UNIT OBJECTIVES
Understand why psychological research is important.
Explain the process of correlation.
Differentiate between experiments and quasi-experiments.
Understand the goals of factor analysis.
WHY DO WE PERFORM RESEARCH?
Students of psychology, including adolescent psychology, often ask why they must learn about the process of research? Frequently people interested in counseling, clinical, or the more applied areas of psychology are particularly bothered. They believe that they possess excellent talent, maybe “great people skills.” Shouldn’t this be where their efforts should be pointed?
Students, especially those new to the field, sometimes think that learning research skills is unnecessary. They believe that intuition, observations, and growing experiences will be sufficient. They also question the time and effort put into research. It is not needed, they think, especially for themselves. However, history proves that this view has been repeatedly incorrect.
No one can genuinely understand psychology without understanding psychological research. No one can understand the behavioral sciences without understanding research. Ultimately, no one can be an effective helper without understanding research and the research process.
Like many other disciplines, psychology has a history of mistakes where well-meaning people used observation and common sense. They often made claims with the best of intentions. Results were sometimes correct, but too often, they were wrong. Research allows us to correct our mistakes.
Even people with research training often made serious errors that continue to have implications for clients and society. One example is mistakes that were made about the causes and the appropriate treatment of autism. Autism is a complex disorder that affects about 3% of children. It can cause developmental problems. These can include slowed speech and long-term social difficulties. People believed that parenting problems, notably “bad mothering,” caused this disorder. However, there was no good evidence for this. It just fit into theories at the time and seemed true.
Once researchers systematically looked for the causes of autism, they found that parenting practices had nothing to do with this disorder. Research has now shown that autism is due to biological factors, including those present early in life. It is independent of parenting styles. Parents do not cause autism, not even unconsciously, as some theorists believed. Nor is there reliable data that it is caused by vaccines.
Adequate research could have helped provide more humane, perhaps even more successful treatment for children and their parents. But, unfortunately, we wasted many years because of a lack of adequate research.
THE SCIENTIFIC PROCESS
You may hear people talk about “the scientific method” as a single formula that can always be applied to any problem. That is not quite true. Science is a process of viewing the world that tries to systematically evaluate evidence. Its’ rules may change through time as we learn whether our past efforts were accurate. Science is about explanation and prediction.
Science starts with observations. But it must go beyond just observations. Psychologists develop hypotheses, which are logical predictions based on these observations. A hypothesis is a proposed explanation for something that can actually be tested. These are often based on a prior theory.
In everyday language, the word “theory” may mean an opinion or a guess. For example, we may say, “I have a theory why that couple isn’t still dating.” By this, we might mean that we have a hunch or speculation, but we don’t necessarily have any firm knowledge.
In science, the concept of theory means quite the opposite. A theory is a comprehensive explanation that has been supported through many series of experiments or other research. Thus, a scientific theory has been frequently tested.
In Psychology and other sciences, you may hear about a hypothesis being “supported.” In more technical language, they may say that evidence “failed to reject” a hypothesis. But psychologists and most other scientists never (or should never) say their hypothesis was “proven.” This is because we can never 100% prove a hypothesis. Future research can (and sometimes does) show us that we were wrong about our causes or about some aspect of our hypothesis. Sometimes our hypotheses need revision or are only partially true. So scientists always need to have an open, skeptical attitude.
Vignette 3.1 Brittany: A SCIENCE DENIER
“I left high school as a “science denier.” That is what you could call me… My family was concerned about making money. Pretty much, that was what we talked about. The exception was my mom, who tended to kind of go down the rabbit hole of conspiracies. So we were brought up thinking that science was just another way to scam people, to make money, one big conspiracy.”
“This wasn’t anything based on our religious beliefs. We barely went to church. It was just how my family was. You make money, and you get cynical.
“I remember my mom telling me once that if scientists were so smart, why did they always change their minds? Why didn’t they have a cure for the disease that killed my grandmother? Why weren’t we all getting around in flying cars? Needless to say, she didn’t believe in vaccines or what doctors say. She thought they were all out to make money, and that was all.
“I had always done okay in science in high school, but I thought it was just something I had to do. It was like history or some other subject I didn’t really like. But my grades were okay. I just wasn’t curious. It all seemed unnecessary, at least to my 18-year-old brain.
