Source: Table created by Dan Laub;Image of stethoscope, PD, http://bit.ly/1RAnV5x; Image of jogger, PD, http://bit.ly/1Joowze; Image of gas pump, PD, http://bit.ly/1O8GL30; Image of SUV, PD, http://bit.ly/1PgcrAO; Image of car, PD, http://bit.ly/1Nyhegs; Image of graduate, PD, http://bit.ly/1NTA7Mi; Image of money bag, PD, http://bit.ly/1RSaTQ3; Image of exam, PD, http://bit.ly/1TcyDge; Image of sleeping man, PD< http://bit.ly/1QBROT0; Image of house, PD, http://bit.ly/1QBRWSt; Image of sale tag, PD, http://bit.ly/1NHfwad
[MUSIC PLAYING] Dan Laub here, and in this lesson, we're going to talk about identifying a reason for performing the experiment. And so let's talk about some of the key objectives for this lesson. One of the first objectives I want to talk about today is using the experimental method and talking about how it involves hypothesis. And we're talking about what a hypothesis is and how it assists us in establishing the cause and effect relationship between an explanatory variable, as well as a response variable.
We are also going to look at how a hypothesis is an educated guess. How it's not a random statement involving the relationship between two variables, the explanatory variable as well as the response variable. However, we're looking at the experimental method, and we're looking at the information that we've gathered, and we're drawing conclusions based upon that method.
And finally, we're going to talk about the difference between the null hypothesis and the alternative hypothesis. What difference exists between the two, and why it's important that we use both in order to come to an educated guess about the relationship between the two variables in question. Before we start talking about hypotheses, let's just kind of take a minute here to bring you up to speed in terms of the experimental method.
So if you recall from previous lessons, the experimental method basically consists of conducting experiments in order to understand the cause behind something, and to understand the relationship about how one variable might affect another one. And the variables in question here are going to be the explanatory variable as well as the response variable. And the idea behind these is pretty clear.
And so the explanatory variable is basically what are we changing? Necessarily, what are we affecting that's going to cause a change to something else. Whereas the response variable is, let's look at how a certain variable responds to the changes that we make with regard to the explanatory variable. And so what I want to do here is go over a few examples just to kind of reinforce this point, as well as to introduce the whole idea of a hypothesis.
So I mentioned a moment ago the idea that a hypothesis is essentially an educated guess. And what we're doing with a hypothesis is basically establishing a parameter for the experimental method. And so we're basically saying, what happens if we change the explanatory variable, and how would that actually affect the response variable? And so this educated guess is not something that we randomly state.
I'll just give you a simple example. We're basically saying, well, maybe we think that a change in variable a, for instance, causes a change in variable b. And essentially, that's our hypothesis. It's an educated guess based upon what we have observed. And the idea behind using the experimental method is largely we're going to set the parameters for the experiment, and then go ahead and determine if there's any cause in order to support that hypothesis.
So to use a relatively simple example of how a hypothesis might work in terms of making a prediction about the cause and effect that exist between a couple different variables, let's assume that we have a particular person that goes to the doctor. And they say their blood pressure is a little high, and they would like to go ahead and see what they can do to keep an eye on that.
And so maybe the person goes back to the doctor in a month, and the next thing you know, their blood pressure is relatively low. Well, one thing that might have happened is they've done a lot of exercise in the interim. And they can hypothesize that, well, since my blood pressure went down, maybe it's the exercise that actually leads to that change taking place. And that's what the hypothesis here would be all about-- making an educated guess about the relationship that exists between these two variable.
So a good example of a hypothesis would be, what happens when gas prices go up? And what effect does that have on the kind of cars that people are buying? So let's use a simple example here. Let's say gas at one point is $1.50 a gallon. And at that relatively low price, we have a lot of people going out and buying expensive vehicles.
And then as the gas price eventually gets up to a higher point, say for instance $3 a gallon, we notice that people drive relatively smaller cars. Now, why would this be? Now, our hypothesis could state, well, it's simply the fact that the gas price is what's actually causing this to happen. Gas prices become more expensive, therefore, it's more expensive to drive a car, or a large vehicle that gets lower gas mileage. And therefore, people opt when they replace their cars, to get smaller cars that get better gas mileage.
However, that might not be the only explanation here. What else could prompt somebody to buy a smaller car? Perhaps the insurance rates on large cars have gone up as well. And that's not necessarily reflected in our initial hypothesis. Or perhaps incomes have gone down and people can't afford the large cars. And as a result, they might want to buy the smaller ones instead.
