Critical Thinking: How to Find Out What Really Works | Andre Klapper, PhD | Skillshare

Critical Thinking: How to Find Out What Really Works

Andre Klapper, PhD, Researcher, Neuroscientist, Psychologist

Critical Thinking: How to Find Out What Really Works

Andre Klapper, PhD, Researcher, Neuroscientist, Psychologist

Play Speed
  • 0.5x
  • 1x (Normal)
  • 1.25x
  • 1.5x
  • 2x
14 Lessons (1h 11m)
    • 1. Quick overview: what you will learn in this course

      2:07
    • 2. Introduction and first steps

      1:48
    • 3. The #1 reasoning fallacy

      9:11
    • 4. Examples of reasoning fallacies in our everyday life

      6:17
    • 5. The most powerful strategy to eliminate alternative explanations

      4:42
    • 6. Types of strategies

      5:06
    • 7. The second-best strategy to eliminate alternative explanations

      7:24
    • 8. A simple way to rule out coincidences

      3:13
    • 9. The elegant way to rule out coincidences

      5:06
    • 10. How to find out whether other things would work better

      4:39
    • 11. How to draw conclusions efficiently

      4:15
    • 12. The Complete Scientific Thinking Blueprint

      4:13
    • 13. Case Study: Will Starting a Business Make You Rich?

      12:00
    • 14. Conclusion

      0:30
  • --
  • Beginner level
  • Intermediate level
  • Advanced level
  • All levels
  • Beg/Int level
  • Int/Adv level

Community Generated

The level is determined by a majority opinion of students who have reviewed this class. The teacher's recommendation is shown until at least 5 student responses are collected.

300

Students

--

Projects

About This Class

Will starting a business make you rich?

Is getting up at 5 AM every morning the key to success?

Does money make people happy?

Life is full of questions and answering these questions is often key to getting what we want.

If we arrive at the wrong answers, we fail to get what we want.

If we arrive at the right answers, success is within our grasp.

Learn the strategies that scientists use to identify the right answers and use these strategies to make better life decisions.

What will you be able to DO after this course?

  • FORMULATE a question: will X get you Y?

  • IDENTIFY relevant evidence and INTERPRET it correctly.

  • AVOID reasoning fallacies that almost everybody falls prey to.

  • ELIMINATE alternative explanations systematically until you uncover the truth.

  • GATHER evidence yourself when no other evidence is available.

  • DRAW conclusions with scientific precision.

After this course, you will have a complete blueprint with simple step-by-step instructions that will enable you to make solid evidence-based decisions in every area of your life.

My first science classes shattered my world.

Afterwards, I saw reasoning errors everywhere.

Things I had believed for a long time turned out to be false.

Other things that I had rejected suddenly made a lot more sense.

It was shocking at times but also incredibly fascinating.

I enjoyed that everything became much clearer to me...

... and I started to feel a lot more confident in my opinions and decisions.

What I love about scientific thinking is that you can apply it everywhere.

Whether you want to be happier, more successful, wealthier, more productive, focussed, ... in all of these areas, scientific thinking can help to discard ineffective strategies and identify what truly works.

However...

This course is for thoughtful decision-makers.

If you prefer to make fast spontaneous decisions, you will not like this course.

If you do not like doing some detective work to uncover the truth, you will not like this course.

This course is for you if...

  • you want to ensure that your beliefs and decisions are correct

  • you want to avoid the mistakes that most people make

  • you are willing to put in the extra effort that most people skip

How does this course work?

Science classes usually involve years of intensive training.

However, you do not need years of training to reap most of its benefits.

This course is designed to provide you with 80% of the benefits for 1% of the effort.

You do not need to learn any equations, measurement theory, or anything like that.

Instead, you can master the key principles in only a few lectures.

Once you get started with this course, the lectures will guide you through the material in easily digestible steps.

At the end of the course, you will have a complete step-by-step blueprint that you can use to make better decisions in every area of your life.

I will also walk you through a case study so that you can practice your new skills.

Also, you can ask me questions anytime and can have my full support when you apply the blueprint to your own cases.

Meet Your Teacher

Teacher Profile Image

Andre Klapper, PhD

Researcher, Neuroscientist, Psychologist

Teacher

Psychology & Neuroscience researcher with more than 10+ years of training and experience.

Learning how our mind and brain work and conducting research on these topics has been incredibly fascinating for me and it definitely enriched my life.

My mission is to share my experience with other people and help them to get the most out of themselves.

I have courses on Psychology, Neuroscience, and research.

Why learn from me?

- 700+ enthusiastic reviews from people all over the world.

- Short and concise lectures - straight to the point without any unnecessary information.

- Simple and easy approach - complex ideas are broken into bite-sized chunks.

- Quality content. PhD, 10+ years of training and experience, scientific publica... See full profile

Class Ratings

Expectations Met?
  • Exceeded!
    0%
  • Yes
    0%
  • Somewhat
    0%
  • Not really
    0%
Reviews Archive

In October 2018, we updated our review system to improve the way we collect feedback. Below are the reviews written before that update.

Your creative journey starts here.

  • Unlimited access to every class
  • Supportive online creative community
  • Learn offline with Skillshare’s app

Why Join Skillshare?

Take award-winning Skillshare Original Classes

Each class has short lessons, hands-on projects

Your membership supports Skillshare teachers

Learn From Anywhere

Take classes on the go with the Skillshare app. Stream or download to watch on the plane, the subway, or wherever you learn best.

