An elementary mantra within the statistics and studies technology try relationship is actually maybe not causation, meaning that simply because several things appear to be associated with one another doesn’t mean this factors others. This can be a training well worth understanding.
If you work with studies, throughout your profession you will most certainly need to re-discover they a few times. However often see the main demonstrated that have a graph particularly this:
One line is something such as a stock market index, and also the other was an (most likely) unrelated go out collection including “Level of times Jennifer Lawrence is actually stated throughout the news.” The fresh outlines browse amusingly similar. Discover constantly a statement eg: “Correlation = 0.86”. Remember that a correlation coefficient is anywhere between +step 1 (the greatest linear relationships) and you will -step one (well inversely associated), that have zero meaning no linear matchmaking at all. 0.86 was a top value, appearing that the statistical relationships of the two date series try good.
The brand new relationship seats an analytical decide to try. This really is an effective instance of mistaking correlation for causality, proper? Really, zero, not: it’s actually a time series problem assessed badly, and you may a blunder which could was in fact avoided. read more
