Data Literacy 101

📋 Overview

Develop practical command of data literacy mindset and decision quality with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply data literacy mindset and decision quality in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: distribution, sample bias, and variance
✓ Common mistakes to avoid when applying data literacy mindset and decision quality
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

distribution sample bias variance correlation causation base rate

💡 Examples

A restaurant owner who checks which dishes sell most on weekdays vs weekends is using data literacy.

Bad data decision: "Sales went up, so our ad campaign worked." (Did anything else change?)

Data literacy is not about being a statistician — it is about asking the right questions before trusting numbers.

Every number you see in the news has a story behind it. Data literacy helps you read that story critically.

💬 Mini-Dialogue

Mentor: Your friend says "I read that 90% of doctors recommend Brand X." Do you believe it?

Learner: That sounds convincing.

Mentor: What if the survey asked only 10 doctors who were paid by Brand X? The number is real, but the data behind it is garbage.

Learner: So the percentage alone isn't enough?

Mentor: Never. Always ask: Who collected this? How? From how many people? With what incentives? THEN decide if the number means anything.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through data literacy mindset and decision quality scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on data literacy mindset and decision quality
Create 5 examples that use distribution, sample bias, and variance
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Data Literacy Mindset and Decision Quality" for one minute and include two concrete examples.

❓ Quick Quiz

1. Data literacy is primarily about:

2. A percentage without context is:

3. The first question to ask about any statistic is:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of types of data and measurement scales with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply types of data and measurement scales in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: sample bias, variance, and correlation
✓ Common mistakes to avoid when applying types of data and measurement scales
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

sample bias variance correlation causation base rate confidence

💡 Examples

Nominal: Categories with no order (eye color, country, blood type).

Ordinal: Categories with order but unequal gaps (movie ratings: 1-5 stars).

Interval: Equal gaps, no true zero (temperature in Celsius — 0 C is not "no temperature").

Ratio: Equal gaps with true zero (height, weight, income — 0 means none).

💬 Mini-Dialogue

Mentor: Is "customer satisfaction: Very Happy / Happy / Neutral / Unhappy" nominal or ordinal?

Learner: Ordinal — there's a clear order from happy to unhappy.

Mentor: Right. But can you average it? Is the gap between "Happy" and "Neutral" the same as between "Neutral" and "Unhappy"?

Learner: Probably not exactly…

Mentor: Exactly. Ordinal data has order but not equal intervals. Averaging it can be misleading — that's a common trap in surveys.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through types of data and measurement scales scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on types of data and measurement scales
Create 5 examples that use sample bias, variance, and correlation
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Types of Data and Measurement Scales" for one minute and include two concrete examples.

❓ Quick Quiz

1. Eye color is what type of data?

2. Temperature in Celsius is:

3. Why is averaging ordinal data risky?

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of reading tables and summary statistics with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply reading tables and summary statistics in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: variance, correlation, and causation
✓ Common mistakes to avoid when applying reading tables and summary statistics
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

variance correlation causation base rate confidence outlier

💡 Examples

Mean: Sum divided by count. Sensitive to outliers — one billionaire in a room skews the "average income."

Median: The middle value when sorted. Resistant to outliers — better for income, house prices.

Mode: Most frequent value. Useful for categories: "The most common shoe size sold is 42."

Range vs IQR: Range = max minus min (includes outliers). IQR = middle 50% (ignores extremes).

💬 Mini-Dialogue

Mentor: Average salary at this company is $250,000. Sounds great, right?

Learner: Very!

Mentor: The CEO makes $4 million. The other 19 employees make $50,000 each. The mean is $247,500 — technically correct but completely misleading.

Learner: The median would be $50,000 — that's the real picture.

Mentor: Always ask: mean or median? One billionaire or one extreme value can make a mean lie to you.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through reading tables and summary statistics scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on reading tables and summary statistics
Create 5 examples that use variance, correlation, and causation
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Reading Tables and Summary Statistics" for one minute and include two concrete examples.

❓ Quick Quiz

1. When data has extreme outliers, which measure is most reliable?

2. The mean is sensitive to:

3. IQR represents:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of visual encoding: why charts work with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply visual encoding: why charts work in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: correlation, causation, and base rate
✓ Common mistakes to avoid when applying visual encoding: why charts work
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

correlation causation base rate confidence outlier dashboard

💡 Examples

Bar charts compare quantities across categories: "Sales by region."