phone

Transcripts

1. Quick overview: what you will learn in this course: well, starting a business make you rich. Well, meditation make you happier. We're getting up at 5 a.m. Making more successful life is a journey full of questions on how you answer these questions can make the difference between succeeding at you goals. Oh, failing of them. Unfortunately, the world is full of doubt worthy opinions, sometimes coming from self proclaimed gurus who say that the X Y diet is the way to lose weight or starting your own business is the way to get rich. Or do the ABC work out. If you want to get ripped, or if you want to feel truly happy and content in your life, then you got to meditate every day. And obviously, if you want to be successful, then you gotta work 16 hours per day now to find out which of these claims are true and which are false, you can use scientific thinking. You can see the impact of signs at the society level. Science has taken us to the moon. It has cured illnesses that were not curable before, and it has given us immense power. Now, this course is about applying scientific thinking to your own life which means that it's not about giving you the answers to your questions, but instead, it's about teaching you how you can uncover the correct answers yourself for any question that you may have. Specifically, you're gonna learn about the common fallacies that almost everybody falls prey to and how you can avoid them. You will learn what evidence you can use to answer your questions, how to interpret that evidence correctly, how to rule out alternative explanations, and in case you cannot find any evidence. I will also show you how you can gather evidence yourself. Now, when you think about science, you may think off a lot off numbers and equations. But that's actually not what this course is about. This is a very practical hands on course in which he will learn in simple steps how you can sharpen your thinking, find the correct answers to your questions and make better decisions. All right, I'm excited to get started and see when the course 2. Introduction and first steps: Hi there. Welcome to this course. My name is Andre. I'm a trained scientist. And in this course I want to show you how you can answer questions such as Will starting a business make you rich? Will meditation help you to feel more happy and content? Or will getting up at five AM every morning help you to be more productive and hands successful? And these are not just interesting philosophical questions, but they're questions that have consequences. The answers you find for such questions can literally make the difference between ending up poorer, being unable to reach your goals or successful and getting the life that you want. Now, unfortunately, when people think about these types of questions, they very often fall prey to reasoning arrows. And as a result of that, they very often end up chasing after things that don't really get them anywhere. And so, in this course, I want to show you how you can sharpen you thinking, find the correct answers to the questions that you want to answer and make better life decisions. So how do we do that? And the first thing that I want to point out is that in all of these questions. There was the same general structure in all questions. It was about whether one variable X causes another variable. Why will this make me rich? Will that make me happy? Will this make me successful? In all of these cases, the question is basically about whether one variable X causes another variable. Why? And because all of these questions have the same general structure, we can use the same type of methodology toe answer all of these questions. And in this course I will teach you that methodology so that you can answer these questions with the highest possible precision. All right. And then the next lecture always show you the number one fellow see that almost everybody falls prey to and how you can avoid it. I see you in the next lecture. 3. The #1 reasoning fallacy: the number one reasoning fallacy that almost everybody falls prey to at some point is the causation. Fallacy. So what is the causation? Fallacy. Now let's say, for example, that you want to decide which university you want to go to. And let's say that you find out that students off University a later earn an average salary off 40 k and you find out that a university be students later earn an average salary off 80 k Well, in that case, most people would think that Oh, probably I should go to university, be university. Be seems to be the better university. Likewise supposed that you learn that people who don't meditate tend to be very unsatisfied with their life, while people who do meditate and to be very happy than most people think almost automatically that Oh, maybe I should meditate more. Maybe that would make me happier. Okay. Another example, which I've seen many times, by the way, is Theo example that people who say get up at 7 a.m. tend to have normal productivity while the people who get up at five AM especially all these CEOs who work like crazy. They tend to have high productivity and they tend to be more successful in life. And then many people come to the conclusion that okay, apparently, if I can get myself to wake up at 5 a.m. in the morning, my productivity will be higher and I will be more successful. And one more example when people hear that employees tend to have average wealth, whereas Andhra preneurs tend to be rich. Okay, Whether entrepreneurs really tend to be rich is actually the question, and I will walk you through that later in a case study at the end of this course. But let's say for the purpose of the example that entrepreneurs are richer on average than employees than many people would draw the conclusion that they should be entrepreneurs so that they can be rich as well. Okay, all the examples I just gave are examples off the causation fallacy in all of the examples I just gave. The conclusion was not necessarily correct. And I'm going to stick to the example off the employees versus the entrepreneurs now, just to illustrate to you what is going wrong here. So the conclusion that people draw here in this case is that starting a business. So becoming an entrepreneur causes people to build wealth, and the evidence that they're using here is what we call a correlation and their interpretation off. That correlation is that there's a causal relationship between the two variables, and that's not necessarily the right interpretation. So in this lecture, I want to show you how you interpret a correlation correctly. So first of all, what is a correlation? A correlation is a statistical relationship between two variables. Where if one variable changes, the other systematically changes with that variable. So a very simple example would be that if one variable increases, the other one increases a swell and have one variable decreases, the other one decreases as well. So, for example, ask people go from not having started the business to starting a business. Their wealth increases, but it can also be the other way around. Another example of a correlation would be that if one variable increases, the other one decreases and vice versa. If the variable decreases, the other one increases. That's also an example of a correlation and what people usually automatically think when they see such a correlation. Is that okay? So starting a business makes people rich or it makes them poor. But in any case, they think that there must be a cause of relationship between the two. Now, as I already said, that's not necessarily the right interpretation. Instead, they're almost always three explanations possible for a correlation. And to make sure that you draw the right conclusions, you need to consider all three explanations. Okay, so suppose that we have two variables. Accent. Why? And they're coronated. Then what other three explanations? The first explanation is that X causes why, If X causes, Why, then whenever X increases, why will increase as well? And so we will see a correlation. However, another possible explanation is that why causes X in that case, whenever why increases? Axel increases well and again we see a correlation and then finally, explanation. Number three is that there's absolutely no causal relationship between X and y at all. But instead there's 1/3 variable. We could call it sets that causes both acts and why. And in that case, whenever Zet increases, axe will increase and why will increase. And so we will see a correlation between X and Y, even though there's absolutely no causal relationship between acts and why. Okay, so these are the three explanations. Now let's apply everything that I just taught you to our example. So let's say that we observe that entrepreneurs are rich. Why am Please only have an average wealth, and the first thing we need to realize is that that is just a correlation. It just means that as one variable increases, for example, as you go from employees to entrepreneur, the other variable systematically various with it. For example, we go from average wealth to being rich. So now the next thing we need to do is we need to walk through all three possible explanations for that correlation, and the first explanation is that starting a business causes people to build wealth. However, another possible explanation is that wealth causes people to start a business, so it could be that as people get rich, they don't really want to have a job anymore. But instead, they're not afraid anymore to take the risk off starting their own business because they're already wealthy, so they don't really have anything to lose. And likewise, when people get poor, they may be less likely to start a business and more likely to take a job instead. So an equally plausible explanation for the correlation between starting a business and wealth is that wealth causes people to start a business. Okay, and finally, explanation Number three is that there's no causal relationship between starting a business and wealth at all. But instead there might be 1/3 variable, and an example of such 1/3 variable could be how ambitious a person is. If a person