Edited, since my original post was not helpful: If you have a ton of random survey data, you might indeed find two variables which correlate on ten data points.

Edited, since my original post was not helpful: If you have a ton of random survey data, you might indeed find two variables which correlate on ten data points. On the other hand, it is much less likely that you accidentally design a study with two variables, and find a similarly strong correlation across 1487 data points. This is particularly true when there is already evidence that this correlation should hold.