Media Literacy
How to navigate the modern information environment: evaluating sources, identifying bias, and building a reliable picture of reality.
The Information Problem
The information environment has fundamentally changed:
| Era | Problem | Filter |
|---|---|---|
| Pre-internet | Too little information | Libraries, gatekeepers |
| Early internet | Finding information | Search engines |
| Current | Too much information, much of it false | You |
You are now your own editor, fact-checker, and curator. Nobody else will do it for you.
Evaluating News Sources
The SIFT Method
A fast, practical approach to evaluating any claim you encounter online:
| Step | Action | Time |
|---|---|---|
| S: Stop | Don't react immediately. Pause before sharing or believing. | 2 seconds |
| I: Investigate the source | Who published this? What's their track record? | 30 seconds |
| F: Find better coverage | Search for other sources reporting the same thing. | 1 minute |
| T: Trace claims | Find the original source of the claim. | 2 minutes |
Total time: Under 4 minutes. This catches most misinformation.
Source Credibility Checklist
| Factor | Strong Source | Weak Source |
|---|---|---|
| Track record | History of corrections and accountability | History of retractions, sensationalism |
| Transparency | Named authors, editorial process, funding disclosed | Anonymous, opaque ownership |
| Evidence | Cites primary sources, links to data | Vague attribution ("experts say," "studies show") |
| Tone | Measured, acknowledges complexity | Inflammatory, absolute, emotionally charged |
| Corrections | Issues corrections promptly | Never corrects or quietly edits |
| Expertise | Reporters with subject-matter knowledge | Generalists covering everything |
Media Bias Spectrum
Media bias exists on multiple axes, not just left-right:
| Axis | One End | Other End |
|---|---|---|
| Political | Left-leaning | Right-leaning |
| Sensationalism | Measured/dry | Clickbait/outrage |
| Depth | Deep analysis | Surface/headline only |
| Selection | Broad coverage | Cherry-picked stories |
| Framing | Multiple perspectives | Single narrative |
No source is unbiased. The goal isn't to find a perfectly neutral source. It's to understand each source's biases and compensate.
Practical Source Rating
| Tier | Characteristics | Examples |
|---|---|---|
| Tier 1: Primary | Original documents, raw data, direct observation | Government datasets, court filings, scientific papers |
| Tier 2: Quality reporting | Professional journalism with editorial standards | Major newspapers, wire services (AP, Reuters) |
| Tier 3: Analysis | Interpretation and opinion from qualified commentators | Think tank reports, expert blogs |
| Tier 4: Aggregation | Summaries of other sources | News aggregators, social media posts about articles |
| Tier 5: Unreliable | No editorial standards, anonymous, agenda-driven | Random blogs, partisan outlets, content farms |
Rule: Never form a strong opinion based solely on Tier 4-5 sources. Always trace back to Tier 1-2.
Identifying Bias
Types of Media Bias
| Bias Type | How It Works | Example |
|---|---|---|
| Selection bias | Choosing which stories to cover | Covering crimes by one group but not another |
| Framing bias | How a story is presented | "Tax relief" vs. "tax cuts for the wealthy" |
| Confirmation bias | Favoring information that confirms existing views | Only interviewing experts who agree with the angle |
| Omission bias | Leaving out important context | Reporting a statistic without the baseline |
| Placement bias | Prominence given to stories | Burying corrections on page 12 |
| Tone bias | Emotional language vs. neutral reporting | "Slammed," "destroyed," "eviscerated" in headlines |
| Source bias | Which voices are quoted | Only quoting one side's experts |
Framing: How the Same Facts Tell Different Stories
The same event, framed differently:
| Framing | Headline |
|---|---|
| Positive | "Unemployment Falls to 5%" |
| Negative | "1 in 20 Workers Still Can't Find a Job" |
| Alarming | "Unemployment Remains Above Pre-Pandemic Levels" |
| Dismissive | "Unemployment Rate Barely Budges" |
All four headlines can be factually accurate. The framing determines the emotional response and the perceived significance.
How to Detect Framing
- Rewrite the headline neutrally. What would a purely factual version say?
- Check for loaded language. Words like "admit," "claim," "refuse" carry implications
- Look for what's missing. What context would change your interpretation?
- Flip the frame. How would the other side describe the same event?
- Check the images. Photo selection heavily influences perception
Fact-Checking Techniques
Quick Verification Steps
| Step | Action | Tool |
|---|---|---|
| 1. Check the URL | Is it a legitimate domain? Watch for lookalikes (abcnews.com.co) | Browser address bar |
| 2. Check the date | Old stories often recirculate as "breaking news" | Article metadata |
| 3. Reverse image search | Is the photo from a different event/time? | Google Images, TinEye |
| 4. Check fact-checkers | Has this specific claim been evaluated? | Snopes, FactCheck.org, PolitiFact |
| 5. Find the primary source | Where did this claim originate? | Follow the citation chain |
| 6. Cross-reference | Do multiple independent sources report the same facts? | Search engine |
Lateral Reading
Don't evaluate a source by reading it deeply. Instead, open new tabs and see what others say about the source.
Vertical reading (ineffective): Reading the article carefully, checking the "About" page, examining the website design.
Lateral reading (effective): Immediately leaving the site to see what independent sources say about it.
Professional fact-checkers spend less than 30 seconds on a site before searching elsewhere for information about it.
