Do two things still change together when we consider something else?
Partial correlation analysis is a statistical method used to measure the relationship between two variables while silencing the effect of one or more other variables.
For example, in our own B2B Buyer Experience Research, we asked buyers to rate their satisfaction with both the contracting phase of the purchase process, and the helpfulness of sales resources. We wanted to understand how the contracting process influenced their satisfaction rating of the overall buying process, but we didn’t want the helpfulness of the selling resources to influence that assessment. The partial correlation analysis calculates the correlation between contract process sentiment and overall satisfaction, while mathematically removing – or controlling for – the effect of seller helpfulness.
In this example, both contract negotiations and seller helpfulness were positively correlated with the buyer’s overall satisfaction with the buying process. However, when the seller helpfulness was partialed out (controlled for), the correlation dropped by nearly a third. It is still a statistically reliable relationship, but it is clear that seller helpfulness influenced how buyers feel about the contracting process.
