Psychologists study mental processes and behavior by observing, interpreting, and recording how humans and animals relate to one another and the environment.
To do this, psychologists look for patterns that will help them understand and predict behavior using scientific methods and principles to test their ideas.
David Myers provides an example of how one might conduct a study on the causes of human happiness. First, he explains, we would formulate a theory and create a hypothesis. In this case, we may theorize that human connection is an important factor in happiness, so we may predict that married couples are happier than never-married people. We would then observe behavior either in the lab or real life and record, analyze, and report the findings. Although scientists look for validity and reliability in their findings, sometimes neither is present. Reliability is the consistency of measurement, while validity is whether the measurement is accurate and true.
In order to test a hypothesis, researchers use various methods, including surveys, naturalistic observation, and case studies. Surveys allow us to generalize to the population by randomly sampling people and asking them specific questions about their thoughts, feelings, or behaviors. Robert Epstein explains that while naturalistic observation tends to be widely used by social psychologists and is the simplest method, it can be somewhat limited in terms of how much we can actually understand about human nature. In case studies, the sample size is extremely low and we cannot generalize to the population, but we can learn detailed information about individual experience.
None of these methods allow us to determine causality. In order to do this, we must conduct an experiment.
An example of an experiment is described in which researchers are testing the effects of gingko biloba on memory. Daniel Gilbert points out the importance of designing experiments that take ethical guidelines into consideration.
A well-designed experiment will also take into account the confounding variables and tease them out using manipulation. In addition, participants should be randomly assigned to either the experimental or control group. When testing drugs, for example, it is common to give a placebo to the control group and to avoid having the researcher know which condition the participant is in, like in a double-blind experiment.
Experiments that get results can be replicated, and when the results are consistent through continual replication, people begin to accept the theory.