Experiments & Their Design

You’ll be able to

Identify the treatments, experimental units, explanatory and response variables
Apply the three principles of experimental design: control, randomization, replication
Explain the purpose of a control group and a placebo

Observational study vs. experiment

In an observational study we measure variables without intervening — we simply watch. In an experiment we deliberately impose treatments on experimental units and observe the response. Only an experiment, by actively assigning treatments, can support a cause-and-effect conclusion. The explanatory variable in an experiment is often called the factor, and its values are the treatments.

The three principles

Good experiments rest on three pillars. Control holds other variables constant (or uses a comparison/control group) so they cannot explain differences. Randomization assigns units to treatments by chance, balancing out lurking variables. Replication uses enough experimental units in each group so that real effects can be distinguished from chance variation. Together these let us attribute differences in the response to the treatments.

Control groups, placebos, and blinding

A control group provides a baseline for comparison — it may receive no treatment or a placebo, a fake treatment that looks real. The placebo effect is a genuine response to the idea of treatment, so a placebo control isolates the treatment’s true effect. Blinding keeps subjects (single-blind) and/or those measuring the response (double-blind) unaware of who got what, preventing expectations from biasing the results.

Completely randomized design

experimental units → RANDOM assignment → treatment groups → compare responses

The defining feature is that chance alone decides which unit gets which treatment. Random assignment is what creates comparable groups and permits a causal conclusion.

Worked example

Researchers test whether a new fertilizer increases tomato yield. They have 40 plots of land. Outline a completely randomized experiment.

1.Identify units and treatments: 40 plots are the experimental units; the treatments are "new fertilizer" and "no fertilizer (control)."
2.Randomly assign the 40 plots to the two treatments — for example, randomly select 20 plots to receive the new fertilizer and the other 20 to serve as controls.
3.Apply the treatments while keeping all other conditions (water, sunlight, soil prep) as similar as possible (control).
4.After the growing season, measure yield in each plot and compare the mean yield of the two groups.

Answer: Randomly split the 40 plots into a fertilizer group (20) and a control group (20), treat all plots identically except for the fertilizer, then compare mean yields. Randomization plus a control group lets any yield difference be attributed to the fertilizer.

Checkpoint

What is the primary purpose of randomly assigning subjects to treatment groups in an experiment?

On the exam

Keep two "randoms" distinct: random selection (sampling) lets you generalize to a population; random assignment (experiments) lets you conclude causation. A study can have one, both, or neither — and that determines exactly what conclusions are allowed.

Checkpoint

In a drug trial, neither the patients nor the doctors measuring symptoms know who received the real drug versus the placebo. This design is described as:

Answer the 2 checkpoints as you read.