is very ambitious, then it could be that that makes the person more likely to start a business. And it could also be that that also makes the person more likely to build well. But the reason why the person builds well may not be that the person started the business, but just because the person is very ambitious at whatever that person is doing. So even if that person hadn't started a business, that person would probably have built well, just because that person is very ambitious. So that's the third explanation. There's no causal relationship between starting a business and wealth at all, but instead they're other variables that cause both. So that is how you interpret a correlation now what if the correlation is not positive, but it is negative. So let's say it's the other way around. And it's not that the entrepreneurs are rich, but the employees are rich. So in that case we have a negative correlation between starting a business and wealth. Now, for this negative correlation, everything works in exactly the same way, with the only difference that you first need to flip the second variable here. So rather than thinking off the second variable as wealth, you can think of it as being poor. And that way the levels off the variable are flipped. Now, when you are high on this variable your poor and when you're low on this variable, you're rich. While on the variable wealth when you were high, you were rich and when he will low, you were poor. And with this simple flip, the negative correlation becomes a positive correlation. Now there's a positive correlation between starting a business and being poor, and now it can interpret this in exactly the same way as I just showed you. So the first explanation for that correlation is that starting a business causes people to get poor. The second explanation is that getting poor causes people to start a business, and the third explanation is that there's 1/3 variable that causes both, such as, for example, risk taking. So it could be that people were very risk taking. So basically, gamblers are more likely to start a business, but they're also more likely to get poor. But the reason why they get poor may not be that they started the business, but just because they do a lot of other risky things that cost them to get poor. Okay, so the key lessons off this lecture are a correlation is a relationship between two variables X and Y. People tend to take a correlation as evidence that acts causes why. But that is only one possible explanation, and the three explanations that you need to consider are first that X causes. Why, Second, that why causes X and third that set So another variable causes both X and y, and you can find the causation fallacy that people immediately jumped Explanation one pretty much everywhere in every area of people's life. And in the next lecture, I want to show you some examples where you can see that 4. Examples of reasoning fallacies in our everyday life: Okay, so now you know about the causation fallacy. And the tricky thing about this fallacy is that is really hiding everywhere. And so in this lecture, I want to train you in detecting the causation fallacy so that you can avoid it. OK, so let's walk through a couple of examples and one example could be that you ask yourself which university is better and we already had the example off University A versus University Be where? After university A people earn an average salary off 40 k whereas after university, be people tend to earn an average salary off 80 cake. And the first thing that is important to notice here is that that is just a correlation. All that you're seeing here is that if the variable university changes from university a university B so does the ever salary off these people. But the question is still what does this coronation meat? And as you learned in the last lecture there, three explanations we need to consider examination One is that the university matters for the salary. University B will cause you to have a higher salary later than university A. Then we also need to consider explanation to, which is that the salary causes the university. But in this case, we're in luck because the salary happens out of the university and assuming that something cannot cause something backwards in time. So the future salary cannot affect what you do in the past. When you choose the university, then this explanation is not possible. So in this case, we can rule explanation to out. Nevertheless, there still explanation three that there's 1/3 variable such as, for example, ambition. And it could be that people were more ambitious, are more likely to choose university, be over university A and at the same time, being ambitious may cause people to earn a higher salary. And so what may be really happening here is not of the university is better at training you so that he can earn a higher salary later, but that the university is just better at attracting ambitious people and that even if these people would go to another university, they would still earn a higher salary. As overall, we don't know whether the highest Saturday is caused by the university. Okay, let's have a look at another example. Does money make people happier Let's say you ask yourself that question and then you look at the world and let's just say hypothetically that you see that people who have not that much money or an average wealth tend to be reasonably happy, while people who are very wealthy tend to be a bit more happy. Let's just say hypothetically that that's what you see. Then again, the first step is to realize that that is just a correlation. Always see here is that if the variable wealth changes from poor to rich than the happiness variable changes from 7 to 8 and now the next thing we need to do is we need to ask yourself, What does this mean? On one possible explanation is indeed that money causes happiness. But another possible explanation is that happiness causes money. So, for example, it could be that people who are more happy are more optimistic and therefore more likely to take the risks that they need to take to earn a lot of money. Or it could be that happy people are more likable, and that also makes them more likely to receive a promotion from their boss, for example, which causes them to earn money, so it could equally well be that happiness causes money. And then, of course, there's also explanation three, which is that there's 1/3 variable that causes both. So it could be that if you're a person who is very often successful than that causes you to earn a lot of money, and it could also make you more happy and not necessarily because of the money that you earn, but just because it's nice if you reach the goals that you set yourself so again, we cannot be sure at all here that money causes happiness. That's just one of many explanations. Okay, one more example, and this is one that I've seen several times. Does getting up at 5 a.m. make people more productive? So let's say again that you look at the world and you see that there. People who get up at a normal time say 7 a.m. And there are people who push themselves to get up very early, such as 5 a.m. and the people who get up at a normal time 10 to have normal port activity, while the people who get out very early tend to have very high productivity. If you see that, then you could think that getting up at 5 a.m. is a very good strategy to get higher productivity, right. But again, we need to realize that that is just a correlation. It just means that if we move from getting up at a normal time to getting about early time , then there's a change in the variable productivity. And now the question is, what does that correlation meat and again? Explanation Number one is that the time you wake up influences how productive you are on that date, then the second explanation we need to consider is whether productivity could cause the wake up time. And here we are in luck again because the productivity happens after the wake up time. First you wake up early or not so early, and then you either productive during that day or not, and assuming that a variable couldn't cause something backwards in time. This explanation doesn't work, so we can rule this one out in this case. But then nevertheless, we still have the third variable explanation. And it could be, for example, that people who are very energetic tend to wake up very early, and they tend to be very productive. And it could be that the reason why they're so productive is not because they wake up so early, but just because they have a lot of energy and I can't speak for you. But I personally tried getting up at 5 a.m. In the morning, once in my life. And I think on the first day and maybe also the second day, my productivity waas higher. But then after that, I just felt more more tired every day. And I think my productivity actually decreased. So does the wake up time matter? So much for productivity. That is actually something that we don't know based on this correlation. Okay. So as you can see, the correlation fallacy is really hiding everywhere. And that leads to the next question. Which is how can we solve this? And that's what I'm gonna cover with you in the next lecture. So I see you in the next section 5. The most powerful strategy to eliminate alternative explanations: all right, So now that you're a little bit trained in seeing these different explanations for correlations that we may observe in the world now the question is, how can we rule out the alternative explanations and figure out whether X causes why and their several ways to do this, each with their advantages and disadvantages? And in this lecture, I'm going to introduce you to one of the most powerful solutions, which is an experiment. Now, when you think of an experiment than you may be thinking about something like this, but this is actually not what an experiment is. This is just how an experiment happens to look like in some branches off signs. But an experiment is something much more simple. So what's an experiment? An experiment is when the levels off the variable acts are randomized. So we had these two variables accident Why? And both variables