“But when I got to college, it was different. I realized that other people didn’t all think the same way I did. Even people I grew up with, like in my town, thought differently. In one of my classes, I listened to the professor talk about what science is. It wasn’t anything like what I had thought. I realized science was logical. It made sense.
COMMON TECHNIQUES IN THE BEHAVIORAL SCIENCES
Case Studies are one method used by psychologists and other behavioral scientists. They are in-depth studies of a single person, an event, or a group. They are popular in the fields of medicine. They provide detailed information that may help generate more research. They allow investigators to get specific details. Sigmund Freud was well known for his case studies. The advantage of case studies is sometimes they are very persuasive, as the vignette of Brittany illustrates. Sometimes these are the only way to gather data.
There are many disadvantages to case studies. The researcher’s biases and beliefs may easily color the interpretations of the findings. A researcher might ignore data that went against their hypothesis or another theory. Furthermore, there is little way to say that the findings would replicate or be found in another study. There is usually no way to say that the findings might apply to someone else in a different situation or time.
One of the problems in early psychological and medical research was a reliance on case studies. Case studies often tell us “What?” but not so much “Why?” They only tell us “What?” for the person or person in the study.
Case studies have flaws, but they are still popular. Sometimes, we can improve the objectivity of case studies by using more structured and objective methods. For example, we can use standardized questionnaires, highly structured interviews, or have multiple people participating in the task to be more objective. Still, it is impossible to claim that a case study will apply in other conditions unless you conduct additional research.
Correlational Studies. A variable is anything that can vary, meaning it changes. Variables can be measured. Correlational studies look for the relationship between two or more variables. Correlation is examining what goes together. Researchers are trying to find or predict what variables correlate or go together in correlational studies and what does not. They usually express these correlational relationships with statistics so that other researchers can understand their findings.
Correlational studies can be exploratory. This happens when a researcher has little idea what she will find. They can also attempt to test hypotheses. This occurs when a researcher predicts that there will be a specific amount or direction of correlation.
Correlations have the advantage that they allow researchers to observe variables that naturally occur. The problem with correlations is that it is impossible to precisely know the cause. Two events may correlate (occur together), and one may cause another. The direction of cause might also be reversed.
The two events might also cause each other. There could always be another cause, perhaps one that is not known or is hidden.
Correlations do not allow us to know the cause. You perhaps have heard the phrase: “Correlation is not causality.” This is precisely why.
The correlation coefficient is the number used most in research to indicate the results of correlational studies. It is a statistic that shows the correlation between two variables. It is a number that ranges from -1.0 to +1.0. When it is close to +1.0, this means there is a positive correlation between the two variables. Both increase or decrease at the same rate and direction.
When there is a negative relationship closer to -1.0, one variable increases, the other decreases.
A near-zero correlation (0.0) indicates no correlation and no relationship (technically no linear relationship).
To determine the strength of the correlation, take the absolute value (remove the sign). The direction is always shown by the sign (the + or -).
An important concept is that in a correlational study, the researcher is not manipulating anything. She is looking at relationships. These can be from present data or from historical data. But the researcher is not causing any relationships to occur or excluding any. Consequently, there is a great deal she cannot control for. That is why we have experimental methods.
Experiments. To logically determine if one variable causes another, we can manipulate or control it, turning it on or off like we might a light switch. An experiment is the type of research where the cause, called the independent variable, is manipulated by the experimenter. First, the effect, called the dependent variable, is measured. Then, other variables that could be the cause are controlled or accounted for.
An experiment can take place in a laboratory or in a natural setting. The key is that the experimenter controls the variables of interest and manipulates or changes what is of interest.
The strength of experiments is that they allow for more precise control. However, in some situations, experimental findings do not generalize or apply to the real world. They may be too artificial. Experimenters also have to be alert for extraneous or outside variables which may influence the results.
Researchers realize that participants need to be randomly assigned to experimental conditions for an experimental result to be as accurate as possible. This means that everyone in the group being studied has an equal chance of participating in each group. This way, accidental and unknown participant characteristics are less likely to influence the study’s outcome. However, in many instances, experiments may also be impractical or unethical to perform.
