In order to prove itself as a legitimate field of study and science, psychology relied heavily on research and experiments. But this research needs to ensure it is correctly studying what it claims to study, and there are many items to consider when designing an experiment. The research methods in psychology are not very different than the research methods used in any other field of study, but as psychologists it is important that we understand all of the terminology when we are conducting our own research or reading the latest research of others. Keeping up on the advances in your filed is every psychologists obligation, and is why it is included on the EPPP.
The topic of Research Methods and Statistics is weighted the lowest at 8%, but just because it is the lowest does not mean it is OK to skip studying this component. You don’t want to spend the most time on this topic, but a basic understanding is still important. The information is best used in collaboration with your own knowledge and what you have already been taught, your previous notes or textbooks, and research if you need more information on a particular topic. Or you could check out the product reviews page for a variety of options to help you prepare for the EPPP.
There are two types of variables in research design:
Independent Variables: are the variable the components that are being compared. For example in the very serious experimental design of “who has more fun, blondes or brunettes”, the hair color is the independent variable. And there are 2 types of independent variables: Non-Manipulated Variables are differences that are pre-existing and the experimenter does not create, for example in the blondes vs brunette study selecting a natural blonde and a natural brunette to compare. Manipulated Variables are differences that the experimenter controls and directly manipulates, for example in the blondes vs brunettes study using the same subject but dying the hair from blonde to brunette for the purpose of the study.
Dependent Variables: are the variable are being measured or the outcome of the comparison. For example in the “who has more fun, blondes or brunettes” the dependent variable is the measurement of fun which may be rated on an interval.
Types of Experiments
True Experimental Research: This is a very narrow window of experimental design and requires subjects to be randomly assigned, and for at least one independent variable to be manipulated by the experimenter. For example, typical clinical trials where groups of people are randomly assigned to the medication trial or the placebo trial.
Quasi-Experimental Research: In this design one independent variable is still manipulated by the experimenter, but the individuals are not randomly assigned to a group because they are already assigned to pre-existing groups. For example if one wanted to test if recess improved academic performance in grade 3 students, the experimenter may let the children in Teacher A’s class out for recess, while the students in Teacher B’s classroom had to continue working through recess.
Observational Research: In this experimental design there is still an independent variable but there is not manipulation, rather the experimenter observes and documents group differences. For example comparing the effects of crack cocaine usage on males and females.
Single Subject Designs: Known as an idiographic approach, where one (or very few) subject is measured multiple times. there are 5 main types of single subject designs.
- AB Design: In this type of research design ‘A’ represents the baseline condition and ‘B’ represents the experimental condition for one subject. So in our previous example of “who has more fun, blondes or brunettes” an individual who is blonde first reports how much fun they have on a standardized scale, then that same individual’s hair is dyed brunette and given the same scale to report how much fun they have.
- ABAB Design: In this design ‘A’ still represents baseline and ‘B’ still represents the experimental condition, they are simply alternated back and forth to improve reliability of the test. For example, as a baseline a baby has 2 naps a day and wakes in the middle of the night, then as an experiment the parents switch to one nap a day and the baby sleep through the night. But then after a few days in the experimental condition go back to 2 nights a day and find the baby does not sleep through the night, and just for good measure after a few days try one nap a day and find the baby sleeps through the night.
- Multiple Baseline Design: In this type of research design treatment in applied sequentially across subjects, behaviors, or situations. In Multiple Baseline Across Subjects the baseline is measured in a small number of subjects, and then one by one the treatment is added. For example in a medication trial for Parkinson’s disease tremors are measured without any medication in 3 subjects, then one subject is given a trial medication while the other 2 remain unmedicated and continue to measure tremors, and after a set amount of time the second subject is given the medication while the third subject remains unmedicated for the same amount of time, and finally the third subject is given the medication. In Multiple Baseline Across Behaviors is the same idea as with the subjects except different behaviors in one subject. For example, a child with autism spectrum disorder has a new behavior plan, it may first be tried as an intervention with head banging, then with self-biting, and then with rocking . In Multiple Baseline Across Situations again similar to subjects and behaviors, but this time the intervention is tested on the same subject in different scenarios. For example treatment for a social anxiety disorder is tested in 3 separate anxiety provoking situation, first at the mall, second at a concert, and third at a private party.
- Alternating Treatment Design: This involves 2 or more interventions being tested at the same time, with each one attempted equal amounts of times and alternating throughout the day. For example a parent trying to increase compliance in a defiant child may alternate between rewarding compliance with either praise or a gummy bear, and compare the effectiveness of each.
- Changing Criterion Design: This design takes into effect the difficulty to change some behaviors too drastically, and instead behavior is changed in increments to meet a changing criterion. For example, someone with an alcohol addiction may drink 10 drinks a day and go into a treatment center, rather than stopping completely they gradually decrease to 8 drinks a day, then 6 then 4 and so on until the individual no longer drinks alcohol.
Group Designs: Rather than testing just one subject, group design takes into account that individual characteristics may skew results and rather test a group of subjects and measure the overall change. There are 3 main types of group design:
- Between Group Design: Basically comparing 2 independent groups with different experimental conditions. For example, which hockey coach teaches better sportsmanship can be determined by observing if hockey team A or hockey team B has better manners after losing a game.
- Within Group Design: The same group of individuals are repeatedly given several different levels of intervention. For example a school classroom is asked to complete a math worksheet in 2 minutes without music, then with rock, country, pop, classical, and ambiant noise to see which noise correlated with a higher score.
- Mixed Design: A mixture of between group and within group can be measured, so individuals are assigned to 2 separate groups, but each group is also given multiple interventions and correlated with each other. take the above answer with the math test and music, and test that in multiple different classrooms, this time alternating the order of which the genres of music are tested.
Behavioral Measurement: There are 2 main types of behavioral measurement
- Time Sampling which is useful when there is no specific beginning or end to a behavior, like how long children pay attention in math class. The recording is broken down into smaller intervals, such as an 80 minute math class in one minute intervals. This can be done with either Momentary Time Sampling where the behavior is recorded either present or not present at the end of each interval. Or Whole Interval Sampling which records whether or not the target behavior was present for the entire interval or not.
- Event Recording is simply tallying the amount of times the behavior occurred in a pre-determined amount of time, such as how many times did a client say “but” in a session.
***Research Design can really be difficult for People studying for the EPPP if this is not something that interests them. This is typically the easier to understand material, so I would recommend if this is a difficult topic for you to learn a couple of the sections and not spend too much times on others as it is weighted the lowest at 8%. In order to help you study, engage in conversation in the comments below. What do you think some of the limitations of certain research designs are? Can you identify any specific research design in any famous psychological studiesB.
You may find the information on this site is not enough to help you feel confident about your ability to pass the exam, That is OK and only you can be the judge of what you need. If this information seems overwhelming to you it does NOT mean you will fail the exam, but you may require a little more in depth material than is offered here. That is why there is a Product Reviews page which will give you a variety of additional options, as well as practice exam questions which I highly recommend as explained on the Study Tips page.