have levels. For example, that can be either high or low. And when the levels off the first variable are randomised, for example, through a coin flip, and we call that an experiment and I know that this sounds very abstract, So let me walk you through an example let's say that you want to know whether money makes people happy. In that case, we have an experiment if the levels off the variable wealth are randomly determined by a conflict, So to illustrate the logic of this, let's say that your gods and you can now determine through a coin flip whether people are poor or whether they're rich. So for every person in the world, you flip a coin, and if the coin says heads, you make that person poor. And if the coin says tail, then you make that person rich. And now let's say that in this very hypothetical scenario, you find that the poor people are very unhappy and the rich people are very happy, which is a correlation, right is a correlation between wealth and happiness. However, the interpretation of that correlation changes completely if you find it in an experiment. So let's walk through this so the two variables you have our wealth and happiness, and now the first variable is randomized through a corn flip. So in that scenario, could the correlation be explained by wealth causing happiness? Absolutely. That's definitely a possible explanation. But what about the reverse effect? Could happiness have caused the wealth? Well, No, because you determined the well through a coin flip and your coin doesn't care about the happiness of a person. No matter whether the person is unhappy or happy, the corn flip makes it equally likely for a happy person at an unhappy person to become wealthy in your experiment. So here. This is not a plausible explanation. Then what about explanation Number three? That there's 1/3 variable influencing both variables, for example, that success in reaching your goals makes people wealthy, and it makes them all so happy. However again, this car's a relationship here, from success to wealth doesn't really work because our coin doesn't care how successful a person is in reaching his or her goals. The coin makes it equally likely that people who are unsuccessful in reaching the goals and people were successful in reaching their goals become wealthy. And because of that explanation, three also doesn't work. So the only explanation that has left in this scenario is explanation Number one, that wealth causes happiness. So in other words, if the levels off the first variable a randomized and you find a correlation between the two variables. Then you can conclude that the first variable cost the second variable. Now you might be thinking grades, but how am I supposed to do this? But sometimes, if you just look around in the world, you confined naturally occurring experiments such as, for example, the lottery. The lottery basically randomly divides its participants into the participants who don't get money, and the participants will become incredibly rich. And so what we can do is simply is we can have a look whether people win the lottery tend to be more happy than people who don't win the lottery. And the finding here is that, temporarily, yes, people who win the lottery are a little bit more happy. But in the long run, it doesn't seem to matter. So it seems to be that money can make you happy temporarily, but maybe not in the long run. All right. To sum it up, the key lessons of this lecture are you can rule out explanation two and three through an experiment, and in an experiment, the levels off the variable X are randomised, for example, through a corn flip, and you can run experiments yourself, which I'm going to show you in the next lecture. But sometimes you can also find naturally occurring experiments such as, for example, the lottery. All right, and then the next lecture. Hubble, zoom further into this with you. So I see you in the next lecture. 6. Types of strategies: in this lecture, I'm going to introduce you to two different types of experiments. So in the last lecture, you learned that an experiment is about random izing the levels off the first variable. And these levels are always about some kind of comparison. For example, university, a Verses, university B or rich people versus poor people while entrepreneurs versus employees and in this lecture wanted zoom further into this and show you that there are actually two different types of comparisons that you confined here. One comparison is a between subject comparison, and the other comparison is a within subject comparison. So a between subject comparison is when you have two groups and you compare these groups to each other, and an example of that would be if you compare students off different universities. But you can also look at a within subject comparison, and in that case, you look at a comparison. Over time you have two or more measurement points per person, and you compare these measurement points to each other. What happened at time, one compared to time to and can do that with several people or even just with one person, so an example of That would be if you compare weeks in which you meditated, two weeks in which he didn't meditate to figure out whether meditation is helping you to feel better or get more focused, for example, and in both of these cases, you need an experiment to be able to draw a clear conclusion. So what is an experiment in these two cases, while in an experiment, you either randomize who goes to which group in a between subject design or when what happens in a within subject design? So let's go through one example. Let's say that you want to know whether getting up at 5 a.m. makes people more productive than one type of experiment you could run or search for would be an experiment where you divide people into a group that has to get up at 7 a.m. In the morning, every time and a group who has to get up at 5 a.m. In the morning every time. And then you can check which of these two groups is more productive, and this would be an experiment. If the dividing into the two groups happens, add random, for example, through a coin flip But if you don't want to bother your friends by telling them when they have to get up, which I can understand very well that he can also run an experiment on yourself. So what you can do is you can have days on which you get up at five AM in the morning, and you can have days on which you get up at 7 a.m. In the morning. And he could turn that into an experiment by deciding each day in advance through a coin flip, whether the next day is gonna be a 5 a.m. day or a 7 a.m. day. So on the first day you may be able to steep longer. Then you have to get up early. Then the conflict says you have to get up early again. Then you can Steve longer than you have to get up early that he can sleep longer and then he conceived longer again and so on and so on. And if what you do on each day is randomized, then that's also an experiment. So let's say that you find that on days on which you wake up early, you're more productive than What could that mean? One thing it could mean is that the wake up time influence your productivity. But normally we would also have to consider the possibility that the productivity influenced the wake up time. Now, we already said that that isn't really possible in this particular case because the per activity happens after the wake up time. But in an experiment, it is impossible anyways, because the wake up time is completely determined by the corn flip, so it cannot be influenced by something else. So this explanation definitely doesn't work. And then the third explanation that we normally need to consider is that there's 1/3 variable that causes both. So, for example, if you wouldn't randomize, then it could be that after nights on which your sleep is really off a high quality and really deep, you automatically wake up early and you're more productive during the day. However, if you determine when you wake up by a coin flip, then again, the conflict doesn't care about your sleep quality, the coin tells. He went to wake up, regardless of whether you feel super sleepy in the morning or actually energetic and ready to start the day, and therefore this cause of relationship doesn't really work again. And therefore the third variable explanation doesn't work. So if you find in this scenario that you are more productive if you get up at 5 a.m. in the morning, then that means that getting up at 5 a.m. in the morning is a good way for you to get more productive. Okay, to sum it up, the key lessons of this lecture are the levels off the variable X can either be between or a within subject comparison, and the between subject comparison is an experiment. If the people are randomly assigned to the two levels off acts and the within subject comparison is an experiment, if the time points are randomly assigned to the two levels off acts. An experiment is the most powerful methods to rule out alternative explanations. But it's not always feasible. For example, you can tell people to wish university they should go based on a conflict. They just won't do that. Therefore, in the next lecture, I will show you a second method to rule out alternative explanations I see in the next election 7. The second-best strategy to eliminate alternative explanations: if you cannot run an experiment, and if you can't find an experiment than the next best alternative is a mixed design. A mix design is when you combine a within subject design with the between subjects. Design. So you have two groups of people, but you also have at least two measurement points in time. So let's go through an example. Let's say that you want to know whether meditation makes people happier. Now suppose that you don't really want to do the effort to run an experiment on yourself. But you also can't find any situation in the world where people are randomly divided into meditators and non meditators. Then the next best thing you can do is a mix design. Let's say that your friends decide to start a new activity every Saturday evening, and let's say that a part of your friends decide that they have a movie night while another part