Causation

• Experiments give us evidence of cause and effect by demonstrating what happens when something is changed while everything else remains the same

• Randomized Clinical Trials (RCT)

  • subjects are randomly assigned (the R) to either a treatment or a control group and given one or more experimental procedures or drugs

  • also referred to as randomized controlled trials, including a control group where subjects receive a placebo or no treatment

  • when ‘clinical’ is used, there isn’t necessarily a control group

  • the golden standard for studies/experiments

Experiments compared to other methods

• Other methods, like surveys, can determine correlations

  • CORRELATION DOES NOT EQUAL CAUSATION

• Triangulation, or the use of different methods to study the same phenomenon, is important to scientific inquiry because it helps give us confidence that what we see using one method can be replicated, or reproduced, using another

Basic criteria for experiments

• 1. the cause must precede the effect

• 2. the cause must also be related to the effect

  • logical or theoretical reason and empirical evidence to suggest causation

• 3. there must be no other plausible alternative explanation for the effect

• good experiments should also include mechanisms for explaining why a certain effect occurred

Elements of experiments

• In experiments, variation is achieved by manipulating the independent variable, the thing researchers will think cause a change

• do not allow factors to covary (vary together)

Confounds

• Things that could provide plausible alternative explanations are called confounds

• Key to a successful experiment is controlling for extraneous influences that might have caused the outcome

• Researchers discover potential confounds two ways: by using common sense and by reading other studies

• experimental control, or carefully managing the variables in play

• statistically control, or take it out of the equation before examining what effect the manipulation had

• Covariates work by taking out the effect of the potentially confounding variable so researchers can see the true effect of whatever variable is being manipulated

Control groups

• The effect is the difference between what did happen when people got the treatment and what would have happened had they not gotten it

Assignment

• gold standard: random assignment

  • a way of placing subjects into the groups in such a way that individual differences are evenly distributed across the different groups