure why, but I thought this would be a fun fact to add.


The study I want to talk about in a little more detail is the effect of emotion.  A new study by Helen Tibboel  in Cognition and Emotion found that when the activity we are switching to is an emotional/arousing activity, we can switch faster.  This is related to the fight or flight instinct.  It also helps switch from texting to braking when someone cuts in front of you on the highway (assuming you notice it).  Our brains have somehow learned that emotional/arousing things are very important and worth dropping the previous task, even if it means forgetting more of it.  So it will take longer to go back to the original activity because we will have to review more of it to get back to where we left it.  


This has good points and bad points.  It means in the workplace, when there is a safety issue (which is usually high arousal) we can switch to it quickly and protect ourselves.  But it means that every time this happens, it takes longer to get back to work, so productivity takes a big hit.  It is well worth it for companies to invest in eliminating safety risks, even if no one is getting injured, just because of this impact on productivity.