And so the idea here is that our hypothesis might state what we believe to be the case. However, there could possibly be other things that don't necessarily mean that simply an increase in the price of gas is causing a change in the type of cars that people drive.
In this next example, I'd like to go over the relationship that exists between education and the income that someone earns throughout the course of their career. I think most people understand that if you have a college degree, it typically means your earnings will be higher over the course of a lifetime. But is getting the degree causing higher incomes, or could there possibly be something else going on here?
And the idea would be that our hypothesis might be, well, simple. College degree translates to higher earnings. And data would tend to back that up if we use the experimental method, that we would see people are earning more typically would have college degrees. But then again, there could be other factors at play here as well.
What about people that choose to get a college degree might be more driven in general, and therefore, they would typically do well in terms of their earnings regardless if they had the degree or not, relative to their peers. And so there's another way of looking at that. We're simply stating what we believe to be the case. And we're basically making an educated guess based upon what the data is actually telling us.
So speaking to the idea of a hypothesis, we actually have two types of hypothesis. One of which is what's called a null hypothesis, basically meaning that there's no existing relationship between the variables in question. Both the explanatory as well as the response variable. And the other is called the alternative hypothesis. And the alternative hypothesis is saying there is a relationship that exists between the explanatory variable and the response variable.
With the null hypothesis, we're largely just saying that it's possible that the explanatory variable is really just affecting the response variable by chance. In other words, there's no direct relationship there. It just happens to be something that's occurring. And it's not necessarily something that's causing one other effect to occur within the response variable.
And with regard to the alternative hypothesis, we're suggesting that there actually is some relationship, there's some cause and effect going on here between the explanatory as well as the response variable. As an example, let's look at test scores. How well a given class might do on a particular exam.
So in a case like this, maybe we're testing to see if there's a relationship between the amount of sleep a student gets and how well they perform on an exam. And so our null hypothesis would be that sleep and student performance are not related. The alternative hypothesis, on the other hand, would be that perhaps increased amounts of sleep would lead to better student performance.
We'd see if the relationship between sleep and student performance actually existed, whether or not there's a cause and effect relationship that exists between the two. It's worth noting that we can never accept the alternative hypothesis. We could reject the null hypothesis. We could fail to reject the null hypothesis. But under no circumstance will we ever accept the alternative hypothesis. And the simple reason is that it might not always be true under all circumstances.
So let's go through a couple different examples of how we would actually look at a situation and identify what a null hypothesis is under those circumstances, as well as what the alternative hypothesis would be. So for the first example we want to look at, let's say we are in the real estate business. And we are in the process of selling a home. And we are interested to see whether or not the home that we're selling is of equal value to a neighboring house.
And so our null hypothesis in a case like this, well, let's assume that they are of equal value. And the alternative in this case would be that, well, one of the homes has to be worth more or less than the other home. That would be the alternative hypothesis in this case.
The second example, let's talk about a business, a store, having a sale. And let's assume that our null hypothesis here is the sales not going to make one bit of difference with regard to bringing in additional customers. The alternative hypothesis in this case would be, well, what's the opposite of that? In this case, it's going to be, well, it could attract additional customers.
And that's what we're looking at in terms of how we define that alternative hypothesis in that individual circumstance. And so when we're looking at the difference between null and alternative, the words we often use, that we see, describing a null hypothesis would be where a situation is equal, or less than or equal to, or greater than or equal to.
The alternative hypothesis, on the other hand, we see words that would suggest unequal, or less than, or greater than, or different. And the idea there is that we have the complete opposite. So if the null hypothesis states that something is equal, the alternative must state that it's inequal.
So to wrap things up, the objectives we discussed covering at the beginning of the lesson-- we use the experimental method and we use hypotheses in the process of establishing some degree of cause and effect. So the idea is that a hypothesis is an educated guess, not simply a random statement. And the hypotheses can be broken down into two types of hypotheses, both the null and the alternative. So hopefully you got some value from today's lesson. And once again, my name is Dan Laub.
(0:00 - 1:13) Introduction
(1:14 - 2:10) Explanatory and Response Variables
(2:11 - 3:35) What is a Hypothesis?
(3:36 - 4:50) Hypothesis Example 1
(4:51 - 5:43) Hypothesis Example 2
(5:44 - 7:30) Null and Alternative Hypotheses
(7:31 - 9:02) Null and Alternative Hypothesis Examples
(9:03 - 9:26) Conclusion
States that two variables are not related.
States that two variables are related, and how they are related.