Best of Both Correlation is the stronger of a factor’s two correlation scores: RankGear computes both a Pearson and a Spearman correlation for the factor, then keeps whichever gives the more usable signal.
| Term | Best of Both Correlation |
|---|---|
| Category | Statistics and Evidence |
| Also known as | Best of Both |
| Where it appears | Workbook Overview and LSI Keywords |
What it means in RankGear
Every factor RankGear measures is scored two ways at once. Pearson correlation looks for a straight-line relationship between the factor and ranking position, while Spearman correlation looks for a consistent rank order regardless of whether the relationship is perfectly linear. The two methods often disagree on how strong the association is, because real SERP data rarely behaves in a clean straight line. Rather than force you to pick a method, RankGear reports the Best of Both value: for each factor it compares the two results and surfaces the stronger, usable one. You will see this figure in the Workbook Overview, where factors are ranked by signal strength, and in the LSI Keywords view, where it helps separate terms that track with the top results from terms that do not.
How to interpret it
Read the value the way you would read any correlation, starting with its sign and scale. A value near +1 means the factor tends to move with better rankings across the SERP you analyzed; near -1 means the opposite; near 0 means no consistent relationship was found. Because the number is the stronger of two methods, treat it as the most favorable honest reading of that factor, not a guaranteed effect. Always weigh it against sample size: a strong correlation drawn from a handful of results is far less trustworthy than a moderate one drawn from a full result set. Check, too, whether the value is genuinely measured, simply missing, or a true zero, since an empty cell and a real zero mean very different things.
| Input | What it captures |
|---|---|
| Pearson | Strength of a straight-line relationship between the factor and rank position. |
| Spearman | Strength of a consistent rank ordering, even when the relationship bends. |
| Best of Both | The stronger, usable of the two — the value RankGear surfaces for the factor. |
Example
Suppose you analyze a query and a factor such as word count returns a Pearson correlation of 0.28 but a Spearman correlation of 0.51. The linear fit is weak, but the higher-ranking pages still order cleanly by length, so Spearman captures a relationship Pearson misses. RankGear reports 0.51 as the Best of Both value and lists the factor higher in the Workbook Overview than the Pearson figure alone would suggest — a prompt to look at how the leading pages actually use length before you act on it.
Important considerations
- Statistical association is not causation. A strong Best of Both value tells you a factor moves with the current top results; it does not prove that changing the factor will move your page.
- Small SERP samples are unstable. With few results, the stronger of two correlations can still be noise, so favor patterns confirmed across a full result set.
- Correlation values are comparative indicators drawn from the pages you analyzed, not scores issued by Google. Use them to prioritize where to look, not as a ranking guarantee.
- Because the figure is the stronger of two methods, it flatters the factor slightly by design — pair it with sample size and distribution before treating it as a target.
Related terms
Part of the RankGear glossary · how RankGear measures · the 870 factors.