Correlation Is Not Cause, and a Trend Is Not a Guarantee
The last chapter showed how to surface what a number assumes and where its limits lie. Even a figure that survives that scrutiny can still be misread in two tempting ways: by treating a link between two things as proof that one caused the other, and by treating a run of past results as a promise about the future. This chapter takes on both traps, because they slip past almost everyone.
See Why a Link Feels Like a Cause
When two things rise and fall together, the mind reaches for a story, and the easiest story is that one is making the other happen. Sometimes that is true. Very often it is not, and the tidy story hides a plainer explanation.
The gap between those two words is where the trap lives. Seeing two things move together is real evidence that something connects them, but “something connects them” is not the same as “this one causes that one.” Before you accept a cause, you have to rule out the quieter explanations that would produce the very same pattern.
Spot the Three Correlation Traps
When you meet a “these two move together, so one must cause the other” claim, three explanations can produce that pattern without the cause being real. Run through them before you believe the story.
Coincidence: with enough things being measured, some will line up by pure chance, and no real connection exists at all;
A third thing behind both: some other factor drives both at once, so they move together without ever touching each other;
The cause running backwards: the link may be real but pointing the other way, with the supposed effect actually doing the causing.
The middle trap, a hidden third thing, is the one that fools people most, so it is worth having a clear example in mind.
Watch It Work: Ice Cream and Sunburns
Picture a town that tracks two numbers across the year: how much ice cream gets sold, and how many people are treated for sunburn. The two rise and fall almost in step. Someone glances at the chart and concludes, “eating ice cream must be causing the sunburns.” Before nodding along, run the traps.
Coincidence? Unlikely, the pattern is strong and repeats every year, so something real is going on;
A third thing behind both? Yes. Hot, sunny weather drives both at once: people buy more ice cream and spend more time in the sun, where they burn. The sunshine is the hidden cause sitting behind both numbers;
The cause running backwards? No, getting a sunburn plainly does not send anyone out to buy ice cream.
Once you see the third thing, the “ice cream causes sunburn” story collapses. The correlation was real, the ice cream and the sunburns genuinely moved together, but the cause belonged to a factor neither number mentioned. That is the whole move: when two things travel together, ask what else could be pushing both before you let one take the blame.
Know Why a Trend Is Not a Promise
The second trap swaps two things moving together for one thing moving over time. A trend is a direction a number has moved across a stretch of the past: three months of rising sign-ups, a week of falling temperatures, a season of wins. A trend is a useful summary of what already happened. It becomes a trap the moment you read it as a guarantee of what happens next.
Past results carry forward only as long as the conditions behind them hold, and conditions change. A streak can end because it was partly luck to begin with, because the thing driving it runs out, or simply because nothing says a line that went up must keep going up. A trend earns you a reasonable expectation, not a promise.
If something has happened five times in a row, isn’t it safe for me to expect the sixth?
It is reasonable to lean that way, but not to bank on it. Five in a row raises your confidence; it does not remove the chance that the sixth breaks the pattern. Ask what has been driving the streak and whether that is still true, rather than trusting the streak by itself.
Watch It Work: The On-Time Train
Suppose the 8:10 train has arrived on time every morning this week, five days running. By Friday you think, “this train is always on time, I can trim my buffer and leave later next week.” The trend is real: five on-time mornings is a genuine pattern. Does it guarantee a sixth?
Not on its own. Ask what has been behind the streak. Perhaps this was a calm week with light traffic and good weather, conditions that will not hold once a storm, a breakdown, or a busy Monday arrives. The five on-time days are a fair reason to expect punctuality, not a promise you can build a tight schedule on. Reading the streak as a guarantee is exactly how you end up stranded on the platform the one morning it matters.
That is the trap in miniature: a past direction is evidence about the past, and only a cautious hint about the future. Grant a trend the confidence it earns, then ask what would have to stay true for it to continue.
Let’s Recap!
A correlation means two things move together; a cause means one makes the other happen, and the first does not prove the second.
A link can still be a genuine clue, so treat a correlation as a prompt to ask why, not as settled proof of cause.
Three explanations can fake a cause: pure coincidence, a hidden third thing driving both, or the cause running the other way.
A trend describes a direction in the past and earns a reasonable expectation, never a guarantee about the future.
Grant a correlation or a trend the confidence it earns, then ask what is really driving it before you rely on it.
You can now judge fit, surface assumptions and limits, and resist the correlation and trend traps; the final teaching chapter turns all of this into an honestly stated claim, with the right amount of confidence and no overclaiming.