of your friends decide that the take meditating classes on Saturday night than in a pure between subject design. You would look at how happy people feel after a while of doing these two activities, and then you may find that the people who didn't start meditating. Do not feel as happy as the people who did start meditating. Now, in a mix design, we add one more measurement point. And in this case, that could be the starting happiness off these two groups of people before they start the new activity. And let's say that the starting happiness off the non meditators Waas a six out of 10 while the starting happiness off the meditators was an eight out of 10. And then after these activities, we see a six out of 10 for the non meditators and an eight out of 10 for the meditators. When that situation, we can see that even though the meditators are happier than the non meditators, the happiness didn't change as a result of any of these activities. The non meditators started with the six and the ended with a six and the meditator started with an AIDS and they ended within eight. So now we can see that the meditation actually didn't help. Instead, this result would be more in line with the explanation that happiness makes people more likely to meditate because here we can see that the happy people went towards the meditation class, and that's the power of a mixed design in the mix design, where you can see these patterns on what you want to see. To conclude that meditation actually does improve your happiness is that the change in happiness over time is larger in the meditation group compared to the non meditation group . So here, for example, the happiness of the meditators increases by one point, while the happiness of the non meditations doesn't change at all. So this result would suggest that meditation causes happiness. However, there are still alternative explanations. So it could be, for example, that the people who went to the meditation classes are people who are very good at getting joy out of new activities. And it could be that if you would have told the meditators to go to the movie nights, they would also have increased in happiness by one point simply because they are very good at enjoying new activities. So even in a mixed design, we still have alternative explanations. Nevertheless, I hope you can also see that not all the alternative explanations that we had to consider before still work and the alternative explanations that still do work are a lot more complicated and not as plausible as the explanations we had to deal with before. So in a mixed design you still have alternative explanations. But they're not as serious as in a pure between subject design or a pure within subject design. Okay, let's walk through another example. Let's say that you wonder which university is better, and this is a tricky situation because you can tell people to wish university they should go, and therefore it's very difficult here to run an experiment. So in this situation, a mixed design is again your best alternative. So what you want to do is you don't just want to look at how capable people are after going to the university. Say we score that as a six out of 10 for the students of University A and an eight out of 10 for the students of University B. But you also want to look at how capable they were already before they went to the university. So when we take that into account than he and this example, we can see that basically the students off neither university improve at University A. They started with an ability level off six, and they end up with a six. And at university be they start with an ability level off eight and they end up with an ability level of eight. So the only difference between these two university seems to be in this example that university B is better at attracting highly capable students. And that's generally a problem, because once the university has a good reputation, that means that good students will go to that university and then automatically. There will be a correlation between the success people have later in life and that university. So what you want to see ideally, is not just that the students of one university are doing better later. But he wanted to see that the students come in relatively incapable and they come out relatively capable, which is something you can on Lee Seon, a mix design. Nevertheless, even if you find that there are still alternative explanations, it could be, for example, that the students who went to university be in this particular case are better learners, and that's the only reason why they improved so much. But if they're better learners than the question is, why weren't they already more capable to begin with. So as you can see here, even though there are still alternative explanations, they are not as plausible as the alternative explanations we had to deal with in a pure between subject design or appear within subject design. Okay, now let's go through an overview off the options that you have. When you want to find out whether one variable is causing another variable, the best option is always to run an experiment on yourself. An experiment immediately eliminates all alternative explanations that we covered at the beginning. And if you run it on yourself and you know that the results will apply to you, the next best option is to look at an experiment on other people. Usually running such an experiment is not really feasible. But sometimes you can find an experiment naturally occurring in the world such as, for example, the lottery. Then the next best option is a mix design. If you cannot find the case where the levels off the first variable are randomised, then at least make sure that you get as much information as possible by not just on Lee comparing groups but also seeing how they change over time, all right, and if even that is not available than the last resort you have, is to make an educated guess based on a correlation. Usually it's possible to find at least some correlation in the world. And if that correlation is neither found in an experiment, nor makes design than the best you can do is simply to go one by one through all the possible explanations and make an educated guess how likely it is that the first variable is causing the second variable. All right, so the key lessons of this lecture are if an experiment is not feasible and mixed design is a good alternative, a mix design minimize alternative explanations without ruling them fully out. And if a mixed design isn't possible either than your last resort is an educated guess based on a correlation, all right, that's his with sexual, and I see you in the next one 8. A simple way to rule out coincidences: okay, And this lecture will show you how you deal with coincidences. So what do I mean by coincidence? So let's say that one day you get up at your normal 7 a.m. time, and then one day later, you get up at five AM and you find that on the second day you're more productive. And let's say that you determined both through a coin flip. So it's an experiment. Then can you really conclude that getting up at 5 a.m. in the morning cause it to be more productive? Truth be told, not exactly because it might just be a coincidence, and he is a way to think about it. If you have done the same on both days, for example, if on both days you get up at 7 a.m. you still would probably be more productive on one day compared to the other, right, because without any clear reasons, sometimes we have more productive days. Sometimes we have less productive days, and therefore, even if you do the same on both days, on one day you will have higher productivity. So another question we need to ask ourselves is that if we find that after getting up at five AM, our productivity is higher. Is that because we get up at 5 a.m. or is it just a coincidence? And there several solutions to rule coincidences out. The first solution is sample size. If we don't just measure on two days, but on many days, then it becomes much less likely that on the days on which we get up at five AM, we just happen to be more productive. And as a general rule, the larger your sample is. So the more measurement points you have, the more you can be sure that is not a coincidence. And what exactly the sample size is depends a little bit on your design. In a between subject designed, the sample size is the number off people you have in each group and in a within subject design, it can be two things. It can either be the number off people you look at, and it can also be the number off measurement points you look at. So if you test on yourself, the sample size would be how often you test over time and an example where sample size is really helping us out is in the lottery example because there's so many people taking the lottery, and over time they have been so many people who won the lottery that we can be really sure that the differences between lottery winners and lottery losers are not just coincidences. So that's the first solution. Sample size. And you might wonder, what's the big sample size and what's a small sample size? And technically, it's quite arbitrary. The rule is simply, the more the better. But just to give you some orientation, I will give you the numbers that I would use in my private life to determine whether I have a large or small sample size. So a small sample size for me would be 20 observations, and that doesn't mean that you cannot go lower. It's just that then the risk gets really high that your results are coincidence and a large sample size would be 100 observations or more. It is more than 100 observations that he can be quite sure that the things you observe are not just coincidence. Okay, so that's the first solution sample size. But sample size isn't always the most feasible solution, and therefore, in the next lecture, I will show you a second solution 9. The elegant way to rule out coincidences: another way to rule out coincidence is is to assess the unexplained variance. So what does that mean? Let's say that you tested over 10 days, whether getting up at 5 a.m. in the morning makes you more productive or getting up at 7 