Line charts show change over time: "Temperature trend over 12 months."

Pie charts show parts of a whole: "Market share by company." (Best with 3-5 slices max.)

The right chart makes data obvious. The wrong chart makes it confusing — or deceptive.

💬 Mini-Dialogue

Mentor: You used a pie chart with 15 slices. Can you read it?

Learner: Not really — they all look similar.

Mentor: Pie charts work for 3-5 categories max. For 15 categories, use a horizontal bar chart sorted by size. The eye can compare bar lengths much more easily than pie slices.

Learner: When should I use a line chart?

Mentor: When time is on the X-axis. Lines imply continuity over time. Never use a line chart for categories — that implies a trend that doesn't exist.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through visual encoding: why charts work scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on visual encoding: why charts work
Create 5 examples that use correlation, causation, and base rate
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Visual Encoding: Why Charts Work" for one minute and include two concrete examples.

❓ Quick Quiz

1. A line chart is best for:

2. Pie charts work best with:

3. Using a line chart for categorical data is wrong because:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of bar, line, scatter, and histogram basics with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply bar, line, scatter, and histogram basics in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: causation, base rate, and confidence
✓ Common mistakes to avoid when applying bar, line, scatter, and histogram basics
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

causation base rate confidence outlier dashboard decision memo

💡 Examples

Histogram: Shows distribution shape — is data clustered, spread out, or skewed?

Scatter plot: Shows relationship between two variables. Tight cluster = strong relationship.

Bar chart vs histogram: Bar = categories (gaps between bars). Histogram = continuous data (no gaps).

Use scatter plots to spot patterns: positive trend, negative trend, no trend, or clusters.

💬 Mini-Dialogue

Mentor: You want to know if hours studied relates to exam scores. Which chart?

Learner: A bar chart?

Mentor: A scatter plot. Each dot is one student: X-axis = hours studied, Y-axis = exam score. If dots trend upward, more study means better scores.

Learner: What if the dots are all over the place?

Mentor: Then there's no clear relationship — which is also valuable information! The absence of a pattern is data too.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through bar, line, scatter, and histogram basics scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on bar, line, scatter, and histogram basics
Create 5 examples that use causation, base rate, and confidence
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Bar, Line, Scatter, and Histogram Basics" for one minute and include two concrete examples.

❓ Quick Quiz

1. A scatter plot shows:

2. The difference between a bar chart and histogram is:

3. A tight upward cluster in a scatter plot suggests:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of distribution, variability, and outliers with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply distribution, variability, and outliers in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: base rate, confidence, and outlier
✓ Common mistakes to avoid when applying distribution, variability, and outliers
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

base rate confidence outlier dashboard decision memo distribution

💡 Examples

Normal distribution: Most values cluster around the mean (bell curve). Heights, test scores often follow this.

Skewed right: A long tail to the right. Income is right-skewed — most people earn modest amounts, a few earn millions.

Standard deviation: How spread out values are. Low SD = tightly packed. High SD = widely spread.

Outlier: A value far from the rest. Not always an error — could be a genuine extreme (like a billionaire in income data).

💬 Mini-Dialogue

Mentor: Test scores: most students scored 70-80, but one scored 15 and one scored 99. What happened?

Learner: The 15 might have guessed randomly, and the 99 really knew the material?

Mentor: Possibly. They're outliers. The question is: do we include them in our analysis? They heavily affect the mean but barely touch the median.

Learner: So we should report both mean and median?

Mentor: Smart. Reporting both tells the full story. If they're very different, outliers are pulling the mean.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through distribution, variability, and outliers scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on distribution, variability, and outliers
Create 5 examples that use base rate, confidence, and outlier
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Distribution, Variability, and Outliers" for one minute and include two concrete examples.

❓ Quick Quiz

1. Income data is typically:

2. A low standard deviation means:

3. An outlier should be:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of correlation vs causation with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply correlation vs causation in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: confidence, outlier, and dashboard
✓ Common mistakes to avoid when applying correlation vs causation
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

confidence outlier dashboard decision memo distribution sample bias

💡 Examples

Ice cream sales and drowning deaths both increase in summer. Ice cream doesn't cause drowning — the sun does.

