Research Methods in Psychology

📘 Science Practices and Research Methods

This background unit introduces the scientific methods psychologists use to investigate questions about human thought, emotion, and behavior. Understanding research methods enables us to evaluate psychological claims, interpret research findings, and develop critical thinking skills. These skills are essential for success on the AP Psychology exam, which is the focus of this AP Psychology Self-Study Hub.

We cover four topics.

First, what is research? What are the principles of scientific research? We discuss how scientists formulate research questions, theory, and hypotheses. A key concept in this topic is the difference between theory and hypothesis. Second, we cover the different research methods, including descriptive research, correlation research, and experimental research. These are the key topics of scientific research, particularly the methods used in psychology. The third topic covers statistics. We discuss descriptive and inferential statistics. Lastly, we cover the ethical principles in research, especially how researchers treat experimental animals and human subjects in humane and ethical manners.

Part 1: What is Research?

Research is something that people do. It is a human activity. Can you replace the word research with another word? Give me another verb to replace the word research. To research is to______. You may say to research is to study, to examine, to investigate. What do these activities have in common? They have one thing in common. These activities involve gathering and analyzing information.

In scientific research we gather and analyze information in systematic ways. What do we mean by “in systematic ways”? Being systematic means you gather and analyze information guided by a research issue or a research question. You have a purpose in your information gathering and analysis. Your research activities are not random.

This approach is different from casual observations. If you travel to a place for tourism, for fun, you don’t have a particular issue to investigate. You people-watch. You enjoy the scenery, maybe you are interested in the local culture, but you don’t have well-formulated questions to answer. You just soak in the information, the sensation, the feelings that are generated by your experience with the place. This approach is different from research. Research is always guided by well-formulated issues or questions. For example, you can do research on why some people get infected by the flu virus in the winter, while other people remain unaffected. After having contact with the virus why do some people become very sick but others remain relatively unaffected. These are research issues. Researchers gather information by making observations to answer these questions. These activities would qualify as research.

Systematic gathering and analyzing information have several characteristics. Number one: you must guide your observation with a research issue. Number two: you must use systematic ways of collecting data, rather than making random observations. There are different methods for collecting data. We will cover the descriptive research, correlation research and experimental research. Experimental research is a more rigorous and controlled way of gathering information.

In the history of psychology, Wilhelm Wunt, a German psychologist was credited as one of the founders of psychological science. He did one thing particularly important in qualifying psychology as a science, which was setting up the first psychology laboratory at a German university. Why is having a laboratory important for psychology to qualify as a science? Because in scientific research we use systematic ways of collecting data and analyzing data. Having a laboratory means researchers use systematic and controlled ways of making observations and gathering information. They can control the conditions for their observations.

In systematic research, we use statistical methods to analyze data and draw conclusions. We will cover the descriptive and inferential statistics later in this chapter.

Part 2: Theory vs. Hypothesis

Next we discuss the concept of theory. In most research activities, researchers not only have a research question, but also a theory. A theory is a tentative answer to the research question. For example, I am interested in studying aggression, in understanding why some people and animals are more aggressive than others. I have observed male animals tend to be more aggressive than female animals. Male mammals fight more than female mammals. Based on these casual observations I generate a tentative answer to the research question, which is what factors cause some animals to be more aggressive. I have a theory which says the level of testosterone in male animals is associated with aggressiveness; the higher the testosterone level, the more aggressive the animal is.

Testosterone is a male hormone. It is produced by the testes in male. Females also have testosterone produced from the ovaries, but at a much lower than in men. Is this a correct theory? How can I test the theory?

We cannot test if a theory is true or false directly, because a theory is too abstract for observation. You cannot make observations to cover the entire scope of the theory. My theory says the level of testosterone in animals is associated with aggressiveness. There are two concepts in this theory. One concept is testosterone level, another is aggressiveness. Can I make observations on these two concepts? No, because they’re too broad. What do I mean by testosterone level? I am referring to testosterone in all animals living today, living in the future and those animals who have lived in the past. I am talking about testosterone in this theory in a very abstract way. I don’t refer to testosterone in particular animals.