When researchers compare two or more groups who have had different experiences, they have no control over the critical variable. This is what is called a quasi-experimental design. In a quasi-experimental design, the research cannot randomly assign participants to groups. Other variables that might co-occur cannot be completely ruled out as a cause. A quasi-experimental design can often tell us a great deal. Still, there will always be a concern that it may not be revealing the true cause of relationships between two variables.
Other types of investigatory methods used by behavioral scientists include the longitudinal study, where a researcher studies something for a long time. As an example, a researcher might wish to study how procrastination changes through the adolescent years. He might use a test for procrastination on a group of sixth-graders and readminister it each year for the next six years. While longitudinal studies help us determine patterns as they develop and are especially valuable in adolescent research, they are expensive and often impractical.
RESEARCH MEASURES
Research measures are the tools that we use to gather the data. Many aspects, including practical constraints, determine the types of tools researchers use. Every kind of research tool has its strengths and also its disadvantages. There are many tools. Below are some of the more common ones found in research with adolescents.
Behavioral measures are behaviors that can be counted. An example might be the frequency in which adolescents use the internet or the hours that they study. Behavioral measures can usually be obtained objectively and accurately. There is little disagreement about what they might mean. The disadvantage is that many variables of interest, such as people’s opinions, cannot be directly measured by behavioral methods. For these, we need different ways like questionnaires.
Questionnaires are surveys with a central theme and a purpose. Questionnaire data are popular in psychology, including adolescent research. An advantage is they can be quickly administered. In addition, they are objective, and this reduces disagreement regarding scoring.
However, there are many disadvantages. People only respond to questions presented to them, often allowing the researcher to miss important data. People that cannot read well or misread the questions are not adequately included. People can misrepresent the truth or be inconsistent or careless in their responses. Confidentiality of responding may be a problem. Furthermore, questionnaires are difficult to objectively design. If a researcher is not careful, she can construct the questionnaire to introduce accidental bias.
Psychological tests are similar to questionnaires, except they have been more standardized. This means that they have been tried out on more people. Usually, they are designed to measure only a few concepts, called constructs. For example, a psychological test of hostility is designed to measure hostile feelings. It is not intended to measure musical knowledge or political attitudes. Usually, such tests are used to predict future behaviors or to test hypotheses. Personality inventories that measure personality sometimes are composed of many smaller tests or subscales. They are usually classified as psychological tests, as are tests of intelligence and ability.
Not all psychological tests are equal. As anyone who has completed a quiz on social media knows, some are not very good. They produce silly results. If a test does something consistently, it is said to be reliable. A reliable test has the same results when it is taken again by the same person.
If a test is accurate, it is said to be valid. If a test does what it claims to do, we say it has validity. Unfortunately, most of the tests on the internet and in the media are neither reliable nor valid.
Constructing reliable tests is often very difficult and time-consuming. Several statistical procedures are applied to make sure a test has reliability. Making sure that the tests are valid, that they perform as claimed is harder still. That is why most of the “personality tests” found on social media are simply unscientific and simply for amusement, about like zodiac charts. They are fun to take, but occasionally they can be bothersome if people believe them without being critical.
Biological measures may tap variables such as heart rate, blood pressure, or respiration. They can include measures obtained from the blood or saliva. They may also include electroencephalograph (EEGs) or brain waves. Another physical measure is neuroimaging, which looks at the brain in real-time, discussed in Unit 1 and later.
ANALYZING DATA
In correlational studies, researchers often compare correlations. Mathematicians have developed several ways to determine whether differences found are likely to have occurred by chance or reflect fundamental differences.
Factor analysis is a way of reducing many correlations into an underlying pattern. It helps find hidden structures in data. Computation and math can be very complex, but computers make it easier. Exploratory factor analysis is a type of factor analysis commonly used when the research has no prior hypotheses. Confirmatory factor analysis is a type of factor analysis that can test hypotheses about factors and compare them to each other or compare them with previous studies.
In statistics classes, you will learn how scientists express the concept of what is known as statistical significance. When a study occurs outside the probability of chance, we may say it is “statistically significant.” Mathematical tables have been worked out to figure all of this out, but this is a part of every psychologist’s training! Significance is expressed as probabilities. When scientists say that a finding is significant at the level of p <.05, they mean that there are less than five chances out of 100 that the probability occurred randomly or due to chance.
