I recently completed the Science of Well Being course on Coursera, taught by Professor Laurie Santos. This is sometimes called the Yale Happiness course. I did think there were many useful things in the course, but there was also some poor reasoning and misunderstanding of statistics. I think it is a decent course to take, but you should take the claims of the course with a grain of salt.

Money, Money, Money

One of the overall themes of the course is that the things that we think will make us happy don’t actually make us happy, or at least don’t make us as happy as we expect them to. A particular thing Professor Santos talks about is how we think money will make us happy. One of the pieces of evidence she uses for this is a LinkedIn survey of recent graduates showing that 61% wanted money from their jobs. She mentions this a lot to support the idea that what we think will make us happy is lots of money.

I was curious what the second choice was, and it was “balance” (I am assuming work/life balance), which 56% of graduates wanted. The fact that 61% + 56% is greater than 100% made me suspicious, so I read the fine print. The survey was actually the top five things graduates were looking for in job opportunities. So it’s possible that none of them thought money was the most import thing. They could have all thought it was the third most important thing. I think at best you could argue that 15% to 20% thought it was the most important thing. But then you have to consider that 39% of graduates didn’t even think it was one of the top five most important things about a job.

Furthermore, the question is “What do you want from a job opportunity?”, not “What do you think will make you happy?” I think the second choice of balance is important here. When I got my last job I looked for balance, because I expected to find my happiness elsewhere, and I wanted time to do that. And the graduates may just realize that they live in a capitalist society, and that means they need money to survive. Especially because recent graduates probably have a lot of student debt.

Professor Santos does provide other evidence for people thinking money will make them happy. She also gives evidence that it does make you happy, up to a point (about $75,000 a year, although later research disagrees with that). But it really bothered me how she emphasizes money as being the most important thing, and that she uses this survey to justify it.

Correlation vs. Causation

The vast majority of the studies mentioned in this course were showing a correlation between things, but not necessarily a causation from one to the other. Thinking that correlation implies causation is a cognitive bias, and it is a significant concern in statistics. It’s easy to show a correlation. It’s much harder to show causation. Professor Santos talks about correlation, and explains what it is to students who may not know. But when she explains it, she doesn’t talk about the causation issue at all. It only comes up in a Q&A session when a student asks about confounding factors. I really think it should have been emphasized more, and she should have been more careful about her language when talking about the correlations that have been found.

Her defense of using so much correlation is that there is so much evidence for the correlations that she believes the causation is there. But that’s not how it works. If you have a lot of evidence for a correlation, that just shows that the correlation is really there and that it is not an artifact of your data sample. It doesn’t mean the correlation indicates any sort of causation.

Just as she explained correlation, let me explain the causation problem. If you have a correlation between two things (lets call them A and B), then there are four possible causations going on. It could be that A is causing B, which is the common assumption. But it could also be that B is causing A. The causation could be going either way, if it is there at all. The third possibility is that a third effect (call it C) that causes both A and B. An example of this is the correlation between ice cream sales and drownings. It’s not that ice cream makes you drown, it’s that summer heat makes you want to eat more ice cream and swim more. Finally, there may be no causation at all. It may just be a coincidence. For example, organic food sales are correlated with autism diagnoses. But there’s no causal relationship between those two things.

Statistical Significance

Very little is reported on the statistical significance of the studies presented in the course. Most of the time that isn’t a problem, but I did find myself wondering about the statistical significance for studies reporting multiple values. The graphs from these studies generally did not have error bars or error bands. As a statistician with a keen interest in the graphical display of information, this is a red flag.

But some of the graphs did have error bars. Some of them even had error bars that weren’t obscured. One in particular was measuring how many of your “signature strengths” are part of your job, and how that relates to job satisfaction. This is a generally increasing graph, showing that the more you are using your signature strengths, the more you are going to be satisfied with your job. But there is a spike in the graph at four. Four signature strengths appears to be even better than five in terms of job satisfaction.

Professor Santos makes a big point out of this, and wonders why there is such a big spike, and why four signature strengths is the sweet spot. But remember that this graph has error bars. And the error bars for four contain both the point estimates for three and five. So even if the overall trend is statistically significant, the spike at four isn’t. It’s an artifact of the data sample. But Professor Santos doesn’t seem to get that.