The same goes with aggressiveness. I’m not just talking about physical aggressiveness, or verbal aggressiveness, or cyberbullying. I’m referring to all types of aggressiveness, maybe even new ways to be aggressive that haven’t been discovered yet but will evolve in the future. So, it’s impossible to make observations on the entire scope covered by a theory. You don’t ask whether a theory is correct or not.

The best we can do is to test a hypothesis derived from a theory. What is a hypothesis, and how does it differ from a theory? A hypothesis is a concrete, precise prediction based on your theory. If your theory is right, what kind of predictions will you be able to make?

If my testosterone theory is correct, I can make a specific prediction: male rats injected with testosterone should attack other rats more frequently than those injected with a saline placebo. This is a specific prediction. It specifies the operational definition of testosterone level, hormone injection versus placebo injuection. Another hypothesis I can make is: if you go to a prison, look at the records of the inmates, and draw blood samples to measure testosterone, you will find a positive correlation between testosterone level and the type of crime they have committed. More specifically, prisoners who committed violent crimes like murder or physical assault will have higher testosterone levels than those who committed non-violent crimes like embezzlement. These are concrete predictions based on the theory. They are so specific that they can be observed.

A hypothesis tells you how to make observations. The first hypothesis mentioned above specifies that you should observe male rats as your subjects, and you should observe those rats under the condition of injecting them with either testosterone or placebo, and you should measure how many times each rat attacks other rats within a period of time.

To test the second hypothesis you need to observe prisoners. You should observe the level of violence in their crimes and measure their testosterone levels. You should find a positive correlation between the levels of violence and levels of testosterone. The difference between hypothesis and theory is that theories are broad and abstract, while hypotheses are testable predictions based on theory. A theory can produce many hypotheses.

Let’s look at another theory about aggression. My theory says frustration causes an increase in aggression. When people want to get something but cannot get it, they tend to be aggressive. We cannot directly test this theory. Fortunately, a theory gives us specific predictions, which are called hypotheses. If my theory is correct, what observation should I be able to make? One prediction is: among football teams, those who want to win really badly but are losing the game should show more aggressiveness than other teams.

This hypothesis tells you how to make observations. It directs you to observe football teams, particularly those who must win to avoid elimination. They want to win badly, they have no alternatives, but they are losing anyway. Under this condition, they should show high aggressiveness. You then define what you mean by aggressiveness—maybe illegal pushes, fighting on the field—and you count how many times these transgressions are committed. If you find support for this hypothesis, your theory receives support. This is how we treat theories in scientific research: we never prove a theory correct, but we can accumulate evidence in support of our theories by testing our hypotheses.

Part 3: The Principle of Empiricism

Hypothesis testing is related to the principle of empiricism. Empiricism is a way to determine the truth value of a theory. It means we use evidence based on experience and observation to verify our ideas. All branches of science follow this principle. Psychology is a science, which means we test psychological propositions based on evidence and observation.

A good way to understand empiricism is by comparing it with other ways of determining truth. Not all people settle the truth value of ideas by making observations. Some people rely on tradition. They hold onto a belief because it has been passed down for generations. For example, in the old days, people believed crazy people were possessed by demons. Today, some people believe certain herbal medicines can prevent or treat coronavirus infections just because of tradition, people in certain cultures have been using these mthods for thousands of years, even though these ideas haven’t been tested empirically.

Some people rely on authority. They accept a belief because it was passed on by an expert or a leader. But leaders and experts can be wrong. For example, in the 1950s, psychologists believed schizophrenia and autism were caused by poor parenting—specifically, that “cold mothers” caused autism. We now know through scientific research that autism and schizophrenia have genetic and biological bases, not poor parenting styles. Authority is not a reliable way to determine truth.

Another way is rationality or logic. Some ideas sound reasonable through reasoning. For example, using transitive reasoning: if A equals B, and B equals C, then A equals C. You don’t need to make observations to prove this; it’s a logical truth. It belongs to philosophy, not science. In science, we are interested in empirical truth, which can only be verified by making observations.

