Two types of Correlation:
- Positive correlation
- Negetive Correaltion
The correlations are all about how two (or more if you use more advanced methods) which relate.
Positive Correlation:
Positive correlations mean that one variable increases, the other variable increases. An example, I remember one of my texts is the possibility that ice cream sales and drownings are positively correlated.
So let's say that the correlation between these two variables is .50.
Negative correlation:
Negative correlations mean that as one variable increases, the other decreases. An example would be smoking and life expectancy, so you smoke more, you will probably have a lower than expected life.
So let's say that the correlation between these two variables is - .80.
The two numbers above examples, these are .50 and - .80, are called correlation coefficients.
Now, one thing to remember is that the negative in the above simply refers to the direction of the correlation. So you can ask me in a test is something like "what the correlation is higher?".
In this case, the second example is the strongest correlation is higher, because .80 is obviously greater than .50.
There is a problem with correlations however. It is with correlations that we can not necessarily say that a variable causes the other. This is summarized by the common phrase "correlation does not equal causation.
What is this?
Well one problem with correlations is the possibility of other variables could be the cause of the correlation. If we take the example of the first ice cream sales and drownings, one could say that sales of ice cream are causing drowning because they are correlated.
But there is a third variable, which is the seasons. More people will buy ice cream and swimming in the summer, and then it would affect the correlation.
The other reason is that correlations do not tell us what the variable is the cause of change in the other variable.
Here, I'll change examples and use anxiety and depression. These two are often correlated in psychological research, but what we do not know is whether high anxiety causes severe depression or if depression causes high high anxiety.
Of course, with every rule there are exceptions. This problem of not being able to identify the causal link can not be treated by trying to account for possible third variables and others. But it requires a little more advanced statistical procedures, in order not important in the first year :)