Correlation coefficient (r): Ranges from -1 to +1. Close to 0 = no linear relationship.

Hidden third variable (confounder): Both shoe size and reading ability increase with age in children. Age is the confounder.

To establish causation you need: correlation + temporal order + no confounders + a plausible mechanism.

💬 Mini-Dialogue

Mentor: "Countries that eat more chocolate win more Nobel Prizes." Causation?

Learner: Probably not… but what's the connection?

Mentor: Wealth. Rich countries can afford both chocolate and research funding. Wealth is the confounder connecting two unrelated things.

Learner: So every correlation should be interrogated for hidden third variables?

Mentor: Every single one. "What else could explain this?" is the most powerful question in data literacy.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through correlation vs causation scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on correlation vs causation
Create 5 examples that use confidence, outlier, and dashboard
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Correlation vs Causation" for one minute and include two concrete examples.

❓ Quick Quiz

1. Ice cream sales and drowning are correlated because:

2. A correlation of r = 0.02 indicates:

3. To claim causation, you need:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of sampling, bias, and representativeness with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply sampling, bias, and representativeness in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: outlier, dashboard, and decision memo
✓ Common mistakes to avoid when applying sampling, bias, and representativeness
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

outlier dashboard decision memo distribution sample bias variance

💡 Examples

Sampling bias: Polling only smartphone users about poverty excludes people who can't afford smartphones.

Selection bias: A gym surveys "customer satisfaction" — only current members respond, not people who quit.

Response bias: People underreport embarrassing behaviors (drinking, screen time) and overreport good ones (exercise, reading).

Survivorship bias: "College dropouts who became billionaires" ignores millions of dropouts who didn't.

💬 Mini-Dialogue

Mentor: A study finds that people who eat breakfast are healthier. Should you start eating breakfast?

Learner: Yes, obviously?

Mentor: Who volunteers for breakfast studies? Health-conscious people who also exercise and sleep well. The breakfast might be correlated with health but caused by an overall healthy lifestyle.

Learner: So the sample is biased toward health-conscious people?

Mentor: Self-selection bias. The people who sign up for the study aren't representative of the general population.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through sampling, bias, and representativeness scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on sampling, bias, and representativeness
Create 5 examples that use outlier, dashboard, and decision memo
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Sampling, Bias, and Representativeness" for one minute and include two concrete examples.

❓ Quick Quiz

1. Survivorship bias means:

2. A gym surveying only current members has:

3. Response bias occurs when people:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of uncertainty, error bars, and confidence with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply uncertainty, error bars, and confidence in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: dashboard, decision memo, and distribution
✓ Common mistakes to avoid when applying uncertainty, error bars, and confidence
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

dashboard decision memo distribution sample bias variance correlation

💡 Examples

Margin of error: "The candidate leads 52% to 48%, +/- 3%." This means it could actually be 49% to 51% — a tie!

Confidence interval: "We are 95% confident the true value lies between 42 and 58."

p-value: The probability of seeing this result if there were actually no effect. Below 0.05 is conventionally "significant."

Statistical significance does not mean practical importance. A drug that works but improves outcomes by 0.1% may not be worth the cost.

💬 Mini-Dialogue

Mentor: "This result is statistically significant (p = 0.04)." Does that mean it's important?

Learner: Yes? Significant means important, right?

Mentor: In statistics, "significant" means "unlikely to be due to chance." It says nothing about size or importance. A massive study could find that a pill raises happiness by 0.001 points — statistically significant but practically meaningless.

Learner: So I need to check the effect SIZE, not just the p-value?

Mentor: Always. Significance tells you it's real. Effect size tells you if it matters.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through uncertainty, error bars, and confidence scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on uncertainty, error bars, and confidence
Create 5 examples that use dashboard, decision memo, and distribution
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Uncertainty, Error Bars, and Confidence" for one minute and include two concrete examples.

