The purpose of this section is to give you a basic outline of the scientific method, as well as a brief intro to research methods for the social sciences.
Overview
The difference between philosophers and psychologists is their methodology, and the transition from philosophy to psychology in the 19th century centered around developing ways to apply the scientific method to the study of the mind. The benefits of using the scientific method are many – it provides a common framework, language, and set of assumptions for all scientists. This allows them to share their research and conclusions, replicate studies, and elaborate on the results. It is the ability to replicate and control elements of a study that differentiates “science” from “anecdotal” evidence.
There are four primary assumptions in the scientific method. First, events are lawful. What this means is that we can observe patterns and then come up with laws that explain the way those patterns behave. Second, events have explanations, which is just a weird way of saying that even if we don’t know why something happens, there IS a reason which can eventually be found. Third, we assume that scientists are going to be open minded and approach research in search of the truth, rather than their own agenda. Lastly, we assume that scientists are skeptics who are willing to critically evaluate their own research, even when the results confirm their hypothesis.
The four steps of the scientific method are (1) create a hypothesis, (2) design and implement your study, (3) analyze the results, and (4) report your findings. The hypothesis is a tentative statement about the relationship between variables, and in its simplest form is really just a theory about how things work. The variables are the observable components of the study. In many studies there are only two, but there can be hundreds of variables. Operational Definitions are used to define the variables in terms of how they are measured; this helps other scientists understand and replicate the study.
Types of Design
The second step of the scientific method is designing and implementing your study. There are two types of research design: descriptive and experimental. Descriptive designs (which include case studies, correlation studies, naturalistic observation, and surveys) focus on looking for patterns and then describing the patterns as they emerge. Experimental designs are a little more complex, and the increased level of control allows the researcher to make statements not just about the existence of a pattern, but about causes.
This leads to the correlation/causation problem. A correlation study may show a very strong relationship between two variables – for example, epidemiological studies show that diagnoses of Autism apparently increased as the percentage of children receiving MMR vaccinations increased. There is a positive correlation between infant MMR vaccination and autism diagnosis, which would seem to indicate that there is a relationship between these events. However, despite the existence of a strong relationship, it is still incorrect to assume that MMR vaccinations cause autism, for a number of reasons. The most important reason is that correlation studies are not tightly controlled enough to rule out the existence of other variables that might impact the results; for example, although the percentage of children receiving MMR vaccinations has increased along with diagnoses of autism, the “third variable” (sometimes called the extraneous or confounding variable) in this scenario could be either (1) increased awareness of autism, (2) autism being a preferred diagnosis over schizophrenia or other disorders with symptom overlap, (3) expanded definitions of autism to include a broader spectrum of behavior, or some combination of all three. Either of those three variables could theoretically account for the increase in diagnosis, and experimental studies that take into account those variables consistently demonstrate that there is no link between childhood vaccinations and autism. So it is possible for a correlation study to show one result, and for an experimental study to confirm the relationship while disproving the causality.
For this reason, experimental designs are both more rigorous and more methodical. Instead of simple variables, experimental designs incorporate independent and dependent variables, which allows the experimenter to directly manipulate the conditions of the study to observe whether there is any change. Experimenters also employ the use of random sampling and control groups to add an additional layer of accuracy to their results. Random sampling allows experimenters to control for “individual differences” such as race, gender, or age. Control groups allow experimenters to control for things like the placebo effect, while blinds and double-blind studies control for things like demand characteristics and observer bias.
Ethical Considerations
Another issue in psychological research methods is that of ethics. In general, APA ethical guidelines require a certain amount of disclosure to any participants in a study. Subjects have to be able to give informed consent to any “treatment” just as they would if they were under a doctor’s care. This means that it is difficult for researchers to avoid the placebo effect, because they may have to reveal the purpose of the research to the subjects, thereby influencing their results either through expectations or demand characteristics. Deception is allowed, but only to a degree – you may only lie to a subject if it is critical to the test and will result in no harm.
Many older studies lacked this element of informed consent, and so they are controversial because many people consider them unethical. Little Albert, the Milgram obedience experiment, and the Zimbardo prison study all raise red flags because of the lack of informed consent. These studies uncovered valuable information, but might not have passed an IRB ethics review under current ethical standards.
General Psychology 