Designing An Experiment Step by step Steps in Designing an Experiment Hypothesis Pick Population: Random Selection then Random Assignment. Operationalize the Variables Identify Independent and Dependent Variables. Look for Extraneous Variables Type of Experiment: Blind, Double Blind etc.. Gather Data Analyze Results Repeat Designing An Experiment Step by step
Let’s do an experiment • Hypothesis: The red pill will reduce anxiety. • We operationalize the definition of anxiety to mean those whose doctors claim they suffer from anxiety. • We find 100 people who fit the operationalized definition
• We randomly assign half the men to the experimental group and half the men to the control group. (Same with women). • I, the researcher, do not know which group will receive the medication and which will receive the placebo. That means this is a double-blind experiment. This will reduce experimenter bias.
The experimental group will receive the actual medication The experimental group will receive the actual medication. The medication is called the independent variable. • The control group will receive a sugar pill (the placebo).
• The research team will ask all participants to measure their level of anxiety on a scale from 1 to 10. Anxiety is the dependent variable (what is measured).
• The control group will usually report a decrease in anxiety even though they received no medicine. This is called the placebo effect.
• Now that the experiment phase is done, you must consider the confounding variables. This is the stuff that will screw up your experiment. Ex: what if the control group had a mean age much less than the experimental group? What if the 2 groups had a different percentage of women?
• Our original hypothesis was: the red pill will reduce anxiety by 40% • Our original hypothesis was: the red pill will reduce anxiety by 40%. • Results: The experimental group reported a mean of 10% reduction in anxiety versus a 5% reduction for the control group. • Theory: After several replications, the medicine has no significant effect on anxiety.
Reliability and validity? • A finding is reliable if it can be replicated. If subsequent studies show that the red pill reduces anxiety then the findings are reliable, thus supporting the hypothesis.
• A study is valid if it measures what it is supposed to measure • A study is valid if it measures what it is supposed to measure. If our experiment measured hypertension instead of anxiety, then the test is invalid, even if it is reliable.