Hedging <title>The Scientists' Press Briefing

What Went Wrong?

A real example of how science gets distorted in headlines:

Headline: "COFFEE CAUSES CANCER — Scientists Warn of Major Health Risk"

What the study actually said: "A preliminary observation in a lab model suggests a potential mechanism by which one component of coffee might, under specific conditions, show limited effects on cellular markers associated with cancer risk. Further research is needed."

The headline is dishonest. The scientist who said "potential mechanism" got turned into "causes cancer." This happens because journalists want certainty, but honest scientists communicate degrees of confidence. Your job in a press briefing is to hold the line—be honest about what your data shows while resisting pressure to overstate.

SPEAK

A journalist asks: "Does your study prove that X is true?" You have 60 seconds. How do you respond honestly without sounding evasive? Don't just say "no"—explain why certainty language is wrong in science.

1:00

Here are the six key dimensions of scientific hedging—the language that matches your actual level of confidence:

1Modal Verbs ▼

What it means: Use modal verbs (may, might, could, tend to) to match the strength of your evidence.

Too strong: "Sleep deprivation causes poor work performance."

Honest: "Sleep deprivation may contribute to reduced occupational performance."

The scale:

may/might — low confidence, preliminary evidence
tends to/appears to — moderate confidence, pattern observed
is likely to — stronger evidence, but not proven
has been shown to — well-established in research

2Epistemic Adverbs ▼

What it means: Adverbs like "possibly," "arguably," "apparently" signal uncertainty without saying "I don't know."

"This possibly indicates a shift in behaviour, though further work is needed."

"The mechanism appears to involve three key stages."

"Tentatively, we can say that..."

3Approximating Language ▼

What it means: Limit the scope of your claims. Don't let journalists turn your sample into a universal claim.

Too broad: "Remote work increases productivity."

Honest: "In our sample of 200 knowledge workers, remote work correlated with increased task completion rates."

Phrases: "in the majority of cases," "under certain conditions," "for populations similar to our sample," "in this context," "in the short term"

4Source Attribution ▼

What it means: Make clear where evidence is coming from. Preliminary data, our sample, current literature—attribute it.

"Based on our preliminary data..." (not "proven")

"The current evidence suggests..." (allows for future contradictions)

"According to the broader literature..." (situates your work)

5The Ladder of Certainty ▼

What it means: Know where your claim sits on the certainty scale. Use language that matches.

"X is proven" — only for laws of nature with universal evidence

"X is well-established" — decades of consistent evidence

"X is supported by evidence" — multiple good studies, some disagreement

"X is suggested by our data" — we found it, but it's preliminary

"X warrants further investigation" — initial signal, needs confirmation

6Correcting Misrepresentation ▼

What it means: When a journalist pushes you toward certainty you don't have, correct them clearly and respectfully.

"That's not quite what our findings show. What we found was [precise statement]. We cannot yet say [overclaim]."

"I want to be precise here: we can say X, but not Y."

"The headline you're suggesting goes beyond what the evidence supports. Here's what we actually observed..."

You are now a scientist preparing to brief journalists about new preliminary findings. Choose a research area and prepare a 2-minute briefing using appropriate hedging throughout.

Research Areas (pick one or use your own):

Option A: Climate Science — New data on regional sea-level rise acceleration. Preliminary, contested, highly politicized.

Option B: Nutrition Research — A study linking processed foods to metabolic changes. Small sample, short-term, but media will want "proof".

Option C: Psychology of Learning — Preliminary findings on AI tutoring effectiveness. Only one school, but journalists want scalability claims.

YOUR TASK

Step 1: Choose your research area (1 min)

Step 2: Prepare your briefing (3 min): Know your findings clearly. Plan how to hedge appropriately.

Step 3: Deliver your briefing (2 min): Explain your findings with honest, careful language.

Prep Time:

3:00

Briefing Time:

2:00

You've finished your briefing. Now the journalists push back. They want certainty, sensational headlines, oversimplification. You have 45 seconds to respond to each question while maintaining honesty and precision.

SCENARIO

Six journalists ask progressively more aggressive questions. Your job: stay calm, be clear, correct misrepresentation, but don't hide your findings behind jargon.

Question 1

Journalist: "Your study found X. Can you just say clearly: is X true or not? Yes or no?"