❓ Quick Quiz

1. "Statistically significant" means:

2. A margin of error of +/- 3% on a poll showing 51% vs 49% means:

3. Effect size tells you:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of common misleading graph techniques with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply common misleading graph techniques in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: decision memo, distribution, and sample bias
✓ Common mistakes to avoid when applying common misleading graph techniques
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

decision memo distribution sample bias variance correlation causation

💡 Examples

Truncated Y-axis: Starting at 98 instead of 0 makes a 2% change look like a 50% drop.

Cherry-picked time frame: Showing stock growth from its lowest point to its peak.

3D pie charts: The "front" slice looks larger due to perspective distortion. Always use 2D.

Dual Y-axes: Two different scales on the same chart can make any two trends look correlated.

💬 Mini-Dialogue

Mentor: This chart shows "Crime surging!" with a dramatic upward slope. Look at the Y-axis.

Learner: It starts at 195 and goes to 210. So the actual increase is from 195 to 210 cases?

Mentor: A 7.7% increase. But by starting the axis at 195 instead of 0, it looks like a 300% surge. The chart is technically accurate but visually deceptive.

Learner: So I should always check the axis ranges?

Mentor: Always. That's the #1 chart manipulation trick. It turns molehills into mountains.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through common misleading graph techniques scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on common misleading graph techniques
Create 5 examples that use decision memo, distribution, and sample bias
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Common Misleading Graph Techniques" for one minute and include two concrete examples.

❓ Quick Quiz

1. A truncated Y-axis can make:

2. 3D pie charts are problematic because:

3. The first thing to check on any chart is:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of percentages, rates, and base rates with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply percentages, rates, and base rates in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: distribution, sample bias, and variance
✓ Common mistakes to avoid when applying percentages, rates, and base rates
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

distribution sample bias variance correlation causation base rate

💡 Examples

"Sales increased by 100%!" From 1 sale to 2 sales. Always ask for the absolute numbers.

Base rate neglect: "This test is 99% accurate" sounds great, but if the disease affects 1 in 10,000 people, most positive results are false positives.

"Per capita" vs "total": India has more English speakers than the UK in total, but a much lower per capita rate.

Percentages of percentages: "Crime decreased by 50%, then increased by 50%." You're NOT back where you started.

💬 Mini-Dialogue

Mentor: "Our new feature increased signups by 200%!" Impressive?

Learner: That sounds huge!

Mentor: Signups went from 3 to 9. In absolute terms, 6 more people. The percentage is real but the impact is tiny.

Learner: So I should always ask "200% of WHAT?"

Mentor: The base rate is everything. A 200% increase from a small base is still small. A 5% increase from a massive base can be enormous.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through percentages, rates, and base rates scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on percentages, rates, and base rates
Create 5 examples that use distribution, sample bias, and variance
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Percentages, Rates, and Base Rates" for one minute and include two concrete examples.

❓ Quick Quiz

1. "A 200% increase" is meaningless without:

2. If crime drops 50% then rises 50%, the final level is:

3. Base rate neglect causes people to:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of interpreting headlines with numbers with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply interpreting headlines with numbers in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: sample bias, variance, and correlation
✓ Common mistakes to avoid when applying interpreting headlines with numbers
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

sample bias variance correlation causation base rate confidence

💡 Examples

Headline: "Coffee linked to longer life!" Reality: Observational study, confounders not controlled, small effect size.

Headline: "Study finds screen time is safe for kids!" Reality: Study funded by a tech company.

Always check: Who did the study? How many people? Was it peer-reviewed? What was the effect size?

The media translates "small correlation in one study" into "X causes Y!" Be the person who checks.

💬 Mini-Dialogue

Mentor: Headline: "New study proves chocolate prevents heart disease!" What do you check?

Learner: Um… the study?

Mentor: Specifically: 1) Was it a randomized controlled trial or just observational? 2) How many people? 3) Who funded it? 4) What was the actual effect size?

Learner: And "prevents" is a strong word.

Mentor: The study probably says "associated with lower risk." The headline converted a correlation into a causal claim. That's the media translation problem.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through interpreting headlines with numbers scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on interpreting headlines with numbers
Create 5 examples that use sample bias, variance, and correlation
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Interpreting Headlines with Numbers" for one minute and include two concrete examples.

