Significant Linear Correlation or Not
- Author:
- David Gurney
- Topic:
- Statistics
Move the Confidence slider to change the confidence.
Move the Sample Size slider to change the sample size.
Move the Number of Correlations Computed to change the number of correlations being computed.
Uploaded 12/22/2022.
This applet computes many linear correlations between randomly generated lists of integers. The correlation values are represented by the circles. Zero correlation or no correlation occurs when a circle is centered at zero. Notice that zero correlation values do not happen very often. This means that for most pairings of lists of numbers, there is some linear correlation.
Linear correlation is significant when there is a strong linear relation between two lists of numbers. When testing for significant linear correlation, the correlation is considered to be significant when the correlation value for the two lists is centered in one of the red regions. If you look at different confidence levels and sample sizes, and compute different numbers of correlations, you will see that significant linear correlation happens much less often than non-significant linear correlation.