a.m. in the morning. And let's say that every day you gave yourself a score from, say, 1 to 10 on how productive you were on that day. Then he is what the results could look like for the 7 a.m. Day. So on this day, your productivity was relatively low on this day here it was a bit higher then of it lower again, then a bit higher and then a bit lower again. And let's say that these other results for the 5 a.m. days So on day one you were very productive. Then your productivity dropped a little bits and then it increased the bids and so on and so on. Now, when I look at these results, I actually get quite convinced by them, even though the sample size is really small and the reason has to do with unexplained variance. So let's break this down there. Two things going on in this data sets first. There is the average difference between the 7 a.m. days and the five AM days. So the average off the 7 a.m. days would be roughly here while the average off the five AM days would be roughly here. And the difference between that is what we think is the effect off the waking up time. Waking up at five AM changed the productivity from here to here. We could call this the explained variance because the variants we can explain through the wake up time. Aside from that, we also have unexplained barrettes. So, for example, why was my productivity lower on this day compared to that day and then lower again on this day? Compared to that day, we don't really know until we could call this unexplained variance. So the unexplained variance is the variants around the mean for which we have no explanation. And we have that for the 7 a.m. days, and we also have that for the five AM days Now, what the unexplained variance basically tells you is how big the effect of coincidences. So here we see that just by coincidence, the productivity goes up and down and up and down in roughly this rich and he is the same. We see that just by coincidence, the productivity goes down and up and down and up roughly in this rich. And what we can see here is that on 7 a.m. days is really unlikely that our productivity goes up until here and vice versa. On five AM days, it's really unlikely that the productivity goes down to here, and from this pattern we conclude that most likely the difference between the means here is not just a coincidence, because coincidences aren't that big based on the unexplained variance that we see. And if you find that pattern, then a small sample is enough. Now compare that to the following situation. Let's say that we have the exact same difference between the means, but the unexplained variance is a lot larger. So therefore, with this amount of unexplained variance, it's quite likely that the difference between the red means is Jessica incidents. So in this situation, the only way to find out whether that difference is not just a coincidence is to collect a larger sample. So the simple rule is that if the effect of the first variable sticks out compared to unexplained variants than a small sample sizes enough. And if it doesn't, then you need a large sample size. Okay, now let me give you some examples off unexplained variance to make clear what unexplained variances in different situations. So let's say that you want to know which university is better then. In that case, unexplained variance is how much students within each university differ from each other. While the explained variance would be how much students between universities differ from each other, let's say that you want to know whether meditation makes you happier. In that case, the unexplained variance is how much your happiness fluctuates within time periods in which you meditate or time periods in which you don't notice it. Basically, it's how much your happiness varies between days on which it did exactly the same thing. Or let's say you want to know whether entrepreneurship makes people rich. In that case, the unexplained variance is how much wealth varies within entrepreneurs and within employees. So how much? Well, various between people who do the same either being entrepreneurs or employees. All right, so the key lessons of this lecture are the first way to rule out coincidence is is to have a large sample. However, small samples can be sufficient if there is not much unexplained parents. And more specifically, if the effect of your variable does not stick out compared to unexplained variance, then the effect may not be real, but just a coincidence. But if the effect off your variable dust stick out, that is probably really even in a small sample. All right, that's it with the sexual and I see in the next one. 10. How to find out whether other things would work better: in this lecture, we're gonna change the question a little bit. And rather than asking, does X cause why, we ask, does x matter enough? So, for example, rather than asking dust, getting up at five AM make you more productive, you could ask, Are there other things that matter more than getting up at five AM Because honestly, getting up at five AM is no fun. So if you do it, then it better be worth it. And you want to make sure that there aren't any better solutions that you haven't considered yet, and you can figure this out using the techniques that you already know. So to illustrate this, let's say again that you test it on yourself whether getting up at 5 a.m. in the morning make you more productive and just to make the results a little bit less cluttered. I have assumed here that the days are not randomized so that everything is next to each other. And here we see the average productivity on days on which you got up at seven AM and here we see the average productivity on days on which you got up at 5 a.m. And we can see that it's a little bit higher, but at the same time there's a lot off unexplained variants, and our effect here doesn't really stick out compared to that unexplained various right. So in that situation we would be worried that the difference is just tickle incidents. But there's more than we can learn from this pattern, because essentially, what unexplained variance is is a measure off the effect off the variables that we haven't considered yet. So let me give you an example. It could be that on the days with high productivity, you regularly took a coffee break and that help you to stay energetic and to stay productive while on the unproductive days you didn't take any coffee break. But she just kept working and kept working until he was so exhausted that you were just staring at your screen without getting anything done on what the pattern that we see here basically tells us is that the coffee break matters a lot mawr compared to when you get up in the morning. Now compare that to the following pattern. Here we have a pattern where the difference between the 5 a.m. days and the 7 a.m. days is relatively big, and the unexplained variance is relatively low, right, and here the same could be happening. It could be that on the higher days you took the coffee breaks while on the lower days, he just kept working. But here the conclusion would be the opposite. Here we would conclude that whether you take the coffee breaks or not hardly matters for your productivity, while the time I wish you get up matters a lot. So by comparing the effect off the first variable to the unexplained variance, you can get an idea how effective that variable is compared to other variables that you haven't tested yet. And if you find a pattern where the effect is very small compared to the unexplained variance, then rather than obsessing about the question whether that difference is a coincidence or not, it makes more sense to ask yourself what is causing that unexplained variance. Why use of productive on this day and this day and this day and this day compared to this day and this day and this day now supposed to do you think about it and you compare the productive days to the unproductive days and you realize that on the productive days you took more coffee breaks. Then does that mean that you should take more coffee breaks and he have to be careful? Because again, this is just a correlation. And unless you tested this in an experiment or at least you have a mix design, it probably doesn't mean so much. So here again, we need to think about our three explanations. It could be that the coffee break costed to be more productive. Or it could be that on days on which you're very productive, you have more time for coffee breaks or the coronation could be due to 1/3 variable. For example, it could be that on days on which you wake up early, you're so tired that you take more coffee breaks, but you also have more time to work and so you're more productive. So if you find that on the more productive days you took more coffee breaks, be aware that it is just a correlation and that the only way to know for sure that the coffee breaks actually cause you to be more productive is an experiment. Okay, so the key lessons off this lecture are if there is a lot of unexplained variance than other variables, probably matter mawr than their variable that you're looking at. And in that case, it's worth asking what other variable could have caused that unexplained variance. And once you have an idea, be sure that you test it ideally in an experiment. 