In qordata’s most recent webinar, Ned Mumtaz revisited a concept mostly familiar to undergrads and professional statisticians. (And increasingly unfamiliar for the rest of us).
The concept differentiated between correlation and causality.
Corporate citizens tend to use ‘correlation’ as a blanket term for any variables that seem to have a corresponding relationship. As a compliance professional, you may hear statements like:
a) There’s a correlation between prescriptions and speaker events.
b) High demand for dermatological treatment is correlated to the Summer Season.
c) Decreasing spend on Lunch & Learn events is correlated with falling sales
The problem with this is that when they say this, they usually mean ‘directly proportional,’ but the statement holds little weight when based on observation rather than scientific validation.
Unfortunately, decisions can take a turn for the worse when observation dominates evidence.
One must add that not enough is said about inversely proportional relationships between variables (or as they’d be called), “negatively correlated” factors, which means that decision-makers miss out on half the view when executing calls to action.
This doesn’t mean missing out on declining demand but failing to proactively identify problem areas. By applying the correlation concept, we can collect evidence for sales deceleration before actual decline takes place. This is one of the factors at work in Compliance Monitoring.
And Now, Causality
Causality is a universal favorite when it comes to rushing to the wrong conclusion. Incorrect information, opinions disguised as facts, logical fallacies, and cognitive biases all play a role in making sure we gain the wrong insight by way of causality.
From the pharma-commercial perspective, for instance, statements like:
· Your company loses an account because a core sales team member left
· Performance declined because competition launched a similar product, and promoted it more aggressively than you did
· Regulatory inquiry is responsible for delays in launching your own drug
