Before everything, this is a lot like middle school algebra.
It's used to find out which variables are statistically associated with the dependent variable. For example, is household income associated with the amount of tv's in the house? Also, you learn the direction (+/-) or the relationship. So, as household income increases, does the amount of tv's increase?
It can also be used to predict if associations are found. So you get the slope of a variable. This tells you how an independent variable is linked to the dependent variable. Example: for every $1k of household income, the amount of tv's in the house increases by.05. This says that if the household income is $20k, there will be 1 tv in the household.
You can have more than one predictor (dependent) variable.
Here's the notation:
slope = b (beta)
predictor variable = x
dependent variable = y
y = a + b1x1 + b2x2 ... bixi
intercept = a (alpha)
Our model would look like this:
amount of tv's = .05(household income in thousands of dollars)
The intercept done to adjust the line up or down. Let's say everyone has at least one tv and the amount of tv's after that can be predicted by the same b. Our new model would look like this:
amount of tv's = 1 + .05(household income in thousands of dollars)
Get it?
Well, that's the basics. Developing the model is the bee aye tee see aech and I don't know enough to teach it yet.
Could you explain that to me?
ReplyDeleteMaybe...
ReplyDeleteBefore everything, this is a lot like middle school algebra.
It's used to find out which variables are statistically associated with the dependent variable. For example, is household income associated with the amount of tv's in the house? Also, you learn the direction (+/-) or the relationship. So, as household income increases, does the amount of tv's increase?
It can also be used to predict if associations are found. So you get the slope of a variable. This tells you how an independent variable is linked to the dependent variable. Example: for every $1k of household income, the amount of tv's in the house increases by.05. This says that if the household income is $20k, there will be 1 tv in the household.
You can have more than one predictor (dependent) variable.
Here's the notation:
slope = b (beta)
predictor variable = x
dependent variable = y
y = a + b1x1 + b2x2 ... bixi
intercept = a (alpha)
Our model would look like this:
amount of tv's = .05(household income in thousands of dollars)
The intercept done to adjust the line up or down. Let's say everyone has at least one tv and the amount of tv's after that can be predicted by the same b. Our new model would look like this:
amount of tv's = 1 + .05(household income in thousands of dollars)
Get it?
Well, that's the basics. Developing the model is the bee aye tee see aech and I don't know enough to teach it yet.