Correlation and causality | Statistical studies | Probability and Statistics | Khan Academy

Main ideas and lessons

• The speaker reviews an excerpt from a WebMD article titled “Eating breakfast may beat teen obesity” and argues that the article frames a cause-and-effect story.
• The study details described suggest the relationship is correlational, not necessarily causal.
• The key educational lesson is the distinction between:
  • Causality: A causes B
  • Correlation: A and B happen together / are observed at the same time
• The speaker emphasizes that correlation can easily be misread as “breakfast prevents obesity,” when the data may only show that breakfast habits and obesity tend to coincide.

What the WebMD article claims (as described)

The WebMD excerpt implies that:

• Teens who eat breakfast regularly are less likely to be obese.
• Teens who skip breakfast are more likely to be overweight/obese.
• Regular breakfast eaters are more physically active.
• Over time, regular breakfast eaters show:
  • less weight gain
  • lower BMI than breakfast skippers

The study details mentioned (what was observed)

• Researchers analyzed dietary and weight patterns of 2,216 adolescents.
• Timeframe: over 5 years.
• Population/source: public schools in Minneapolis–Saint Paul, Minnesota.

The speaker’s critique: why this may not prove “breakfast causes obesity prevention”

The results described can support the claim that variables move together, such as:

• Breakfast skipping ↔ obesity (correlated)
• Breakfast eating ↔ physical activity (correlated)
• Breakfast eating ↔ lower weight gain / lower BMI (correlated)

However, the speaker argues the evidence does not show:

• whether breakfast directly causes obesity outcomes
• or which direction any relationship runs

The speaker also notes that multiple alternative explanations are plausible, including reverse causation and shared underlying factors.

Alternative explanations / hypotheticals offered

• Reverse causation (examples):

  • Maybe physical activity increases morning hunger, leading teens to eat breakfast.
  • Maybe obesity/high body fat affects morning appetite, leading teens to skip breakfast.

• Third-variable (confounding) explanations (shared causes):

  • Physical activity could underlie both:
    • higher likelihood of eating breakfast
    • lower likelihood of obesity
  • Sleep could be a factor:
    • insufficient sleep leads to obesity and also causes people to wake late and skip breakfast
  • Poverty / household constraints could influence both:
    • ability to eat breakfast consistently
    • overall diet quality and obesity risk
  • Stress and family structure could affect breakfast habits and health generally.

Implications for what advice should (and shouldn’t) be concluded

The speaker argues you cannot confidently conclude:

• “Eating breakfast will prevent obesity”
• “Skipping breakfast will cause obesity”

solely from correlational evidence.

More reliable conclusions would require evidence establishing causality (for example, studies designed to show that breakfast itself changes obesity risk).

The speaker also suggests blanket advice might even be harmful in some situations:

• If obesity is driven mainly by lack of activity, poor sleep, or unhealthy dinners, then forcing breakfast (especially if it’s unhealthy or if someone isn’t hungry) may not help and could potentially worsen outcomes.

Methodological emphasis (checklist)

• Distinguish clearly between:
  • Correlation (A and B occur together)
  • Causality (A produces B)
• Ask what the data can actually show:
  • Does it demonstrate the direction of effect?
  • Does it identify a mechanism?
  • Does it rule out other causes?
• Consider alternative possibilities:
  • Reverse causation (B may influence A)
  • Confounding variables (a third factor influences both A and B)
• Avoid turning correlational findings into absolute recommendations (e.g., “you must eat breakfast”).
• Treat broad dietary/behavior advice as conditional on underlying drivers (activity, sleep, stress/poverty, overall diet).

Speakers or sources featured

• Speaker/Instructor: Unnamed narrator (teaching-style commentary, likely from Khan Academy, based on the video title).
• Source discussed: WebMD (specifically an article excerpt).
• Study source mentioned: Pediatrics (journal title referenced; researchers analyzed adolescent data).