Basically, if there is a relationship between two variables, it is called a correlation. If a correlation between two characteristics of the same attribute can be observed over time, this is called autocorrelation.
An example: Unemployment statistics show that in February, 12.00 million people were unemployed. In March, that number fell to 11.72 million before falling still to 11.66 million in April. These figures are related. A large proportion of the unemployed in March were still out of work in April and May as well. The amount of unemployed people has not suddenly a change in eleven million people. The number of unemployed in one month is always related to the number in the previous month – this represents an autocorrelation.
A counter example is the random number in roulette. If the 4 or 17 occur before the 36, neither value has an impact on the subsequent run of the ball. It is not possible to analyze the numbers that fall on a roulette night through a wavy curved graph in the same way as the unemployment statistics. Instead, such data takes the form of a jagged curve where values fluctuate wildly back and forth between the numbers 0 to 36.
Please note that the definitions in our statistics encyclopedia are simplified explanations of terms. Our goal is to make the definitions accessible for a broad audience; thus it is possible that some definitions do not adhere entirely to scientific standards.