# Sampling Techniques

Psychology I · Research Techniques · https://tryals.app/learn/psychology-i/sampling-techniques

## Who Actually Gets Studied

A **population** is the entire group you want to understand, while a **sample** is the subset actually studied. Because studying everyone is impractical, how you sample determines if your findings **generalize** back to the population.

| Method | How it selects | Generalizability |
| --- | --- | --- |
| **Random** | Equal chance for all | Strong |
| **Stratified** | Proportional by subgroup | Strong, protected |
| **Convenience** | Whoever is easiest to reach | Weak |

**Random sampling** minimizes systematic bias. **Stratified sampling** divides the population into strata (age, gender) and samples proportionally. **Convenience sampling** (like psychology undergrads) is fast and cheap, but creates **sampling bias**.

## Validity and Sample Size

Poor sampling threatens **external validity**: findings from a biased sample may not hold for the broader population. This fuels the WEIRD-samples critique, where psychology risks over-generalizing from narrow slices of humanity.

> **Common pitfall:** assuming a LARGE sample automatically means a REPRESENTATIVE sample.

Sample size and sampling method are separate issues. A huge convenience sample of 50,000 self-selected online volunteers can still be badly biased, while a smaller, carefully drawn random sample generalizes far better.

## Practice questions

6 of this lesson's 11 practice questions, with answers. The full set is in the app.

### 1. Sort each described sample by whether it generalizes well to the broader population.

**Answer:**

- Strong generalizability: A carefully drawn random sample of 400 from the whole population, A stratified sample matching age and region proportions
- Weak generalizability: 30,000 self-selected online volunteers, Only intro-psych undergraduates at one university

**Why:** Random and stratified samples generalize well because their selection method controls bias; large self-selected or single-classroom convenience samples generalize poorly despite their size. Method, not headcount, drives external validity.

Page: https://tryals.app/practice/psychology-i/sampling-techniques/sort-each-described-sample-by-whether-it-generalizes-well-to-the

### 2. A population is 60% women and 40% men. Simple random sampling could, by chance, produce a sample that is 75% women. Stratified sampling cannot. Why is removing that possibility worth giving up pure randomness for?

A. Because stratified samples can be smaller, since each stratum needs fewer participants.
B. Because random sampling cannot be used when the population contains identifiable subgroups.
C. Because stratification guarantees the sample matches on a key variable.
D. Because random sampling is biased, and stratification corrects the bias.

**Answer:** C. Because stratification guarantees the sample matches on a key variable.

**Why:** Random sampling is unbiased in the long run and you only get one sample, which can still land at 75%. Stratification builds a fact you already know into the design so that error cannot occur. Its weakness is variance rather than bias, and it protects only the variables you stratify on, which is a judgement about what matters for the outcome.

Page: https://tryals.app/practice/psychology-i/sampling-techniques/a-population-is-60-women-and-40-men-simple-random-sampling-could

### 3. The 1936 *Literary Digest* poll collected 2.4 million responses and called the US election for the wrong man. Gallup got it right with a fraction of that. What went wrong?

A. Respondents lied about their voting intention in a way Gallup's did not
B. The sampling frame, which in the Depression meant the affluent
C. The sample was too large to process accurately with the technology available
D. Opinion shifted between the poll and the election

**Answer:** B. The sampling frame, which in the Depression meant the affluent

**Why:** The frame decided the answer before a single reply arrived: sampling from car and phone owners in 1936 sampled wealth. Size cannot repair exclusion, it only measures the wrong population more precisely, which is the most expensive way to be wrong.

Page: https://tryals.app/practice/psychology-i/sampling-techniques/the-1936-literary-digest-poll-collected-2-4-million-responses-and

### 4. Select every statement that is TRUE about sampling.

A. A biased sampling method is not repaired by collecting more data.
B. Convenience sampling is the strongest method for generalizing to a population.
C. Every member of the population has the same chance of being picked, so no group can be systematically left out.
D. A large sample size guarantees a representative sample.
E. Stratified sampling deliberately matches the sample's composition to the population's.

**Answer:** A. A biased sampling method is not repaired by collecting more data.; C. Every member of the population has the same chance of being picked, so no group can be systematically left out.; D. A large sample size guarantees a representative sample.

**Why:** Stratified proportionality, random sampling's equal chance, and convenience sampling's bias are all true. A large sample is not automatically representative, and size is not the same as method - those two false options are the exact confusion the lesson warns against.

Page: https://tryals.app/practice/psychology-i/sampling-techniques/select-every-statement-that-is-true-about-sampling

### 5. An online study recruits 50,000 self-selected volunteers. A rival study randomly samples 400 people from the target population. Why does the smaller study support better generalisation, despite having one hundred and twenty-five times fewer participants?

A. Because size controls random error and method controls systematic error.
B. Because the larger study cannot achieve statistical significance with so heterogeneous a sample.
C. Because 400 is closer to the ideal sample size, and samples that are too large introduce their own distortions.
D. Because online data collection is inherently less accurate than data collected in person.

**Answer:** A. Because size controls random error and method controls systematic error.

**Why:** Only one kind of wrongness responds to sample size: random error shrinks as n grows and systematic error does not, so 50,000 self-selected volunteers locate a wrong answer with impressive precision. The 1936 Literary Digest poll had 2.4 million responses and called the election for the wrong candidate; Gallup got it right with a fraction of that.

Page: https://tryals.app/practice/psychology-i/sampling-techniques/an-online-study-recruits-50-000-self-selected-volunteers-a-rival

### 6. A carefully drawn random sample of 400 people can generalize to the population better than a self-selected convenience sample of 40,000.

**Answer:** True

**Why:** True. A smaller random sample can generalize better than a much larger convenience sample, because generalizability depends on unbiased selection, not on size. A huge self-selected sample keeps its bias no matter how many people it contains.

Page: https://tryals.app/practice/psychology-i/sampling-techniques/a-carefully-drawn-random-sample-of-400-people-can-generalize-to-the