The scientific method, in brief, is a means of explaining and predicting cause and effect. In psychology, the effects (also called dependent variables or DVs) are always some human thought, feeling, or behavior, or combination thereof. When someone’s thought, feeling, or behavior is observed, the scientific method is the best means of answering the “why” questions; e.g., Why does Aunt Maud drink so much alcohol? Why can I love someone I don’t even like? What motivated the terrorists in the World Trade Center and Pentagon attacks? Why are some people highly intelligent, while others are not? Why are some people humanists, but most people are not? What are the key causes that determine why most people act as they do? That means we research the cause(s) — i.e., the independent variable(s) (IVs) — of each effect, so we can better understand (explain and predict) what causes that effect, and better influence and even control that effect in the future for the benefit of the person and other people. In psychology, an application of the scientific method might look something like this:
In the “Maud’s drinking” example, first the scientific method operationally defines and measures the quality and quantity of her drinking behaviors. (What does she drink? When? How? How much? What precipitates her drinking? What are the consequences?) This process generates a list of possible causes of Maud’s drinking, as well as a list of the possible consequences of that drinking behavior.
Second, scientists form causal hypotheses — based on that causal list and their prior knowledge of these kinds of evidence. (Since causes must always precede effects, which of the already identified precipitating events, or any other known precipitating events, are the most likely causes of Maud’s drinking behaviors?) In this example, let’s say that we determine that most of Maud’s drinking seems to be correlated with three precipitating events: being alone, being depressed, and instances of family conflict.
Third, those causal hypotheses are rigorously tested — by research design and statistical analysis techniques — to determine not only whether one or more of the hypothesized causes really accounts for the observed results, but exactly to what extent the results are accounted for by those causes. (Do our causal hypotheses account for most of the observed results — maybe 70% or even 90%, or more? — or do they only account for 50% of the results, or 20%, or even less?) Let’s say that “aloneness” doesn’t account for much of Maud’s drinking — i.e., she drinks as frequently in the presence of others as she does by herself — but instances of depression account for ~85% of her drinking.
Fourth, if most of the effect’s causes have been identified statistically, we proceed to experimental testing (i.e., subject the causal IVs to rigorous controlled experimentation to determine exactly how each IV effects the DV). We subject Maud and/or other drinkers like Maud to carefully controlled experiments on each cause, each effect, and each consequence of Maud’s drinking behaviors. (NOTE: If at Step 4 we haven’t accounted for most of the causal agents, we recycle back to the original evidence to find more causal hypotheses, to then be tested the same way with Steps 1-4.)