PSYC100 slides 8

Welcome to PSYC100: Understanding Research Designs

Retrieval Practice Prompts

To challenge yourself and ensure retention, try to answer the following without looking at your notes:

• Can you recall the timeline of the foundations of psychology?

• How would you describe the three validities in research?

Overview of Research Designs

Today, we'll talk about the three main ways psychologists design their studies to learn about behavior and the mind:

• Descriptive methods: These are like taking a picture of what's happening.

• Correlational methods: These look for connections between different things.

• Experimental methods: These are used to figure out if one thing causes another.

Part 1: Descriptive Methods

Descriptive methods are used to simply describe behavior, thoughts, or feelings as they naturally occur. They don't try to change anything or find causes; they just report facts. They provide a clear description of a phenomenon.

Types of Descriptive Methods:

• Observational studies: This is when researchers watch and record behavior without interfering, usually in its natural environment.

  • Definition: Observing and documenting behavior in its original setting.

  • Example: A researcher watches children on a playground and writes down how often they share toys. They don't tell the children what to do; they just observe.

• Self-reports: Participants give information about themselves, often through surveys, questionnaires, or interviews.

  • Definition: Gathering information directly from individuals about their thoughts, feelings, or behaviors.

  • Example: A survey asking students about their study habits or an interview asking people about their favorite types of music.

• Case studies: This involves a very detailed, in-depth look at one specific person, a small group, or a unique event.

  • Definition: A thorough investigation of a single subject (person, group, or event) to understand complex details.

  • Example: Studying a patient with a rare brain injury to understand how that injury affects language abilities.

Part 2: Correlational Methods

Correlational methods are used to explore whether two or more things are related and how they change together. They help us see if changes in one variable tend to go along with changes in another.

Operational Definitions

Before we can study relationships, we need to be very clear about how we're going to measure each variable. An operational definition specifies the exact steps or procedures used to measure or manipulate a variable. It turns an abstract concept into something measurable.

• Variable 1: Distance To Professor

• Variable 2: Academic Performance

An operational definition for "Distance To Professor" could be: the average number of feet a student sits from the front of the classroom during lectures. An operational definition for "Academic Performance" could be: the total points earned on all exams in a specific course.

Examples of Correlational Studies

• FORM 1: "Is there a relationship between how far you sit from the professor and the points you earn on the exam?"

• FORM 2: "Is there a relationship between a person's IQ level and their brain size?" (It's important to remember that in correlational studies, researchers do not assign people to specific groups. They just measure existing characteristics or groups.)

The Correlation Coefficient

The strength and direction of a linear relationship between two variables is shown by a number called the correlation coefficient, which is often written as $r$.

• Range: The correlation coefficient can be any value from $-1.0$ to $+1.0$.

• Interpretation:

  • $+1.0$: This means a perfect positive correlation. As one variable goes up, the other variable goes up by exactly the same proportion. They move in perfect sync in the same direction.

  • $+.5$: This indicates a medium positive correlation. Both variables tend to increase together, but not perfectly.

  • $0$: This means no linear correlation. There's no consistent straight-line relationship between the variables. Changes in one don't predict changes in the other in a linear way.

  • $-.7$: This suggests a medium negative correlation. As one variable increases, the other tends to decrease, and vice versa. They move in opposite directions to a moderate degree.

  • $-1.0$: This signifies a perfect negative correlation. As one variable increases, the other decreases by exactly the same proportion. They move in perfect sync in opposite directions.

Examples of Correlation Interpretation:

• If a researcher finds that the more students use metacognition (thinking about their thinking), the higher their grades are, this would show a positive correlation. For instance, an $r$ value of $.6$ would mean that as metacognition increases, grades tend to increase.

• If a researcher finds that the less often parents used diapers, the less times their babies had urinary tract infections, this would also depict a positive correlation. Why positive? Because both variables are decreasing together. (If one decreased and the other increased, it would be negative).

Why Correlation Does Not Equal Causation

This is a super important point in psychology: just because two things are related (correlated) does not mean one causes the other. There are two main reasons why we can't make cause-and-effect claims from correlational studies:

1. Directionality Problem: We can't tell which variable is causing which. Even if A and B are related, we don't know if A causes B, or if B causes A.

   • Example: If we find a correlation between sitting closer to the professor and getting higher exam points, we don't know if:

     • Sitting closer to the professor (A) causes higher exam points (B).

     • OR, if being a good student who gets higher exam points (B) causes them to choose to sit closer to the professor (A).

2. Third Variable Problem: An unmeasured or hidden third variable might actually be causing both of the observed variables to change, making them look like they're related when they aren't directly causing each other.

   • Example: The correlation between distance from the professor and exam points could actually be due to a "student motivation" (third) variable. Highly motivated students might sit closer and also study more, leading to better grades. So, motivation is affecting both factors.

   • Another Big Example: A strong positive correlation is often found between homicide rates and ice cream sales. Does eating ice cream make people commit crimes? No. The third variable here is heat. When it's hot, people buy more ice cream, and higher temperatures are also linked to more social interactions and sometimes increased aggression, leading to higher crime rates. Heat causes both, not that one causes the other.

Part 3: Experimental Methods

Experimental methods are the only research design that can establish a cause-and-effect relationship; they are designed to figure out what causes an outcome. Researchers actively manipulate one thing to see if it directly changes another.

Key Principles of Experimental Design

• Random Assignment: This is a crucial step where participants are sorted into different groups in the experiment purely by chance (like flipping a coin). This helps ensure that, on average, all the groups are pretty much the same at the beginning of the study, so any differences we see later are likely due to our manipulation, not pre-existing differences.

• Independent Variable (IV): This is the variable that the experimenter manipulates or changes. It's the "cause" that the researcher is testing.

  • Example: If studying the effect of different teaching methods, the teaching method (e.g., lecture vs. group work) would be the IV.

• Dependent Variable (DV): This is the variable that the experimenter measures to see if it was affected by the IV. It's the "effect" or outcome.

  • Example: In the teaching methods study, the students' exam scores (how well they learned) would be the DV.

• Experimental Group: This is the group of participants who receive the specific treatment or manipulation being tested (the IV).

• Control or Comparison Group: This group either receives no treatment, a standard treatment, or a placebo. They serve as a baseline to compare against the experimental group, helping us see if the manipulation had a real effect.

Experimental Design Diagram

1. Population of interest (e.g., all college students)

2. Sample (a smaller group selected from the population)

3. Randomly assigned to groups within the study:

   • Control Group (e.g., takes notes by hand, DV measured: exam scores)

   • Experimental Group (e.g., takes notes on a laptop, DV measured: exam scores)

Application Exercise (Based on a "Brain Games" Episode)

When you look at a study or demonstration, ask yourself these questions:

• What is the IV? How was it manipulated? (What did the researchers intentionally change or vary?)

• What is the DV? How was it measured? (What was the outcome they measured to see if it changed?)

• What are some potential confounds? (What other uncontrolled things might have influenced the results, making it hard to be sure the IV was the only cause? These are like hidden variables that can mess up your experiment.)

• What kind of validity is affected by having all of these confounds? Confounds primarily hurt internal validity. This refers to how well an experiment shows that the IV actually caused the change in the DV. If there are many confounds, it's difficult to be confident that the IV, and not something else, was the real reason for the observed effect.