What can a chart show about association without proving cause?
Quick answer
A chart of paired observations can show that two measured variables tend to change together in the data. That is an association, not proof that one variable caused the other: another factor could be connected to both. Describe what the chart displays, including its direction and scope, and reserve causal claims for evidence that can address other possible explanations. [S1]
The question, the details, the next step.
I made a scatterplot from a small set of paired observations. How can I say what it shows without claiming too much?
Start with what is visibly measured: name both variables, say whether the points generally rise, fall, or show no clear pattern, and limit your statement to these observations. For example, “In this sample, higher X values tend to appear alongside higher Y values” describes an upward association. It does not say X made Y increase.
A relationship between variables alone does not establish that one causes the other. A third factor could be related to both, so a chart by itself may not distinguish possible explanations. [S1] A neutral description is useful even when the cause is unknown.
Could you show that with invented numbers, so I can see how to phrase it?
Assume the following entirely invented example: a class records minutes spent reading and pages read by five students in one week.
- (10 minutes, 4 pages)
- (20 minutes, 8 pages)
- (30 minutes, 11 pages)
- (40 minutes, 15 pages)
- (50 minutes, 17 pages)
A cautious caption could be: “In these five hypothetical observations, more reported reading time tends to occur alongside more pages read.” This describes the paired values and their general upward pattern. It does not prove that adding reading time caused the page counts to rise. The example is illustrative, not research evidence, and it cannot establish what explains the pattern.
What wording should I avoid, and what can I say instead?
Avoid verbs such as “causes,” “leads to,” or “results in” unless your evidence supports a causal conclusion. An upward pattern in a scatterplot is not enough on its own.
Try these revisions:
- Instead of “Reading longer causes students to read more pages,” write: “In these observations, longer reported reading time is associated with more pages read.”
- Instead of “X improves Y,” write: “Higher X values occur alongside higher Y values in this dataset.”
- If the pattern is unclear, say: “These observations do not show a clear association.”
Mention that you are describing this dataset, rather than making a claim about every person or situation. The chart does not establish whether another factor is involved or settle a causal explanation. [S1]
Is there a quick way to check my caption before I turn it in?
Use this fill-in template: “Among [who or what was observed], [variable A] tends to [rise/fall/show no clear pattern] alongside [variable B]. This describes an association in [the dataset or period]; it does not by itself show that [A] caused [B].”
Then check that the caption matches the plotted points, names the variables correctly, and does not imply a wider result than your observations support. If you do not know how the data were collected or what other factors may be involved, say so rather than filling in a cause. The chart can describe a pattern; the available information may not determine why it appears.
A practical checklist
Name both measured variables and identify the observations being described.
Describe the visible direction or say that no clear pattern appears.
Use association language such as “occurs alongside” or “is associated with.”
Avoid cause-and-effect wording unless the evidence supports it.
State limits: the chart alone may not reveal other explanations.
A few more useful answers
Does a strong-looking pattern prove causation?
No. A strong pattern in paired observations still does not, by itself, show that one variable caused the other; another factor may be related to both. [S1]
Can I say that two variables are correlated?
You can use “correlated” as a description of an observed relationship when it fits the data, but do not use it as a synonym for “one causes the other.”
What if my chart has only a few points?
Describe only the pattern visible in those observations and make their limited scope clear. Do not turn a small dataset’s apparent pattern into a general or causal conclusion.
Sources
References used for this guide.
- [S1] 1.2 Data, Sampling, and Variation in Data and Sampling — Introductory Statistics openstax.org · Checked