Example
Suppose that you kept a record of the periodic changes in the price you paid for each delivery of heating oil and also the average temperature setting of your thermostat during the period covered by the specified price.
Given the following table:
A
B
1
Price
Setting
2
4.50
64
3
4.20
65
4
3.91
5
3.22
66
6
3.09
7
3.15
8
2.98
68
9
2.56
70
10
2.60
11
2.20
72
=CORREL(B2:B11,A2:A11) evaluates to approximately -0.907629573252938, indicating a close correlation (as prices rose, the thermostat was lowered). Correlation is a measure of how closely two variables (in this case, the price of heating oil and the setting of the thermostat) change together. A correlation of –1 (decreasing slope) or 1 (increasing slope) indicates perfect correlation. A correlation of 0 indicate the data sets are uncorrelated.
Example—Survey results
To see an example of this and several other statistical functions applied to the results of a survey, see the FORECASTSLOPE