Topic: Methods of Research
When a researcher is ready to formulate a study, he or she chooses from several different methods. The best method depends on the research question and hypothesis. The different methods are:
1. Naturalistic Observation
Definition: participants are carefully observed in their natural setting without interference by the researchers.
Researchers should be inconspicuous and do nothing to change the environment or behavior of the participants.
(a) an anthropologist unobtrusively observing wild
(b) a researcher sitting in a fast food restaurant and
observing the eating habits of men vs. women
This method is good if a researcher wants participants to be reacting normally but it can be time consuming, the "sought-after" behavior may never occur, there is no control over the environment (e.g., fast food restaurant runs out of fries), and it is difficult to know if the researcher will be able to be completely unobtrusive.
2. Survey Method
Definition: questioning a large group of people about their attitudes, beliefs, etc.
Conducting a survey requires a representative sample, or a sample that reflects all major characteristics of the population you want to represent. If you are attempting to survey "America's attitude towards exercising", then your sample cannot include only caucasian, upper-class college students between the ages of 18 and 22 years. This does not represent America.
Surveys must also use careful wording in the questions to prevent confusion or bias.
(a) survey of recent retired citizens on their major concerns
about life without work
(b) survey of first-time pregnant women on their beliefs about
their efficacy on being a mother
This method is very quick and efficient; however it is sometimes difficult to gain in-depth knowledge from a survey and there is no guarantee that the person taking the survey is being open and honest.
3. Case Study
Definition: obtaining detailed information about an individual to develop general principles about behavior.
It is sometimes very helpful to study one person (or a very small group of people) in great depth to learn as much information as possible. This method is particularly useful in studying rare disorders or circumstances.
(a) studying the life history of a man who acquired
schizophrenia at the age of 20
(b) following one child from conception to adulthood to
examine this specific lifespan development
Case studies require a lot of time, effot, and attention to detail. Yet, they reveal more about a particular subject than any other research method. Generalizing the findings to other people or groups is usually difficult.
4. Correlational Design
Definition: measuring the relation between two variables.
Sometimes correlation studies are seen as a separate research method while other times it is subsumed under another category.
Correlations are stated as either positive or negative.
Positive correlations mean that as the value of one variable goes up, the value of the other variable goes up (or, vice versa: as one goes down, the other goes down)
Negative correlations mean that as the value of one variable goes up, the value of the other variable goes down. See the examples below for further clarification.
(a) there exists a positive correlation between intelligence and grade point average such that the more intelligent a person is, the higher their grade point average
(b) there exists a negative correlation between eating junk food and overall health such that the more junk food a person ingests, the less they are healthier
CORRELATION DOES NOT MEAN CAUSATION. The most a reseacher can state about 2 variables that correlate is that they relate to one another. There is no test of cause-effect. In the second example above, it might be tempting to assume that consumption of junk food causes a decline in health. However, it is conceivable that the less healthy one is and feels, the more likely it is they'll give up on trying to be healthy and eat junk food. We do not know the direction of influence (eating junk food leads to poor health or poor health leads to eating junk food) and cannot know using a correlation alone. This is one limitation to this method.
Correlations can be deceiving. Finding a significant correlation between 2 variables does not guarantee that they are the only 2 variables. There may be an intervening variable that wasn't measured. Consider the first example above: perhaps the more intelligent a person is, the more likely they are to study for tests, which then translates into a higher grade point average. "Studying for tests" is a potential intervening variable that was not examined.
5. Experimental Method
Definition: a study in which the investigator manipulates (at least) one variable while measuring (at least) one other variable
This method is often used in psychological research and can potentionally lead to answering cause-effect questions.
(a) Testing the effects of ritalin medication on the attention spans of children with ADHD
(b) Examining the reliability of eyewitness testimony in young children
Participants in an experiment are usually randomly assigned to different groups. The group that receives the independent variable is called the experimental group and the group of participants are treated in the same manner as the experimental group but do not receive the independent variable is called the control group. Sometimes a preexisting characteristic already exists in the participants, such as sex, age, clinical diagnosis, etc. In this case, there is no random assignment and the type of research is referred to as differential research.