In the history of psychology, some ideas were not suitable for empirical verification. For example, Sigmund Freud proposed that human behavior is driven by unconscious instincts, such as unconscious sex and death instincts. Can we conduct empirical investigation on unconscious drives? It’s difficult. If it’s unconscious, the person themselves doesn’t know they have it. How can anybody observe that? If something is unobservable, you cannot conduct empirical verification. Behaviorism went to the opposite extreme. Behaviorists are only interested in things we can directly observe: observable stimuli (such a loud noise) and observable responses (like a baby crying).

Part 4: Key Components of Research – Theory, Hypothesis and Variables

Let’s look at the key components of a research project. What are the steps we need to take in order to complete a research project?

First, we must have a research issue. All research starts with questions. Psychology asks many questions about human behavior and human functioning. For example, I am interested in anxiety, particularly in anxiety disorders. I can ask why some people suffer from anxiety disorders, why some people worry so much, why they become nervous and stressed out easily. In order to answer these questions I come up with theories. Many theories have been proposed in psychology to explain human differences in various characteristics, such as anxiety, depression, friendliness, conscientiousness and so forth.

For example, I have a theory that says anxiety disorders are caused by early traumatic experiences.

How do I test this theory?

We cannot test whether a theory is true or not directly. We test a hypothesis derived from the theory, which is a specific prediction based on the theory. If the hypothesis is supported by evidence, we say the theory receives empirical support. My theory says anxiety disorders may be caused by earlier traumatic experiences. I can derive many specific predictions based on my theory. One prediction is that if I study people with anxiety disorders, I should find that a higher percentage of them have had earlier traumatic experiences than the general population. If I investigate 50 individuals who have been diagnosed with anxiety disorders by qualified doctors, then if I interview the 50 people, I should find that many of them had suffered traumatic experiences before they developed anxiety disorder. They might have been bullied at school, or experienced domestic violence, or witnessed violence. I can generate another hypothesis which says if we give research animal traumatic experiences they will later develop anxiety reactions similar to human anxiety disorders.

Once you have figured out a specific prediction, a hypothesis, you know how to make observations based on your prediction. You will be able to design a research and to test your hypothesis.

A research design is a way to gather information, to collect data, to make observations and analyze data. If your data analysis supports your hypothesis, then you can say my theory receive empirical support.

To summarize, these are the key steps or components of scientific research.

First, you identify a research issue, you ask a specific question. Then you come up with a theory which is a tentative answer to your question. Then you figure out the prediction based on your theory which is a hypothesis. Next you develop a research design to guide you to gather information, collected data, to test your hypothesis. Then you analyze the data and draw conclusions. You conclude whether your hypothesis is supported or not.

Variables – Categorical and Numerical

When you compose a research issue and formulate a theory, you use variables as raw materials. A research issue in psychology is a question about the relations between psychological variables. A psychological theory is a proposition about relations between psychological variables.

Let’s look at what variables are. A variable is a characteristic that varies.

Can you replace the word “vary” with another word? To vary is to change, to differ, to take on different values. A variable is a characteristic that can take on more than one value.

Can you think of a physical characteristic of the human body that varies from person to person? Body height is a variable. Some people are taller, some people are shorter. Biological sex is also a variable, with male and female as different values. Men on average are taller than women. These two variables are related.

Body weight is another variable. My value on this variable is about 80 kilograms.

People vary in personality, such as introversion-extraversion. Some people are more outgoing, some people are more introverted and shy. Some people are socially anxious.

Intelligence is a variable that can be measured with scores on IQ tests.

There are characteristics that only have one value and therefore are not variables. They are constants because they do not vary. An example is the number of head(s) each person has. All people have the same value on this characteristic. When people lose their head, when they are decapitated and have zero head, they stop being people and become carcases or corpses.  Psychologists are interested in variables, in how people differ from each other. Variables are important in psychological research.

Categorical and Numerical Variables

There are two broad ways we can measure a variable. One type of variable is called a nominal  or categorical variable. They measure qualitative differences.

Gender is a qualitative or categorical variable. In some cultures this variable has only two values, male and female. In other cultures this variable has more than two values, beside male or female, a person’s gender may take on other values, such as non-binary, trans-gender etc. These are qualitative differences. You cannot do math with these values. You cannot add or subtract with the values on those values.

