Science Literacy
How science actually works, what it can and can't tell us, and how to evaluate scientific claims without a PhD.
Why Science Literacy Matters
You don't need to be a scientist. But you need to understand how science works because:
- Medical decisions depend on interpreting scientific evidence
- Policy debates center on scientific claims (climate, vaccines, nutrition)
- Products are marketed using "scientifically proven" claims
- Pseudoscience exploits scientific language to sound credible
- Science reporting in media is often misleading
The goal isn't to become a scientist. It's to become a competent evaluator of scientific claims.
How Science Works
The Scientific Method (Simplified)
Observation → Question → Hypothesis → Experiment → Analysis → Conclusion
↑ |
└────────────────── Revision ←───────────────────────────────┘
This is the textbook version. In practice, science is messier, more iterative, and more social.
How Science Actually Works
| Textbook Version | Real Version |
|---|---|
| Hypothesis → test → accept/reject | Multiple competing hypotheses tested over years |
| One experiment settles the question | Dozens of experiments gradually build consensus |
| Objective, bias-free | Researchers are human; bias exists but is corrected over time |
| Linear progress | Messy, with dead ends, retractions, and paradigm shifts |
| Individual genius | Collaborative, building on others' work |
The Key Components
| Component | What It Does | Why It Matters |
|---|---|---|
| Hypothesis | A testable, falsifiable prediction | If it can't be tested, it's not science |
| Experiment | Tests the hypothesis under controlled conditions | Controls allow causal inference |
| Controls | Comparison groups that differ only in the variable tested | Without controls, you can't attribute effects |
| Replication | Others repeat the experiment to verify | Single results could be flukes |
| Peer review | Other experts evaluate methods and conclusions | Catches errors and weak reasoning |
| Publication | Results made public for scrutiny | Transparency enables correction |
Falsifiability: The Demarcation Line
A claim is scientific only if it can, in principle, be proven false.
| Falsifiable (Scientific) | Unfalsifiable (Not Scientific) |
|---|---|
| "This drug lowers blood pressure" | "Everything happens for a reason" |
| "The universe is expanding" | "Invisible, undetectable forces guide us" |
| "Meditation reduces cortisol" | "The universe wants you to succeed" |
Not being scientific doesn't mean something is worthless. Ethics, art, and personal meaning are important but not scientific questions. The problem comes when non-scientific claims dress up in scientific language.
Scientific Consensus vs. Individual Studies
What Consensus Means
Scientific consensus isn't a vote. It's the convergence of evidence across many studies, methods, and researchers over time.
| What Consensus IS | What Consensus IS NOT |
|---|---|
| The conclusion best supported by the totality of evidence | A majority opinion or vote |
| Something that can change with new evidence | Permanent or unquestionable |
| Developed over years through multiple independent lines of evidence | Based on a single study or authority |
| Strongest when methods are diverse | Valid just because many people agree |
How to Think About Consensus
| Situation | What It Means |
|---|---|
| Strong consensus (e.g., evolution, climate change, vaccines) | Overwhelming evidence from multiple independent lines; extremely unlikely to be overturned in fundamentals |
| Emerging consensus (e.g., gut microbiome effects) | Consistent direction of evidence, but details still being worked out |
| Genuine debate (e.g., optimal diet composition) | Evidence is mixed; reasonable experts disagree |
| Fringe claim against consensus | Possible but unlikely; burden of proof is very high |
Contrarian is not the same as correct. Galileo was right and persecuted, but millions of contrarians have been wrong. The fact that you disagree with consensus doesn't make you Galileo.
"A New Study Shows..."
The most misleading phrase in science reporting.
| What You Hear | What It Means |
|---|---|
| "A new study shows X causes Y" | One study found a correlation; replication and meta-analysis needed |
| "Scientists discover..." | One team published a finding; other teams haven't verified it yet |
| "Breakthrough" | Often incremental; "breakthrough" sells headlines |
| "Experts say" | Which experts? How many? What's the consensus? |
| "Studies show" | Which studies? How many? How strong? |
Rule: Never change your behavior based on a single study, especially if it contradicts established consensus.
Reading Scientific Papers
You don't need to read papers often, but knowing the structure helps you evaluate claims.
Paper Structure
| Section | What It Contains | What to Focus On |
|---|---|---|
| Abstract | Summary of the whole paper | Read first, but it often overstates findings |
| Introduction | Background, why this matters | Context for the study |
| Methods | How the study was conducted | Most important section. Sample size, controls, blinding |
| Results | What was found | Look at effect sizes, not just p-values |
| Discussion | Interpretation and limitations | Check if they acknowledge limitations honestly |
| Conclusion | Summary of implications | Often goes beyond what the data supports |
Quick Assessment Questions
| Question | Why It Matters |
|---|---|
| What was the sample size? | Small samples = unreliable results |
| Was there a control group? | No control = can't attribute causation |
| Was it randomized? | Without randomization, groups may differ systematically |
| Was it blinded? | Knowing group assignment introduces bias |
| Who funded it? | Industry funding correlates with favorable results |
| Was it pre-registered? | Pre-registration prevents after-the-fact hypothesis changes |
| Has it been replicated? | Single studies are unreliable |
| What's the effect size? | Statistical significance ≠ practical significance |
Understanding Risk Communication
Science communicates risk in ways that are easily misunderstood.
Absolute vs. Relative Risk
| Type | Example | Impact |
|---|---|---|
| Relative risk | "This doubles your cancer risk!" | Sounds terrifying |
| Absolute risk | "Risk goes from 1 in 10,000 to 2 in 10,000" | Not very scary |
Both describe the same thing. Relative risk is used to sensationalize. Always ask for the absolute numbers.
