Planning a Study
Introducido| English | Significado |
|---|---|
| observational study/ɒbzəˈveɪʃənl ˈstʌdi/ | estudio observacional |
| experiment/ekˈsperɪmənt/ | experimento |
| cause-and-effect/kɔːz ænd ɪˈfekt/ | causa y efecto |
| sample/ˈsæmpl/ | muestra |
| confounding variable/kənˈfaʊndɪŋ ˈveərɪəbl/ | Variable de confusión |
Watching vs. doing
- Two ways to gather data: an observational study 观察性研究 or an experiment 实验.
- In an observational study you watch and measure — you don't change anything.
- In an experiment you impose a treatment and see what happens.
- The difference decides whether you can claim cause-and-effect 因果关系.
Population and sample
- The population is the whole group you want to learn about.
- The sample 样本 is the smaller part you actually study.
- You measure the sample, then hope to generalize back to the population.
- Good design keeps the sample a fair stand-in for the population.

Simple random and stratified and cluster and systematic samples of the same population
Randomized experiments support cause
- To show $X$ causes $Y$, you must actively assign $X$ and hold other things comparable.
- That's what an experiment does — and why only it can establish cause-and-effect.
- An observational study can reveal an association but never prove the cause.
- If you didn't control who got the treatment, you can't rule out other explanations.
The confounding trap
- In an observational study, a confounding variable 混杂变量 can create a misleading link.
- A confounder is tied to both the explanatory and the response variable.
- It makes two things look causally linked when a third factor drives both.
- This is exactly why "observed association" ≠ "cause."
The single most important distinction in Unit 3: only a well-designed experiment (with random assignment) can establish cause-and-effect. An observational study — however large or careful — can show association but is always open to a confounding variable, so it cannot prove that the explanatory variable caused the response.
Do vitamins improve health?
- Observational: compare people who choose to take vitamins vs. those who don't. Vitamin-takers may also exercise more — a confounder. No causal claim.
- Experiment: randomly assign people to vitamins or placebo. Random assignment tends to balance exercise across groups; chance imbalances remain, so assess whether the response difference exceeds chance variation.
An observational study watches without intervening; an experiment imposes a treatment. You study a sample to learn about a population. Only an experiment with random assignment can establish cause-and-effect — an observational study is always vulnerable to a confounding variable.
Write from this case
- A study randomly assigns a tutoring method.
- Another study only surveys students who already chose a method.
One study randomly assigns a tutoring method; another surveys students who already chose a method. Which design can support a causal conclusion, and why?
Random assignment is the experimental route to cause.
One study randomly assigns tutoring; another surveys existing tutoring choices. Which design can support a causal conclusion?
Dificultad 1/5 · 100 monedas + un pequeño bono en el primer intento de hoy
Random assignment helps separate the tutoring effect from confounding, allowing causal inference after accounting for chance variation. Observing existing choices alone does not.
A sample drawn from a population
The sample is a slice we study to learn about the whole.
Which type of study can establish a cause-and-effect relationship?
Dificultad 1/5 · 100 monedas + un pequeño bono en el primer intento de hoy
Only random assignment in an experiment supports causation.
In an observational study, the researcher imposes a treatment on the subjects.
Dificultad 1/5 · 100 monedas + un pequeño bono en el primer intento de hoy
Observational = watch and measure without intervening; experiments impose treatments.
The whole group you want to learn about is the ___.
Dificultad 1/5 · 100 monedas + un pequeño bono en el primer intento de hoy
You study a sample to learn about the population.
Vitamin-takers also exercise more, muddying a study of vitamins and health. Exercise is a...
Dificultad 1/5 · 100 monedas + un pequeño bono en el primer intento de hoy
It's tied to both taking vitamins and to health — a confounder.
Which are true of experiments (vs. observational studies)?
Dificultad 2/5 · 140 monedas + un pequeño bono en el primer intento de hoy
Experiments impose treatments and, with randomization, support causation.