The first documented experimental design […] The king assigned them a daily amount of food and wine from the king’s table. They [the soldiers] were to be trained for three years, and after that they were to enter the king’s service. Daniel resolved not to defile himself with the royal food and wine, and he asked the chief official for permission not to defile himself this way […] but the official told Daniel, “I am afraid of my lord the king, who has assigned your food and drink. Why should he see you looking worse than the other young men your age? The king would then have my head because of you.” Daniel then said to the guard  “Please test your servants for ten days: Give us nothing but vegetables to eat and water to drink. Then compare our appearance with that of the young men who eat the royal food, and treat your servants in accordance with what you see.” So he agreed to this and tested them for ten days. At the end of the ten days they looked healthier and better nourished than any of the young men who ate the royal food. So the guard took away their choice food and the wine they were to drink and gave them vegetables instead. Daniel 1:1-16

OBSERVATIONAL STUDY Observational study: researcher observes and records behavior. No attempt to impose treatment or change environment An observational study does NOT help prove causation.

Experimental design TREATMENT A well-designed experiment design imposes a treatment to the experimental units (if humans they are called subjects). Idea: to determine the extent to which treatments (explanatory variable(s)) affect outcomes (response variable(s)). Experiments can help prove causation

Experimental design: CONTROL A well-designed experiment controls for confounding variables (Statistical control = the act of holding variables not under study which could influence the outcome)

Experimental design: RANDOMIZATION A well-designed experiment randomizes. The goal is to use chance (=randomization) to make sure that the groups being studied differ only in the effect of the treatment variable.

Experimental design: REPLICATION A well-designed experimental study replicates (= repeats) the experiment on enough units or subjects . The goal is to reduce the effect of chance variation on the outcomes.

Experimental design CONTROL RANDOMIZATION REPLICATION

Completely Randomized Design Comparison Group 1: treatment Compare results Volunteers Group 2: control Administration of different treatments Random allocation

Explain the reason for: To minimize the effect of variables that are not under study, on the response variable (control for confounding variables).  COMPARING two or more treatments. CONTROL: RANDOMIZATION: REPLICATION: To minimize the effect of lurking and confounding variables by equalizing groups. To minimize the effect of chance variation on the outcome by repeating the experiment of enough units / subjects.

Placebo effect The placebo effect refers to the phenomenon in which some people experience some type of benefit after the administration of a placebo. A placebo is a substance with no (known) medical effects, such as sterile water, saline solution or a sugar pill.

Blind / Double Blind Experiment Blind: Subjects/ units do not know which treatment they will be given. Double blind: nether subjects nor researchers know what type of treatment each subject/unit will receive.

Statistical Significance The difference between what we would expect if no treatment were imposed and what we actually found is statistically significant if the probability of obtaining a difference this big (or bigger) is extremely low.

BLOCKING Another way to control for confounding variables is to use block design. A block is a group of subjects / units that we suspect to be similar in some way that is expected to affect the response to the treatment. Divide the subjects / units into blocks, Perform a completely randomized design on each block Compare each block independently (we do not compare blocks)

Block design Group 1: treatment Compare results Block 1 Group 2: control Volunteers Group 1: treatment Compare results Block 2 Group 2: control

MATCHED PAIR Compares just two treatments. 1. Choose blocks of two units that are as closely matched as possible. 2. Randomly assign one treatment to the first subject/unit and the other to the second. 3. Compare Sometimes the block can consist of just one subject (example: to test the effectiveness of a cream for hands; use cream on one hand and nothing on the other hand of the same subject)