Ethnicity is a categorical variable. It has different values, Asian, African, European, etc.

Attitudes can be measured as a qualitative or categorical variable. During the COVID-19 pandemic, people have positive or neutral or negative attitudes toward social distancing and wearing face masks.

Another type of variable is numerical. The values of these variables are numbers. They measure quantitative differences. For example, generosity can be measured by how much money people donate to a charity. You can donate one dollar, two dollars, 10, 20 dollars etc.

IQ score is a numerical variable.

The prevalence of mental disorders can also be measured as a numerical variable. We can measure what percentage of people suffer from depression, anxiety, schizophrenia and etc. in a community.

There are two types of numerical variables. In other words, numeric variables can be measured in two different ways. Rank order variables are a relatively less precise way of measuring quantities. In a marathon race, the champion gets the gold medal, the second place gets the silver medal and the third place gets the bronze medal. These are rank orders. You know the gold medalist ran faster than the silver medalist, who ran  faster than the bronze medal winner, but you cannot tell how much the difference is between them simply by knowing the rank order.

A more precise way of measuring numerical or quantitative differences is equal interval variables.

The following table was published by the Word Health Organizations. It shows statistics of cases of coronavirus infection worldwide on July 24, 2020. It shows the rank order of the different regions.

The regions are a categorical or nominal variable. The values on this variable are the names of the regions, such as Americas and Europe.

The Americas, including North America and South America combined, had the largest number of cases and ranked number one. The second place was Europe, which had fewer cases than the Americas, but more cases than the regions ranked number three or lower. Number three was Southeast Asia, followed by number four Eastern Mediterranean, number five Africa, and Western Pacific, which had the least number of cases. This ranking is a rank order variable.

You can measure the cases more precisely by counting how many cases each region had.

On that day the Americas reported 8,121,700 cases, Europe reported 3,170,382 cases, and Western Pacific had only 276,845 cases. The number of cases is an equal interval variable, because the interval between one case and two cases is equal to the difference between two cases and three cases. The unit of increment is constant, always one case at a time.

In fact the number of virus infection cases is a special type of equal interval variable; it is a ratio variable. A ratio variable not only has equal intervals, but also has a meaningful zero point. When there is no new case reported, the value is zero. This value of zero is meaningful. If Europe reports two cases and Africa has four cases, then Africa has twice as many cases as Europe. Money is also a ratio variable, so is people’s age.

Some equal interval variables are not ratios. They have equal units of differences but no meaningful zero point. IQ scores are not a ratio variable, because there is no meaningful 0 IQ. A person with an IQ score of 120 is not twice as high as another person with a score of 60. Temperature in Celsius  is not a ratio variable, because 0°C does not mean no temperature or no heat. Moreover, 20°C is not twice as hot as 10°C.

Part 5: Descriptive Research – Case Studies, Surveys and Naturalistic Observations

Next, let’s talk about research methodology, which refers to how you collect data. In psychology, we have three broad types: descriptive, correlational, and experimental research.

Descriptive research aims at determining how things are. We can study how many spoken words are mastered by average normal developing children three years of age. We can investigate if women are more talkative than men. These studies describe phenomena without explaining them.

Descriptive research includes case studies, survey research and naturalistic observations.

Case studies

A case is a person with a particular characteristic. We can talk about cases of certain diseases, such as virus infection. Psychologists are interested in cases of different mental disorders.

In a case study, researchers investigate individuals in detail in order to understand the person’s conditions.

For example, we can investigate what factors contribute to the development of post-traumatic stress disorder (PTSD). Worldwide and throughout history, many soldiers have been signed to combat and many people have suffered trauma as victims of violent crime or natural disasters. They have a higher likelihood than the general population to develop Post Traumatic Stress Disorders. However, not all of them subsequently develop post-traumatic stress disorder. I am interested in why some people are more vulnerable for developing post-traumatic stress disorders and what kind of factors contribute to their development of this disorder.

