Correlation, Mechanism and Causation in Mushroom Research

Research language often blurs correlation, mechanism and causation. This article separates the three, explains markers versus outcomes and applies it to mushroom research.

Correlation, Mechanism and Causation in Mushroom Research

Key takeaways

  • Correlation is a pattern; it doesn’t show cause.
  • A mechanism is a possible explanation, often from lab studies.
  • Causation in people is best shown by randomised controlled trials.
  • Changes in markers don’t guarantee meaningful outcomes.

“Linked to”, “shown to affect”, “works by…” Research language can blur three quite different ideas: correlation, mechanism and causation. This article separates them, explains the traps that catch researchers and readers alike, and shows how each applies to mushroom research.

Correlation

Two things are correlated when they tend to occur together. Observational studies find correlations, for example between diets and health outcomes. But correlation doesn’t show cause. For a worked example, see Mushrooms in the Diet and Cognition.

A classic example: ice cream sales and drowning rates rise and fall together. Ice cream doesn’t cause drowning. Both are driven by a third factor, hot weather, which brings people to both the ice cream van and the sea.

Why correlations mislead

  • Confounding. A third factor causes both things, as with the ice cream and hot weather. People who eat more mushrooms may also eat more vegetables, exercise more and smoke less.
  • Reverse causation. The effect may run the other way. People who are unwell may change what they eat, rather than their diet causing their illness.
  • Chance. Look at enough patterns, and some will appear by coincidence.

A mushroom example

A large Swedish study found that people with higher blood levels of ergothioneine, a compound found in mushrooms, had lower rates of heart disease and death over two decades. That is a genuinely interesting correlation. But ergothioneine was also the compound most closely linked to a health-conscious diet overall. So higher ergothioneine may simply mark people who eat well, rather than doing anything itself. Only controlled trials can separate the two. See Ergothioneine Explained.

Mechanism

A mechanism is a proposed explanation of how something might work, often found in laboratory studies: a compound binds to a receptor, switches on a gene, or blocks an enzyme in cells. A plausible mechanism makes an effect possible, not proven.

Many mechanisms seen in cells never translate into effects in people. The compound may not be absorbed, may be broken down before it reaches the relevant tissue, or may never reach the concentrations used in the dish. Cordycepin is a good example: it has many effects on cells in the laboratory, but in the body it is broken down quickly by an enzyme, which limits how much survives. See Cordycepin Explained and In Vitro, Animal and Human Studies.

Causation

Causation means changing one thing reliably changes another. The strongest evidence for causation in people comes from well-designed randomised controlled trials, ideally repeated independently and reviewed together. Randomisation is powerful because it spreads confounding factors, known and unknown, evenly between the groups, so the only systematic difference between them is the treatment.

Where trials aren’t possible, epidemiologists weigh several considerations together when judging whether a correlation is likely to be causal: how strong and consistent the association is, whether the cause comes before the effect, whether more exposure brings a bigger effect, whether there is a plausible mechanism, and whether experimental evidence supports it. No single one is enough on its own.

Markers versus outcomes

Studies often measure markers, such as levels of something in the blood or activity of certain cells, rather than outcomes people notice or care about. Markers are useful, faster and cheaper to measure, but a change in a marker doesn’t guarantee a meaningful change in health. For example, a Cordyceps trial reported changes in several immune markers in the blood over four weeks, but that doesn’t tell us whether participants were any less likely to fall ill. See Cordyceps Research.

Applying this to mushroom research

  • Laboratory mechanisms are a starting point, not a conclusion.
  • Observational patterns can’t isolate the effect of a mushroom from everything else about the people who eat it.
  • Marker changes need to be linked to real outcomes.
  • Controlled human trials are needed before drawing conclusions about effects in people.

Spotting the language

Phrase Usually means
“Linked to”, “associated with” Correlation
“May work by”, “activates”, “inhibits” Mechanism, often in cells
“Increased levels of”, “modulates” Often a marker
“Reduced”, “improved”, compared with placebo Possibly causation, if from a good trial

See How to Read Mushroom Research.

Last reviewed . How we research and write

Join the Spore & Bloom Journal

Receive occasional notes on mushrooms, botanical extraction, new releases and events.

We write occasionally and never share your address. Unsubscribe at any time.