Averaging sleep spindle occurrence in dogs predicts learning performance better than single measures

Correlations carry exciting implications for researchers, but in addition to only indirectly hinting at causality, they can be entirely misleading, as well. Consider, for example, the association between the decline in active pirates worldwide and the rise in global temperatures, or the link between ice-cream sales and murders during summer.

 
It becomes clear, looking at these examples, that some correlations are indeed only false positives, or in the language of statistics „Type I errors”. They are likely artifacts of sampling error, i.e. the fact that we deal with limited samples of the phenomena we try to study. What we see in these samples can sometimes deviate very much from what is really going on in the bigger picture.