Statistics as a Process for Inference

Lesson 1 of 2: Foundations

In this lesson:

• Name the population, parameter, sample, and statistic
• Run the four-step inference loop on real data

Learning Objectives for This Unit

By the end of this two-lesson unit, you should be able to:

1. Distinguish population, sample, parameter, and statistic
2. Describe statistics as a four-step inference process
3. Explain why a random sample is necessary
4. Recognize sampling variability across samples
5. State what one sample can and cannot conclude

A Question You Can't Fully Answer

Lincoln High has 2,400 students.

What fraction of them would re-enroll in the elective if it were offered again?

• Nobody is going to ask all 2,400
• Yet we still want a trustworthy answer

The true fraction is real — but out of reach. So what do we do?

We Can Only Ask a Few

We randomly ask 60 students. 39 say yes.

• This number we can compute — it's right here
• The whole-school fraction we can't — it's hidden

Two numbers: the one we want, and the one we have. They need names.

The Whole Group and a Part

• Population: the whole group — all 2,400 students
• Sample: the part we observed — the 60 asked

The Truth and the Number We Have

• Parameter: a fixed fact about the population
  • the true proportion who would re-enroll — unknown
• Statistic: a number computed from the sample
  • — known, we just calculated it

The parameter is the truth we want; the statistic is the number we have.

The Asymmetry That Drives Everything

One fixed truth we can't see; one visible number that depends on which sample we drew.

Notation That Sticks From Day One

• Plain/Greek = population truth; hat/bar = sample estimate

Label This Sleep Scenario Yourself

A district wants the average sleep hours for all its students. They survey 200 students and find a mean of 7.1 hours.

Identify all four:

• Population? Parameter? Sample? Statistic?

This is a mean context — think and . Commit to all four before advancing.

Four Objects Are One Machine

Check: population = all district students; parameter = ; sample = the 200; statistic = .

• These four are moving parts, not a list to memorize
• Sample → statistic → estimates → parameter

You've named the parts. Now watch the machine run.

Statistics Is a Four-Step Loop

1. Pose a question → 2. Sample → 3. Compute → 4. Infer back

Walk the Loop: Question and Sample

A town has 50,000 workers. What is their mean commute time ?

• Step 1 — Question: estimate , the true mean commute (unknown)
• Step 2 — Sample: randomly select 80 workers

We've named the truth we want and pulled a fair slice. Now we measure.

Walk the Loop: Compute and Infer

From the previous slide: 80 workers sampled.

• Step 3 — Statistic: their mean is minutes
• Step 4 — Infer: conclude is around 27 minutes

We reasoned from the sample we have back to the population we want.

Which Way Does the Arrow Run?

In probability, you know the population and predict the sample.

In inference, which direction does the reasoning go?

• A. Population → sample
• B. Sample → population

Commit to A or B before advancing.

Probability and Inference Run Opposite

• Probability: population → sample (predict the data)
• Inference: sample → population (estimate the truth)

Step 4 Is an Estimate, Not Proof

The parameter is unknown — so step 4 is a reasoned estimate.

• Algebra gives one certified answer
• Inference gives "around 27, give or take"

This uncertainty isn't a weakness — it's the truth about reasoning from a part to a whole.

Which Step Carries the Doubt?

The four steps: question → sample → compute → infer.

Which one is uncertain — and why?

The arithmetic in step 3 is exact. So where does "give or take" enter?

Your Turn: Two Full Scenarios

Label population, parameter, sample, statistic — with symbols.

1. A factory checks 400 bulbs; 12 are defective ()
2. A biologist weighs 50 fish from a lake ()

Do both before advancing. Proportion uses ; mean uses .

Key Takeaways From Lesson One

✓ Population = whole group; sample = the part observed
✓ Parameter = fixed unknown truth; statistic = computed estimate
✓ Inference runs sample → population, a four-step loop

Watch out: estimates — it is not
Watch out: inference reverses the probability arrow

Coming Up Next: Why Random Matters

In Lesson 2, you'll find out why the sample must be random — and what breaks when it isn't.

You'll also see why even a perfect random sample can never give you a certain answer.