❓ Quick Quiz

1. Media headlines about studies often:

2. A study funded by a company that benefits from the results has:

3. Before believing a health headline, always check:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of experiment design and a/b testing basics with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply experiment design and a/b testing basics in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: variance, correlation, and causation
✓ Common mistakes to avoid when applying experiment design and a/b testing basics
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

variance correlation causation base rate confidence outlier

💡 Examples

A/B testing: Show version A to half the users and version B to the other half. Measure the difference.

Control group: The group that doesn't get the intervention. Without it, you can't isolate the effect.

Random assignment: Ensures the groups are similar before the test. Prevents selection bias.

Sample size matters: Testing on 10 people is unreliable. Testing on 10,000 reveals real patterns.

💬 Mini-Dialogue

Mentor: Your team changed the button color from blue to green and sales went up. The green button works, right?

Learner: Obviously!

Mentor: Was it an A/B test — green for half, blue for the other half, at the same time? Or did you just switch everyone?

Learner: We switched everyone on Monday.

Mentor: Then maybe sales went up because it was a new week, or because of a promotion, or seasonal trends. Without a control group running simultaneously, you can't isolate the button color effect.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through experiment design and a/b testing basics scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on experiment design and a/b testing basics
Create 5 examples that use variance, correlation, and causation
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Experiment Design and A/B Testing Basics" for one minute and include two concrete examples.

❓ Quick Quiz

1. A/B testing requires:

2. A control group is needed to:

3. Random assignment prevents:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of building simple dashboards for insight with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply building simple dashboards for insight in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: correlation, causation, and base rate
✓ Common mistakes to avoid when applying building simple dashboards for insight
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

correlation causation base rate confidence outlier dashboard

💡 Examples

A dashboard should answer ONE question at a glance: "Are we on track this month?"

KPI (Key Performance Indicator): The 1-3 numbers that matter most. Revenue, conversion rate, customer satisfaction.

Traffic light coding: Green = on target, Yellow = watch, Red = action needed. Instant readability.

Less is more: A 3-chart dashboard beats a 12-chart dashboard. Every chart should earn its space.

💬 Mini-Dialogue

Mentor: Your dashboard has 15 charts. How quickly can you answer: "Are sales on track?"

Learner: I'd need to look at several charts and do some mental math…

Mentor: Too slow. A good dashboard answers that in 3 seconds. One big number showing sales vs target, color-coded green/yellow/red. Supporting details below, only if needed.

Learner: So the dashboard is for decisions, not data exploration?

Mentor: Exactly. Dashboard = at-a-glance decisions. Detailed analysis is a separate tool.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through building simple dashboards for insight scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on building simple dashboards for insight
Create 5 examples that use correlation, causation, and base rate
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Building Simple Dashboards for Insight" for one minute and include two concrete examples.

❓ Quick Quiz

1. A dashboard should answer key questions in:

2. KPI stands for:

3. A good dashboard has:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of communicating data to non-experts with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply communicating data to non-experts in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: causation, base rate, and confidence
✓ Common mistakes to avoid when applying communicating data to non-experts
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

causation base rate confidence outlier dashboard decision memo

💡 Examples

Lead with the conclusion: "We should expand to Market B. Here's why." Not "Here's 30 slides of data."

The "So What?" bridge: After every data point, state the implication for the audience.

Use comparisons: "Our churn rate is 5%" means nothing. "Our churn rate is 5% vs industry average of 8%" means something.

Avoid jargon: "Standard deviation" -> "how spread out the values are." Match your audience's vocabulary.

💬 Mini-Dialogue

Mentor: You're presenting quarterly data to executives. How do you start?

Learner: With the methodology?

Mentor: Executives want answers, not methodology. Start with: "Revenue grew 12%, beating our 8% target. Here are three drivers." Methodology goes in the appendix.

Learner: What if they have questions about the data?

Mentor: They'll ask. Then you pull up the supporting detail. Lead with insight, follow with evidence. Not the reverse.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through communicating data to non-experts scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on communicating data to non-experts
Create 5 examples that use causation, base rate, and confidence
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Communicating Data to Non-Experts" for one minute and include two concrete examples.

❓ Quick Quiz

1. When presenting data, lead with:

2. "Our rate is 5%" is improved by adding:

3. For non-expert audiences, you should:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📋 Overview

Develop practical command of capstone: data-driven recommendation memo with short drills, retrieval checks, and real-world transfer tasks.