Statistical significance does not mean that something is necessarily meaningful or of value. It does not even mean it is interesting! It only means that the observed differences, even small ones, did not likely occur by chance alone. Whether results are truly meaningful is for the researcher and reader to decide.
Psychologists and other scientists have many statistical tests that they use to test statistical significance. Some are complex. Many are based on the concept of averages. Most people know that there are different types of averages, including the mean, median, and mode. The mean makes it easier to compare two groups, especially with a smaller amount of available data. However, there are many situations when the mean is not an appropriate statistic, and other measures must be employed.
Psychologists have other ways besides statistical significance to determine whether results occurred by chance. For example, more current research often uses several approaches simultaneously. This may produce findings more likely to be replicated, which means that they will be found by other studies.
Meta-analysis is a process of combining the results of various studies into one final statistic. This is important because the conclusions of studies sometimes contradict each other. There are many reasons for this other than bad science or careless research. For example, studies conducted during specific years or certain locations may only hold for particular groups. Even the time of year a study was conducted might accidentally influence its results. Unfortunately, researchers might have no way of recognizing this unless they perform a meta-analysis.
A meta-analysis combines previous research into a more usable whole. It allows researchers to look for patterns where there are exceptions when data did not come out as expected. It also enables us to see how powerful the effects of findings are compared to other studies’ results. The disadvantage is that since meta-analyses combine other studies, there must first be many studies that have already been completed.
RESEARCH ETHICS
Psychologists and other scientists must abide by ethical codes and guidelines. This is to prevent foreseeable harm to their subjects and to society.
Proposed research projects are evaluated for their potential risk. Colleges and universities require that researchers obtain approval for their research from internal review boards. This includes projects that students design and carry out.
Projects are evaluated to assess their safety and relative risks. Sometimes researchers have excellent research ideas, but their proposed projects are potentially too risky. Conversely, some potentially tricky projects can be approved if the worth to society of potentially new knowledge is favorable.
Ethical guidelines state that research subject participation is always voluntary. Subjects may withdraw from research without negative consequences. In most cases, people need to sign an agreement to participate in a research study. This is called Informed Consent. Adolescent people under 18 who are under parental care can give written consent to participate in research, called assent or permission, and their parents or guardians must also consent.
Ethical guidelines also require that researchers try to ensure the privacy of participants. Thus, researchers are required to protect anonymity when necessary. This is particularly important for information that may be at all embarrassing or controversial. For people who are legally minors, which includes many adolescents, this imposes additional research restrictions.
Vignette 3.2 Kayla: A PROJECT TOO RISKY?
Kayla was a college junior who was majoring in the physical sciences. She was an exceptional and eager student. Her interests were in eventually developing a drug to prevent sexual diseases in gay men. She hoped to be accepted into medical school and to eventually conduct research in this area. She figured she would get a head start and begin research early.
“I had an idea I thought was really good. I wanted to research gay men’s first sexual experiences and relate this to the knowledge of current risks of sexual diseases. I believe this is an important issue. The problem was, I knew that many parents did not want to discuss this topic, at least in my part of the country.
“So, what to do? My idea was to put up a web page and have gay high school kids complete anonymous surveys. I would leave information in their schools telling them where to find the survey. That way, it would all be anonymous. The best part, their parents wouldn’t even know. That’s important because I was trying to get information from kids who were 14 and 15.”
Kayla eventually realized this plan had problems. This realization occurred when she received research training in one of her psychology classes. She learned that research with human subjects must be approved by a review committee. Furthermore, research regarding children and adolescents may have additional legal and ethical restrictions.
“I’m glad I wasn’t naïve enough to collect that data. If I had been a parent or even a subject, I imagine I might have gotten outraged. Funny, I didn’t think of this at the time. I still think the project is important. But when I do it, I’ll do it the right way.”
CRITICAL THINKING
Can you think of research that is appropriate for a case study?
Think of examples of positive and negative correlations. Can you think of an instance when correlation and cause are not the same?
Factor analysis is a statistical procedure that helps us determine the relationship between variables, such as personality traits? Can you think of some personality traits in adolescents that might be related?
Should adolescents be allowed to give consent to participate in research without their parents’ permission? If so, at what age?
What ethical problems might researchers have when they tried to examine the social media posts of adolescents? Are there ways of overcoming these problems? How could we protect people’s privacy while we perform research on their internet behaviors?