Chapter 2: Section 2: Experimental Methods
Starting from the general and moving to the more specific, the first concept we need to discuss is Theory. A theory can be defined as a “general principle proposed to explain how a number of separate facts are related.” In other words, a theory is an “idea about a relationship.” In order to test whether a theory is correct or not, we need to do research. Theories are stated in general terms, so we need to define more accurately what we will be doing in our experiment.
To do this, we need to define the variables in our theory so that they are testable, and every experiment has two types of variables:
- Independent Variable (IV) – the variable that is manipulated by the experimenter (input variable)
- Dependent Variable (DV) – the outcome variable (results of the experiment)
By defining our variables that we will use to test our theory we derive at our Hypothesis, which is a testable form of a theory.
As an example of this, lets say that we have a theory that people who drive sports cars are more aggressive in theory interactions with others. Our independent variable would be the type of car you drive (sports, sedan, SUV, etc.). Our dependent variables, the outcome of our research, would be aggression. We would need to further define aggression so that it is something we can test such as speeding or cutting other people off in traffic. We now have the basics of our very simple experiment and can write our Hypothesis: People who drive sports cars drive over the speed limit more frequently than people who drive other types of cars.
Now we’ve got a hypothesis which is the first step in doing an experiment. Before we can continue, however, we need to be aware of some aspects of research that can contaminate our results. In other words, what could get in the way of our results in this study being accurate. These aspects are called research biases, and there are basically three main biases we need to be concerned with.
- Selection Bias – occurs when differences between groups are present at the beginning of the experiment.
- Placebo Effect – involves the influencing of performance due to the subject’s belief about the results. In other words, if I believe the new medication will help me feel better, I may feel better even if the new medication is only a sugar pill. This demonstrates the power of the mind to change a person’s perceptions of reality.
- Experimenter Bias – The same way a person’s belief’s can influence his or her perception, so can the belief of the experimenter. If I’m doing an experiment, and really believe my treatment works, or I really want the treatment to work because it will mean big bucks for me, I might behave in a manner that will influence the subject.
Controlling for Biases
After carefully reviewing our study and determining what might effect our results that are not part of the experiment, we need to control for these biases. To control for selection bias, most experiments use what’s called Random Assignment, which means assigning the subjects to each group based on chance rather than human decision. To control for the placebo effect, subjects are often not informed of the purpose of the experiment. This is called a Blind study, because the subjects are blind to the expected results. To control for experimenter biases, we can utilize a Double-Blind study, which means that both the experimenter and the subjects are blind to the purpose and anticipated results of the study.
We have our hypothesis, and we know what our subject pool is, the next thing we have to do is standardize the experiment. Standardization refers to a specific set of instructions. The reason we want the experiment to be standardized is twofold.
First, we want to make sure all subjects are given the same instructions, presented with the experiment in the same manner, and that all of the data is collected exactly the same or all subjects. Second, single experiments cannot typically stand on their own. To really show that are results are valid, experiments need to be replicated by other experimenters with different subjects. To do this, the experimenters need to know exactly what we did so they can replicate it.