Red Flags for Fake News
| Red Flag | What to Look For |
|---|---|
| Extreme emotional reaction | If a story makes you furious or ecstatic, verify before sharing |
| Too good to be true | Stories that perfectly confirm your worldview |
| No named sources | "Sources say" or "people familiar with" without specifics |
| No other outlets reporting it | Major news would be covered by multiple outlets |
| Satire misidentified | Check if the source is a satire/parody site |
| Manipulated images/video | Look for artifacts, inconsistencies, and reverse search |
| Outdated information | Real stories from years ago presented as current |
Headlines vs. Articles
Headlines are written to get clicks. They systematically misrepresent the content.
Common Headline Tricks
| Trick | Example | What the Article Actually Says |
|---|---|---|
| Omitting qualifiers | "Coffee Causes Cancer" | "One study in rats suggested a correlation at extreme doses" |
| Question headlines (Betteridge's Law) | "Is This the Cure for Aging?" | Almost certainly no |
| Cherry-picked quotes | "'Disaster' Says Expert" | Expert called one minor aspect "potentially disastrous if unchecked" |
| False urgency | "BREAKING: New Study Reveals..." | Study was published last month |
| Misleading numbers | "Risk DOUBLES" | From 1 in a million to 2 in a million |
Rule: Never form an opinion from a headline. Read the article. Then check the source it cites.
Social Media Information Quality
Why Social Media Is Unreliable
| Problem | Mechanism |
|---|---|
| Engagement algorithms | Outrage and novelty get promoted over accuracy |
| No editorial oversight | Anyone can post anything |
| Speed over accuracy | First beats correct |
| Echo chambers | Algorithms show you what you already believe |
| Emotional contagion | Emotional content spreads faster than factual content |
| Anonymity | No accountability for false claims |
| Bot networks | Artificial amplification of narratives |
Social Media Hygiene
| Practice | Why |
|---|---|
| Don't share without reading | Headlines are misleading |
| Don't share when emotional | That's exactly when you're most likely to share misinformation |
| Check the date and source | Old stories recirculate constantly |
| Be skeptical of screenshots | Easy to fabricate |
| Verify viral claims | The more viral, the more likely it's been distorted |
| Follow diverse sources | Break your echo chamber deliberately |
Propaganda Techniques
Propaganda isn't just wartime posters. It's used daily in politics, advertising, and media.
| Technique | How It Works | Example |
|---|---|---|
| Bandwagon | "Everyone else agrees" | "Join millions who already switched" |
| Name-calling | Labels to dismiss without argument | "That's just socialist/fascist/woke" |
| Glittering generalities | Vague positive words with no content | "Freedom," "innovation," "common sense solutions" |
| Transfer | Associate with respected symbols | Politicians with flags, celebrities with products |
| Testimonial | Celebrity or authority endorsement | "Doctors recommend..." |
| Card stacking | One-sided presentation of facts | Only showing favorable statistics |
| Plain folks | "I'm just like you" | Politicians eating at diners, using folksy language |
| Fear appeal | Exaggerating threats | "If we don't act NOW, everything will be destroyed" |
| Repetition | Say it enough times, people believe it | Repeating a slogan or talking point relentlessly |
Deepfakes and AI-Generated Content
The Problem
AI can now generate:
- Realistic images of events that never happened
- Video of people saying things they never said
- Convincing articles with fabricated quotes and sources
- Fake academic papers with plausible-sounding findings
Detection Strategies
| Content Type | What to Check |
|---|---|
| Images | Reverse image search; check hands, teeth, backgrounds for artifacts; look for inconsistent lighting/shadows |
| Video | Look for unnatural facial movements, lip sync issues, background distortion; check source |
| Text | Verify specific claims and quotes independently; AI text often lacks specific verifiable details |
| Audio | Verify with other recordings of the same person; check for unnatural cadence |
Principles for an AI-Content World
- Provenance matters more than ever. Where did this content originate?
- Verification is non-negotiable. If it matters, verify it independently
- Be skeptical of "perfect" evidence. Conveniently perfect proof may be fabricated
- Trust networks of sources, not individual pieces of content. One source can be faked; consistent reporting across many independent sources is much harder to fake
Building a Balanced Information Diet
Principles
| Principle | Practice |
|---|---|
| Diversity | Read sources from different perspectives |
| Quality over quantity | Better to read 2 good articles than 20 bad ones |
| Primary sources | Go to the original data/documents when possible |
| Slow consumption | Weekly/monthly analysis is more valuable than hourly updates |
| Active, not passive | Choose your sources; don't let algorithms choose for you |
Practical Information Diet
| Category | Approach |
|---|---|
| Breaking news | Wait 24-48 hours for facts to settle; initial reports are often wrong |
| Daily news | 1-2 quality sources with different perspectives |
| Weekly analysis | Long-form journalism, newsletters from subject experts |
| Deep dives | Books, academic papers, primary documents |
| Social media | Treat as entertainment/social, not as news |
The 24-Hour Rule
For any story that triggers a strong emotional reaction:
- Don't share it immediately
- Wait 24 hours
- Check if the story has been updated, corrected, or debunked
- If it still holds up after 24 hours, engage with it
This single habit prevents most misinformation sharing.
Key Takeaways
- You are your own editor. No one else will filter information for you
- Use SIFT. Stop, Investigate, Find better coverage, Trace claims
- Read laterally. Check what others say about a source, not just the source itself
- Headlines lie. Never form opinions from headlines alone
- Social media optimizes for engagement, not truth. Treat it accordingly
- No source is unbiased. Understand biases rather than seeking a mythical neutral source
- Wait 24 hours. Time kills most misinformation
- Verify before sharing. You are responsible for what you amplify