Number Needed to Treat (NNT)
How many people need to take a treatment for one person to benefit?
| Treatment | NNT | Meaning |
|---|---|---|
| Aspirin for heart attack prevention | ~120 | 120 people take aspirin for 1 to avoid a heart attack |
| Statins for high-risk patients | ~50 | 50 people take statins for 1 to avoid a cardiac event |
| Antibiotics for ear infection | ~8 | 8 children take antibiotics for 1 to recover faster |
Lower NNT = more effective treatment. An NNT of 1 would mean everyone benefits (extremely rare).
Risk vs. Hazard
| Concept | Definition | Example |
|---|---|---|
| Hazard | Can this cause harm? | Shark (yes, sharks can bite) |
| Risk | How likely is it to cause harm? | Shark attack (1 in 3.7 million per year) |
Something can be hazardous but not risky (sharks), or risky but not perceived as hazardous (driving).
Common Misunderstandings About Science
| Misunderstanding | Reality |
|---|---|
| "It's just a theory" | In science, a theory is the highest level of explanation, supported by extensive evidence (e.g., theory of gravity) |
| "Science proves things" | Science supports or falsifies claims with evidence; it rarely "proves" in the mathematical sense |
| "Scientists always agree" | Disagreement is normal and healthy; consensus emerges over time through evidence |
| "Natural = safe, chemical = dangerous" | Everything is chemical; natural substances can be deadly; synthetic ones can be lifesaving |
| "If it's not 100% effective, it doesn't work" | Nothing in medicine is 100%; effectiveness is a spectrum |
| "Correlation = causation" | See 05-evidence.md for detailed coverage |
| "Science is always self-correcting" | It's self-correcting in the long run, but errors can persist for years |
| "Published = reliable" | Publication is a first step, not validation |
Pseudoscience Red Flags
How to distinguish science from pseudoscience.
| Red Flag | What It Looks Like | Example |
|---|---|---|
| Unfalsifiable claims | No possible evidence could disprove it | "The energy is there but can't be measured" |
| Appeal to ancient wisdom | "This has been used for thousands of years" | Age doesn't equal effectiveness |
| Conspiracy as explanation | "Big Pharma is suppressing this" | Conspiracy explains lack of evidence |
| Anecdotes over data | "Thousands of testimonials!" | Stories instead of controlled studies |
| Immune to revision | Never updates with new evidence | Same claims for decades despite contrary evidence |
| Hostility to criticism | Attacking critics personally | "You're closed-minded" instead of addressing evidence |
| No peer review | Published in books, websites, or predatory journals | Bypasses scientific scrutiny |
| Special pleading | "Science can't measure this" | The effect conveniently can't be tested |
| Grandiose claims | "Cures everything" | Real treatments target specific conditions |
| Celebrity endorsement | Promoted by famous non-experts | Expertise in entertainment ≠ expertise in health |
The Pseudoscience Spectrum
Not everything is clearly science or pseudoscience. There's a spectrum:
Science ←————————————————————————→ Pseudoscience
| | | |
Established Emerging Unproven Debunked
consensus research but possible or unfalsifiable
Examples along the spectrum:
- Established science: Vaccines, evolution, germ theory
- Emerging research: Gut microbiome effects on mood, epigenetics
- Unproven but possible: Some supplements, certain alternative therapies
- Pseudoscience: Homeopathy, astrology, crystal healing
Science vs. Scientism
| Science | Scientism |
|---|---|
| A method for understanding the natural world | The belief that science is the only valid way to know anything |
| Acknowledges its limitations | Claims science can answer all questions |
| Descriptive: "How does this work?" | Prescriptive: "Science tells us we should..." |
| Humble about what it doesn't know | Dismissive of non-scientific knowledge |
Science can tell you:
- What the effects of a policy are
- Whether a treatment works
- How the climate is changing
Science cannot tell you:
- What we should value
- Whether an outcome is "good" or "bad"
- How to live a meaningful life
These require philosophy, ethics, and human judgment informed by, but not replaced by, science.
Practical Science Literacy
Evaluating Health Claims
| Claim Type | Reliability | Example |
|---|---|---|
| "Studies show" (no citation) | Very low | Marketing copy |
| Single study, dramatic results | Low | "New study: chocolate prevents cancer" |
| Multiple studies, consistent results | Moderate | "Several studies suggest exercise improves mood" |
| Systematic review / meta-analysis | High | Cochrane review of a treatment |
| Established medical consensus | Very high | "Smoking causes lung cancer" |
When You Encounter a Scientific Claim
- What's the source? Published paper? News article? Social media post?
- Is it peer-reviewed? Not all journals are equal; check the journal's reputation
- What's the sample size? Bigger is generally better
- Has it been replicated? Don't trust a single study
- Who funded it? Potential conflicts of interest?
- What are the limitations? Honest researchers state them clearly
- Does it match consensus? If it contradicts established science, the bar for evidence is very high
- Am I motivated to believe this? Watch for your own confirmation bias
Key Takeaways
- Science is a method, not a body of facts. It's the best method we have for understanding the physical world
- Consensus > individual studies. Never change behavior based on one study
- Falsifiability is the key test. If nothing could disprove a claim, it's not science
- Ask for absolute risk. Relative risk is almost always misleading
- Methods section matters most. How a study was done tells you more than what it found
- Pseudoscience has reliable red flags. Unfalsifiable claims, anecdotes over data, hostile to criticism
- Science has limits. It describes "what is," not "what should be"; values require philosophy
- Healthy skepticism ≠ denial. Question claims, but accept well-established evidence