To answer this question I want to study cases of soldiers who have returned from active duty in the battlefield and received PTSD diagnosis. I will gather detailed information from those individuals. I will analyze the different factors that might contribute to their vulnerability for developing PTSD.

I am interested in the level of stress or trauma the soldiers have experienced on the battlefield. Did they have combat experiences, did they fire their guns, did they get shot at, did they witness other soldiers or civilians being killed? These are the operational definitions of the level of stress they experience. My theory is that the higher stress they experience, the more likely they will develop post-traumatic stress disorder.

Another factor that interests me is their perception of danger or threat. Some people tend to exaggerate the level of a threat or danger they encounter. This trait might make them more likely to suffer from PTSD after a traumatic experience.

I am also interested in their environment after they return to their normal life, the level of social social support they receive may influence their subsequent adaptation and development of post-traumatic stress disorder. After the soldiers returned to civilian life, did they have a family to go back to and friends that support them? Was it easy for them to find civilian jobs? My theory is that strong support can protect people from developing post-traumatic stress disorder.

I am also interested in their pre-stress psychological adjustment. Before they encountered the traumatic event, were they well-adjusted, or already struggling with mental disorders? Were they already anxious or depressed? Those pre-existing conditions might contribute to their later development of post-traumatic stress disorder.

I am interested in all these factors. With each case I want to conduct detailed interviews and review their records to gather information about  those factors. I predict that I will see patterns and trends. I predict that people who later developed PTSD are more likely to have pre-existing conditions of maladjustment, such as anxiety or depression in their early development. I also predict that social support they receive after the traumatic experience of military deployment also contributes to their later development of PTSD.

Case studies are useful at the beginning of investigating a particular issue. A good example of case studies used for understanding new issues is this coronavirus infection which started at the end of 2019. At the beginning of the COVID-19 pandemic, health professionals did not know much about the disease. The only way they could gather information about the disease was to study the small number of cases confirmed at that time.

Case studies were also useful in the early investigation of the HIV infection in the early 1980s. AIDS is a disease of weakened immune system that was first reported in the 1980s. At that time scientists and doctors did not know what caused the diseases. They studied the few cases that they had access to in detail to look for patterns. At the beginning they found that most of the cases reported were from the gay men community. Based on this information they inferred that the AIDS disease might have something to do with a gay lifestyle.

A weakness of case studies is that the cases being studied are not representative samples of the population. The sampling is biased. As new cases accumulated, scientists discovered cases of AIDS among women and heterosexual men, falsifying the theory that AIDS was related to homosexuality. In 1983 scientists isolated HIV and then they understood that the virus was the cause of AIDS.

Science Practices Video Lesson Part 1

In this video I explain the concepts of scientific research, research issue, theory and hypothesis.

Science Practices Video Lesson Part 2

In this video I explain the scientific principle of empirical verification.

Science Practices Video Lesson Part 3

In this section, I explain different types of variables, including numerical and categorical variables and how to show values of variables.

Science Practices Video Lesson Part 4

In this video I explain the concepts of Descriptive and Correlational Research. Descriptive research includes case studies, surveys and naturalistic observation. Correlational research focuses on linear correlation.

Science Practices Video Lesson Part 5

In this video I elaborate the concepts of correlation research, including scatterplots, correlation coefficients, correlation direction and strength, and why correlation is NOT equal to causation.  

Science Practices Video Lesson Part 6

In this video I explain the concepts of experimental research, including independent and dependent variables, experimental and control groups, random sampling and assignment.

Science Practices Video Lesson Part 7

In this video I elaborate the principles of experimental research, especially the manipulation of the independent variable and control of potential confounding variables.

Science Practices Video Lesson Part 8

In this video I show an example of an experimental research, explaining how to design and experiment and how to analyze data collected from the experiment, such as calculating the means and standard deviations.

Complete the Multiple Choice Questions Quiz to test your mastery of the Sicence Practices materials.

Science Practices Video Lesson Part 9

In this video I explain the concepts of blinding in experimental design, descriptive statistic and statistical significance, which belongs to inferential statistics.  

---

Multiple Choice Questions Quiz – Science Practices