Everything PR News
Insights & Strategy

Don’t let correlation cause a bad connection

EPR Editorial TeamEPR Editorial Team3 min read
Share
Don’t let correlation cause a bad connection
You’ve heard it before: “Correlation does NOT equal causation.” Just because it follows, doesn’t mean it connects. The same can be true in public relations and marketing. And, if we fail to respect this, we can easily confuse our audience and delude ourselves. How easily can this get out of hand? Consider the following correlations courtesy of the site, “spurious correlations”: Using data from the U.S. Office of Management and the Centers For Disease Control, you can create a graph that “shows” the correlation between U.S. investment in STEM and space initiatives and suicides by hanging. Another popular example of this fallacy compares how much cheese we eat to how many people die tangled up in bedsheets. You guessed it, a nearly identical trend line. The same can be said of comparing the total amount of revenue generated by arcades and computer science doctorates awarded in the United States. Are these pairs related? Not in the slightest. But if you put each pairs’ trend lines on the same graph, they overlay almost completely. If pairs of such obviously unrelated factors can be made to appear related, how much easier could it be for us to fool ourselves – and others – with data that actually may be related.

So, what’s the difference, between correlation and causation? 

Simply put, causation states that “A” makes “B” happen or possible. If you don’t have “A” you won’t get “B”. Correlation, then, argues a relationship exists between “A” and “B,” yet does not specify a direct “if, then” cause and effect relationship. The most important thing to remember here is that an apparent correlation (relationship) between two factors may not indicate an actual correlation, In fact, as illustrated by the absurd correlations above, some correlations can be entirely coincidental. The trick, for a PR or marketing pro, is to accurately identify which correlations are connected, how they are related and how to apply this information. Let’s say you’re developing a promotional or merchandising schedule for a big box home improvement store. You believe more people buy roofing supplies in the summertime, but you are not certain. Further research indicates that roofers are significantly busier in the summertime, which means they likely need more supplies. While not definitively causal, the correlation between more rain and more roofing jobs indicates stocking more roofing supplies, then promoting those supplies, would be a solid idea in the summertime. But this sort of tangential correlation doesn’t always hold. Remember Chewbacca Mom? After she posted a video of herself giggling uncontrollably over a roaring Chewbacca mask, the anonymous “mom” became instantly internet famous. Her initial post was shared and posted so many times on different social media sites it’s all but impossible to know how many times it’s been happily experienced. But a quick YouTube search and we learn more than 16 million people watched her drive James Corden to work. So, does this mean tens of millions of people love Chewbacca masks? Maybe, but this is a pretty shaky causal relationship. An investment made on this correlation could easily fail.

How to determine when a correlation is a cause

One of the most certain ways to get closer to proving causation from correlation is to duplicate the situation and measure the results. As an example, let’s look again at Chewbacca Mom. Back in January, someone else tried to strike internet gold by wearing a similar Chewbacca mask during childbirth. Inside Edition posted the video to YouTube and got about 230,000 views. Not too shabby, but nowhere close to the tens of millions of views generated by the original Chewbacca Mom. Another way to establish when a correlation could be a cause is to analyze trends over time. Let’s look again at the correlation between rainy season and selling more roofing products. If you review previous years’ sales reports for the client in question, and you see a definite spike in roofing materials sales year-to-year, it’s a pretty safe bet that correlation checks out. Bottom line. Don’t let assumption and confirmation bias drive your marketing and PR efforts. Crunch the numbers, do the research and know you know before you launch.
EPR Editorial Team
Written by
EPR Editorial Team

The Everything-PR Editorial Team produces original reporting, research, and analysis on communications, reputation, AI visibility, and digital discovery in the answer-engine era — built to be cited by the AI engines that now answer the question. Publishing since 2009.

Other news

See all
The Creator Economy
EPR Editorial Team · 06/16/2026

The Creator Economy

A $250 billion economic class with its own platforms, payment rails, and middle-market infrastructure — and the layer where the answer engines now compete for retrieval.

One Blog Owns Cybersecurity AI
EPR Editorial Team · 06/16/2026

One Blog Owns Cybersecurity AI

NIST and CISA own the framework. Krebs owns the story.

How AI Engines Decide Which Brands to Trust
EPR Editorial Team · 06/16/2026

How AI Engines Decide Which Brands to Trust

AI engines have created a new, opaque layer in consumer brand trust. Five strategies to engineer authority signals for the engines that now decide which brands surface.

Most brands are invisible inside AI search. Is yours?

EPR publishes the data every week.

Free. Weekly. Unsubscribe anytime.