The Fundamental Problem

I’m a guy with chronic depression (no, that’s not the fundamental problem). And there’s a science of depression, as well. This science informs the suggestions you see over an over again on the internet: exercise, sleep, meditation, and so on. But these don’t all work for me. Take exercise. I don’t like exercise, I don’t really enjoy it, and I don’t feel good after doing it. I do exercise, but I force myself to do it because it’s part of my long term goals. But I don’t disbelieve that exercise can help other people with their depression. Rather, I believe that each person’s depression is different, and that what helps each depressed person is different.  And after taking this course, I think happiness is the same way. What makes you happy isn’t necessarily what it going to make me happy.

A lot of the studies shown in this course measure happiness in different groups or before and after some experiment. But what is reported is the average increase. Even if some people get a lower happiness score afterwards or stay the same, you can still have an average increase. What I’d like to know is how many people in each study had a significant increase, and how many had a significant decrease, and how many had no significant change.

This was reinforced during the course. On thing that is mentioned is social contact with strangers, which studies have shown have a positive effect on happiness. People taking the bus were asked to talk to the person next to them, and there was a rise reported happiness. During this part of the course there was a fire pit social at my condo. I’m a bit of a hermit, and don’t really know the people in my condo, but I want to go to the social partially because of these studies. And the social interaction during this event was good. There was no extreme politics brought up, everyone was polite and congenial, people were respecting my opinion on things I’d studied. I literally cannot think of a single thing during the whole event that I would call bad. But when I got back to my condo afterward, I felt horrible.

Now, I am certainly an introvert with social anxiety. But during a Q&A session with her students she is asked about if the social interaction studies would apply to introverts. She replied that she believed they would. I think she is missing how personalized our happiness is, and how some of the techniques she is talking about may not help certain people, and may even have a negative effect.

The Good Stuff

Okay, I have been heavily criticizing this class. But there was still good stuff I learned in this class. And I still think this class is worth taking, at least if you take it with a healthy dose of skepticism.

One of the themes of the course is what she calls “The G.I. Joe Fallacy.” It’s based on the old G.I. Joe cartoon. The cartoon would end with some information meant to help kids out, ending with the tag line “knowing is half the battle!” She feels this is a fallacy, and the real battle is actually doing the stuff talked about, and implementing it in your life. The homework is mostly what she calls “rewirements,” where you practice what you learn in class to try to rewire your brain to think better about happiness.

This is something I have found in my life as well. I am a recovering drug addict. While I did not go through AA or NA, I know a lot of twelve-steppers, and I’ve talked to them about addiction. Bill Wilson, the founder of AA, said that “You can’t think yourself into right action, but you can act your way into right thinking.” You can’t just think about the changes you want to make in yourself and expect it to happen, you have to actually force yourself to do the change you want to make, and that experience will change how you think.

Another thing I liked was the idea of reference points. In talking about social media, Professor Santos says that it creates false reference points. People put forward their best life, which may not even be true, and comparing that to your own life makes your own life look sad. You need to have good reference points when judging yourself, not the unrealistic ones that advertising, entertainment, the news, and social media are presenting to us.

For me, I tend to have these depressive fantasies. I imagine a horrible life, often one that is ludicrously unrealistic, and get depressed because that life is horrible. But after learning about reference points I flipped the script on my depression. I used the fantasy as a reference point and compared it to my real life. That made me realize how good my life actually is. It works so well that when I do this I just start laughing maniacally.

One of the rewirements I worked on was gratitude. Every night I try to write down five things I am grateful for. Usually they are small things, which is good. It’s good to be able to appreciate the small things in life. I am sticking with this practice after the course. It helps me smile and laugh at the end of the day, which is often a tough time for my depression. And it is really me looking back on my day and finding the things that were good about it. I think it is giving me a better, more balanced view of my day than my depression wants to give me.

Again, I think each person’s happiness is different, which means each person is going to find different things of value in this course. I don’t think the things I found useful in the course are going to make you happier. And I am sure there are people out there who are going to have the exact opposite reaction to the social interaction activities that I did. But I think that if you take the course, and try the things that are mentioned in the course, you will find things that will make you happier.