✅ Can-Do Goal

I can apply capstone: data-driven recommendation memo in a practical task with confidence.

🎯 Key Points

✓ Core principle: Interpret quantitative signals without overconfidence or narrative distortion
✓ High-frequency pattern set: base rate, confidence, and outlier
✓ Common mistakes to avoid when applying capstone: data-driven recommendation memo
✓ A compact review loop to move this lesson into long-term memory

🧩 Key Terms

base rate confidence outlier dashboard decision memo distribution

💡 Examples

Your capstone: Pick a real question, find relevant data, analyze it, and write a 1-page recommendation memo.

Structure: 1) Question, 2) Data sources, 3) Key findings (with charts), 4) Recommendation, 5) Limitations.

Example: "Should our campus cafeteria add a vegan option?" Analyze survey data, sales data, competitor offerings.

A good recommendation acknowledges uncertainty: "The data suggests X, though Y could affect the outcome."

💬 Mini-Dialogue

Mentor: For your capstone, choose a question that matters to you personally.

Learner: "Should I switch my major to computer science?"

Mentor: Perfect. What data would help you decide?

Learner: Salary data by major, job market trends, my grades in CS vs my current major, interest surveys?

Mentor: Now analyze it honestly — including data that might argue AGAINST switching. A strong memo presents both sides and explains why you still recommend one path.

🧠 Core Concept

Interpret quantitative signals without overconfidence or narrative distortion. Apply it directly through capstone: data-driven recommendation memo scenarios.

✏️ Practice Exercises

Complete a 10-minute focused drill on capstone: data-driven recommendation memo
Create 5 examples that use base rate, confidence, and outlier
Run a self-check and record one improvement for the next session

🗣️ Speaking Prompt

Explain or demonstrate "Capstone: Data-Driven Recommendation Memo" for one minute and include two concrete examples.

❓ Quick Quiz

1. A data-driven recommendation memo should include:

2. Acknowledging limitations makes your recommendation:

3. The best capstone question is:

🔎 Practical Insight

Always ask what data is missing before deciding what the data means.

Launch Interactive Lesson →

📚 Course Lessons

🎯 Key Points

🧩 Key Terms

💡 Examples

💬 Mini-Dialogue

🧠 Core Concept

✏️ Practice Exercises

🗣️ Speaking Prompt

❓ Quick Quiz

🎯 Key Points

🧩 Key Terms

💡 Examples

💬 Mini-Dialogue

🧠 Core Concept

✏️ Practice Exercises

🗣️ Speaking Prompt

❓ Quick Quiz

🎯 Key Points

🧩 Key Terms

💡 Examples

💬 Mini-Dialogue

🧠 Core Concept

✏️ Practice Exercises

🗣️ Speaking Prompt

❓ Quick Quiz

🎯 Key Points

🧩 Key Terms

💡 Examples

💬 Mini-Dialogue

🧠 Core Concept

✏️ Practice Exercises

🗣️ Speaking Prompt

❓ Quick Quiz

🎯 Key Points

🧩 Key Terms

💡 Examples

💬 Mini-Dialogue

🧠 Core Concept

✏️ Practice Exercises

🗣️ Speaking Prompt

❓ Quick Quiz

🎯 Key Points

🧩 Key Terms

💡 Examples

💬 Mini-Dialogue

🧠 Core Concept

✏️ Practice Exercises

🗣️ Speaking Prompt

❓ Quick Quiz

🎯 Key Points

🧩 Key Terms

💡 Examples

💬 Mini-Dialogue

🧠 Core Concept

✏️ Practice Exercises

🗣️ Speaking Prompt

❓ Quick Quiz

🎯 Key Points

🧩 Key Terms

💡 Examples

💬 Mini-Dialogue

🧠 Core Concept

✏️ Practice Exercises

🗣️ Speaking Prompt

❓ Quick Quiz

🎯 Key Points

🧩 Key Terms

💡 Examples

💬 Mini-Dialogue

🧠 Core Concept

✏️ Practice Exercises

🗣️ Speaking Prompt

❓ Quick Quiz

🎯 Key Points

🧩 Key Terms

💡 Examples

💬 Mini-Dialogue

🧠 Core Concept

✏️ Practice Exercises