11. How to draw conclusions efficiently: Okay, So now you have seen what kind of things you need to look at and what kind of things you need to be aware off when you draw conclusions and try to figure out whether acts can get you Why? And as you probably realized, it's a lot more complicated than people typically think. So this is a good moment to ask ourselves. How can we be efficient about this or, more precisely, are really all lessons of this course always necessary? For example, what happens if you ignore alternative explanations for correlations? Or what happens if you ignore unexplained variance? And the short answer is, of course, that he will make more errors. But he will also make less Eros because they're actually two types off errors that we can make when we draw conclusions. The first type is what we call false positives, which is believing in something that is not true. For example, if you believe in ghosts, even though goes stone really exist, then that would be a false positive. The other type of error are false negatives. False negatives is when you reject something, even though it is true. On an example of that is what happens in flat Earth society. The people in flat Earth Society reject the idea that the earth is a sphere. Even though we know with very high certainty that the earth is a sphere and you can draw conclusions in two ways, you can either be conservative, which means that you don't easily believe things, and in that case you minimize false positives. So you minimize the possibility that the things you believe are not true and the other strategy is to be liberal, which means that you easily believe in things, and in that case, you're minimizing false negatives. So you prevent that you reject ideas that are actually true. So both of these strategies have their advantages and disadvantages. If you're very conservative, then you're basically a neigh Sayer who will miss a lot of opportunities. But if you're too liberal than you are, a yes, Sayer and you start believing in things that totally on true, such as believing that you can predict the future by reading carts. So both of these have their strength and they're weaknesses. And the question that you need to ask yourself for your particular question is which type of error is more costly to you. And by answering their question, you can figure out how efficient you can be about answering your question. So let's go through two examples. Let's say that Peter wonders whether the productivity ad for $1 will help him to get more things done. Well, in that case, what would have false positive be? It would be that he buys the app so he wastes $1 he doesn't get more productive. Is that really so bad? Not really right. On the other hand, a false negative would be that he rejects the app, even though it would help him. And in that case, his productivity will stay low. And that may actually be a lot more costly to him. So in this case, it would make absolute sense to take a leap of faith. And just by the up, even at the risk that this turns out to be a false positive. Now compare that to the second example. Let's say that Maria wonders whether it's worth to risk all of her savings to start our own business. So what would a false positive B In this case, it means that she loses all of her savings and her business doesn't work. And what would be a false negative in this case? It would be that she keeps her savings. But she fails to start a business that otherwise would have been successful. And he could argue in this case that losing all of your savings is really costly and that false positives in this case are actually more costly. So for Maria, it would make a lot of sense to take all of the things that I covered in this course very, very seriously when assessing the question, whether starting a business can give her what she wants to have. So to sum it up, the key lessons of this lecture are whether you need to apply all lessons from this course depends on which type of error is more costly to you. If false negatives are more costly than take a leap of faith, However, if false positives are more costly than be sure that you apply all or at least most off the lessons off this course. All right, that's it for this lecture. And I see you in the next one 12. The Complete Scientific Thinking Blueprint: Okay, so now you know the principles off scientific thinking. And in this lecture, I'm gonna put these principles together for you into a step by step blueprint. So the first step is always to formulate a causal hypothesis off the form that X causes. Why does money make people happy? Does entrepreneurship make people rich? So, in other words, dust the first variable X cost, the second variable. Why, then, once you have that that the next step is to look at the costs off the two types of arrows. What if you accept the hypothesis and it turns out to be false, which would be a false positive? And what if you reject the up offices and it's true, which would be a false negative? And if false positives are more costly than I would recommend that you go through all or at least most off the subsequent steps off this blueprint. And if false negatives are more costly, then it can make sense to just follow only a few off the steps off this blueprint and to just take a leap of faith. Then the next step off the blueprint would be to select a test and most of the time, the ideal test would be an experiment on yourself. But if that's not possible or if you don't want to do it, then you can also look at an experiment on other people. Then the next best choice is a mixed design on other people. So design, in which you look at both two groups and the change over time and then finally, your last resort is an educated guess based on a correlation. Then, once you have selected a test and once you see some results, then the next step will be to compare the effect off the first variable to the unexplained variance. And if the effect of the first variable sticks out from the unexplained variance, and that could mean that the variable is a major cause off the second variable. And if the effect of the first variable is overshadowed by unexplained variance, then that suggests that the variable you're looking at is probably not very important. Now, if you still want to know whether that variable is a cause off the second variable, even though it doesn't seem very important, then you can still figure that out. If you have a large sample size. But in most cases, if the unexplained variance is much larger than the effect of your first variable, it's probably not worth looking at it further. And instead it makes sense to ask what other variable could explain that variants that still isn't explained? So we have this example where you looked at your productivity on days on which you get up at 7 a.m. In the morning and days. Always you got up at 5 a.m. In the morning. And if the effect of the first variable so the average difference here sticks out from the unexplained variance, then that means that the time in which you wake up is a major correlate of productivity compared to other possible correlates. And in that case, the final step would be to assess the three possible explanations for this correlation. So one explanation could be that X causes y. Another explanation could be that why causes X? And finally, the third explanation could be that 1/3 variables that causes both X and Y and which of these explanations you need to consider depends on the test that you have chosen earlier in an experiment. It's very easy. It can only be explanation number one in the mix design. It can be several explanations, but explanation one is one off the more plausible explanations. And if you have neither an experiment nor mix design, then you need to accept that all of these explanations can be true. And you just need to do an educated guess, okay? And the other situation that he could have in step four is that you find that the effect of your first variable. So this difference here doesn't stick much out compared to unexplained barriers. And that suggests that you are variable is not very important and that it makes sense to find a new variable that explains that remaining variants. And if that's the situation you end up with, then the final step for you would be to go back to step one and to formulate a new causal hap offices for a new variable. Okay, so that's the whole scientific thinking blueprint. And in the next lecture, I'm gonna go through a case study with you, in which I'm gonna show you how you can apply this blueprint I see in the next election 13. Case Study: Will Starting a Business Make You Rich?: all right, let's apply The scientific blueprint to a case study and the case study have chosen is the relationship between entrepreneurship and wealth, which is, I think, a very popular topic at this moment. So, in other words, is starting a business the way to get rich. And there are definitely some books that say that the answer is a definite yes, such as the book Rich Dad, Poor Debt or also the book The Millionaire Fast Lane, both very popular books and both books that I happen to read. And even though I thought that they are very smart books, we need to keep in mind that these books present opinions, and we also want to look what does the evidence actually say? So let's answer this question in this lecture. Let's apply the whole scientific thinking blueprint to get an evidence based opinion on this matter. Okay, so what's the first step? The first step is to formulate a causal hypothesis, and the causal app offices, in this case is simply that entrepreneurship causes wealth. Now the next step is to determine the costs off different kinds of errors. So the first era we can make is a false positive. We can think that entrepreneurship is the way to get wealthy, even though it actually isn't. And in that case we may waste a lot of money and we may waste a lot of time on a business that just most likely isn't going to make us rich. Okay, so that's a false positive. And the other possible error is a false negative, which means that we conclude that entrepreneurship isn't the way to get wealthy, even though it actually ISS. And in that case there's, of course, a huge missed opportunity. We could have made ourselves a lot more wealthy by starting a business, but we didn't now which of the two is worse and honestly, in this case, I think it's kind of hard to answer because it really depends on the business that you want to start. If your business is a very big business idea, that involves a lot off capital costs and a lot of risk, the risk of wasting a lot of money and time is quite huge. But on the other end, if you really want to get rich, if that's really important to you, then the missed opportunity may actually weigh heavier for you is really subjective here, and I can really say which one is worse. That's really a personal choice. And I'm going to say for now that both are equally bat. So I will try to be as neutral as possible and will not try to be biased towards either. Concluding that entrepreneurship does make people wealthy or concluding that entrepreneurship does not make people wealthy. All right, that was Step two. Now the third step is to select an appropriate test to figure out whether entrepreneurship really causes wealth. So what can we do here? Well, in an ideal scenario, we would want to run an experiment on ourselves. But that's of course, not very realistic, because for that you would actually have to start the business first and take the risk off wasting your money and wasting your time, while the whole point here is that we want to prevent that if it isn't really effective in making us wealthier. So even though this would be an ideal test if we want to get a definitive answer, it's not ideal overall. So the next best thing we can do is an experiment on other people. And that would mean that we flip a coin and then decides that you become an entrepreneur and you become an employee based on what the coin says. But of course, again, that's not very realistic, because then we need to force other people into taking the risk that we don't want to take ourselves. That's again not very realistic, and therefore we need to move to the next best option. So was the third best option. The third best option is a mix design on other people, which means that we don't flip a coin. But we just look at people who choose themselves to be entrepreneur or choose themselves to be employees, and we just follow them over time and see who does better over time. So in other words, we check how well they have been doing before. They either became employees or entrepreneurs, maybe at school, for example, and then we check how well they do after being, say, an entrepreneur or an employee for 10 years. Well, that's the third best thing, but that would cost us, in my example just now, 10 years, which again is not idea. So therefore we moved to the fourth best thing, which is an educated guess based on a correlation. So in essence, what that means is that we just look whether entrepreneurs are wealthier than employees. And if they are, then we speculate whether that is because entrepreneurship causes people to be wealthy or whether something else is going on here. So basically, the question I've chosen here is a worst case scenario where we need to resort to the weakest method that is available. But he will see that even that can be very informative. So what is the correlation between entrepreneurship and wealth? And to answer this question, I found a relatively recent research paper that investigated exactly that relationship the correlation between what they call household wealth and entrepreneurship. And we're gonna apply the scientific thinking blueprint now to interpret what they have found, and I want to show you their results now. But first I want to prepare you a little bit because they display their results in the form off a history Graham. So what's the his to grab? Here's an example. And in this history, Graham, the more you go to the rights, the higher the wealth and the more you go to the left, the lower the wealth and the bars tell you how many people you confined in each wealth range. So here we can see, for example, that most people are in the medium wealth range. And when we go to higher wealth and it gets less people and less people, And finally, in the very rich category, we find very, very few people. So that is how you read a history, Graham. And we can use a hist a gram now to compare employees to entrepreneurs. And here's another hypothetical example just to train you in reading these hissed a crops. So in this example, we can see that the entrepreneurs are much more on the wealthy side off the hissed a gram and the effect off being an entrepreneur rather than being employee. He is essentially the distance between this point and this point, and the unexplained variance is all the remaining variants here and here, and in this example, the effect of entrepreneurship is clearly sticking out from the unexplained variance. So this would be a history Graham that will be very much in line with a picture that gets sketched by these. Get rich books. Okay, here's another hypothetical results. In this example, the entrepreneurs are wealthier than the employees. But that effect doesn't really stick out from the unexplained various. So in this situation, even though it will be true that entrepreneurship has a higher potential for making you rich, it doesn't really matter that much. Then, in this example, there would be no difference between entrepreneurs and employees. And I want to show you one more example, which is this one. And I think this is the pattern that most people would expect in this better. And the employees are all in the medium wealth range, while the entrepreneurs spend all the wealth wage so they can be very, very poor and they can be very, very rich. And this will be most in line with the idea that entrepreneurship can make you rich. But that is also very, very risky so that he can also end up bankrupt and very, very poor. All right, With that in mind, let's have a look at the real pattern that the authors of the research paper found in the real world. And I've recreated the pattern here for you in a history crap And as you can see, there is hardly any difference between entrepreneurs and employees. If you look very closely, you can see that the history graham here off the entrepreneurs, the green one is actually shifted a little bit more to the wealthy site compared to the employees. But the difference is really small. If I would guess where the averages of these two groups are, I would say that the average off the employees is roughly here, while the average off the entrepreneurs is roughly there and you can see that there's hardly any difference. On average, entrepreneurs are not much more wealthy than employees. And that brings us to Step four checking the unexplained various. And we can conclude here that even though there is a weak correlation between entrepreneurship and wealth, there is also a lot of unexplained variance, which suggests that there are other factors which impact wealth a lot more than just the question whether a person is an employee or an entrepreneur. So to explain this a little bit more. A way to look at this is that the important question is not whether it is better to be an entrepreneur or an employee, but more in which of these two distributions can you end up in the upper range? Are you rather an above average employee, or are you rather an above average entrepreneur in which one can you outperform other people? All right, in principle, we could already stop here because we can already conclude that there isn't a strong correlation and that other things seem to matter much more than the question whether you want to be an entrepreneur or an employee, but just to train you in the whole scientific thinking blueprint, I also want to move to Step five and assess the three explanations that we can give for a correlation. So we found now that there is a weak correlation between entrepreneurship and wealth, and the question is now. What is the explanation for that correlation? And one explanation could be that entrepreneurship causes wealth. It wouldn't be a strong effect. It's not that entrepreneurship makes you super wealthy, but it could make you a little bit wealthier than being an employee. However, that's just one possible explanation. Another possible explanation is that wealth causes entrepreneurship because all we observed was a coordination, and so it could be that people who are wealthier are more likely to start a business and become an entrepreneur. That's another possible explanation. And finally, it could be that there is 1/3 variable that causes both. For example, it could be that ambitious people are more likely to start a business and become an entrepreneur and also to get wealthy, but not because they are entrepreneurs, but just because there are more ambitious. So there are many possible explanations for this correlation. And so we cannot even concludes that entrepreneurship makes people a little bit wealthier. And now, at this point, if your goal is to find a way to get wealthy, the next step would be to go back to Step one and formulate a new causal hypothesis. Find another possible cause off people getting wealthy and then test this one. All right, So to concludes, based on the evidence we have seen, we can conclude that there is a small correlation between entrepreneurship and wealth. Entrepreneurs are a little bit wealthier than employees, however, that correlation is very small, and it seems to be that other variables are not more important than whether a person is an entrepreneur or an employee. In addition, it remains unclear whether the small correlation actually reflects a cause of relationship . For example, it could just be that wealthy people are more likely to become entrepreneurs, and not that entrepreneurs are more likely to become wealthy. So what we can see here is that the popular opinion that has spread through many books that entrepreneurship is the way to become wealthy doesn't hold up that well when compared to evidence. And overall, a much more reasonable conclusion based on the evidence would be not that you should become an entrepreneur, but that you should ask yourself, Where can you perform better? Are you more suited to be an entrepreneur, or are you more suited to be an employee? In which of these two groups are you more likely to beat the average person? All right, that's the end of the case study, and I see you in the next lecture 14. Conclusion: Hey, congratulations on finishing this course. I hope you got a lot of value out of this course. I certainly try to do my best to put a lot of useful material into it. Even nevertheless, you still have open questions. Don't hesitate to ask me. And finally, please don't forget to leave a rating for this course. This is super important for me and also for future students who would still need to decide whether to take this course or not. All right, that's it. Thank you for the time and effort you put into this course and let me know if I can help you with anything.