Types of Research Methods

Why This Matters

Research methods are the backbone of everything you'll study in AP Psychology. Every claim about memory, development, emotion, or social behavior rests on how psychologists gathered their evidence. When you encounter a study about infant depth perception (like the visual cliff) or implicit attitudes (like the IAT), you're being tested not just on what researchers found, but on why they chose that particular method and what conclusions it actually supports.

The AP exam tests your ability to evaluate research quality and identify methodological limitations. Can this study establish cause and effect? What's a potential confound? Why might results not generalize? These questions come up across every unit, from operant conditioning to stereotype formation. Don't just memorize method names; know what each method can and cannot tell us about human behavior.

---

Establishing Cause and Effect

These methods involve manipulation and control, allowing researchers to determine whether one variable actually causes changes in another. The key mechanism is isolating variables so that alternative explanations can be ruled out.

Experimental Method

The experimental method is the only research design that can establish a true cause-and-effect relationship. It works because the researcher actively changes one thing (the independent variable, or IV) and measures the effect on another (the dependent variable, or DV), while holding everything else constant.

• Random assignment is what makes this possible. Participants are placed into conditions by chance (like flipping a coin), so any pre-existing differences between groups are distributed evenly.
• Control groups provide a baseline for comparison, helping isolate the effect of the IV while minimizing confounding variables (outside factors that could offer alternative explanations).
• Replication strengthens findings. If results can be reproduced across different samples and settings, confidence in the causal relationship increases.

Quasi-Experimental Design

Sometimes random assignment isn't possible. A quasi-experimental design still compares groups and measures outcomes, but the groups are pre-existing rather than randomly formed.

• Researchers compare groups that already differ in some way (e.g., smokers vs. non-smokers, different age cohorts).
• This design is useful when true experiments aren't feasible or ethical, such as studying the effects of childhood trauma or comparing developmental stages. You can't randomly assign someone to experience trauma.
• The tradeoff: quasi-experiments are more vulnerable to confounds because group differences may have existed before the study began. Results suggest but don't prove causation.

---

Measuring Relationships Without Manipulation

These methods examine associations between variables but cannot determine causation. The core principle: observing a relationship doesn't tell you why it exists or which variable (if either) influences the other.

Correlational Method

A correlational study measures two or more variables and calculates how strongly they're related, without manipulating anything.

• The correlation coefficient ($r$) ranges from $-1$ to $+1$. It tells you two things: direction (positive means variables move together; negative means they move in opposite directions) and strength (closer to $-1$ or $+1$ means a stronger relationship; closer to $0$ means weaker).
• Does not establish causation. A correlation between sleep and test scores doesn't tell you if poor sleep causes lower scores, if stress causes both, or if the relationship is coincidental.
• Third variables (confounds) may explain observed relationships, which is why correlational findings always require cautious interpretation.

Cross-Sectional Studies

A cross-sectional study compares different groups at a single point in time. It's like taking a snapshot.

• For example, a researcher might test memory in 20-year-olds, 40-year-olds, and 60-year-olds all during the same week.
• This approach is efficient and cost-effective compared to tracking the same people over decades, making it practical for studying age-related differences.
• The major limitation is cohort effects: differences between age groups may reflect generational experiences (e.g., growing up with or without the internet) rather than actual developmental changes.

---

Tracking Change Over Time

These methods follow the same participants across extended periods, allowing researchers to observe developmental trajectories and long-term effects. The key advantage is capturing within-person change rather than just between-group differences.

Longitudinal Studies

A longitudinal study makes repeated observations of the same participants over months, years, or even decades. This is essential for studying development, as seen in research on critical periods for language acquisition.

• Because you're tracking the same people, you can observe whether Variable A precedes and predicts Variable B over time. This helps identify causal patterns, though it still can't prove causation as firmly as a true experiment.
• Participant attrition (dropout) is a major threat. People who stay in the study may differ systematically from those who leave (e.g., healthier or more motivated participants stick around), potentially biasing results.

---

Observing Behavior in Context

These methods prioritize ecological validity, meaning they study behavior as it naturally occurs rather than in artificial laboratory settings. The tradeoff is that reduced control makes it harder to isolate specific causes.

Naturalistic Observation

Researchers observe subjects in their natural environment without interference. They watch and record behavior as it spontaneously occurs (e.g., studying children's play on a playground or primate social hierarchies in the wild).

• High ecological validity means findings reflect real-world behavior, but the lack of control makes cause-and-effect conclusions impossible.
• Observer bias can distort findings if researchers interpret ambiguous behaviors in ways that confirm their expectations. Using multiple observers and checking for inter-rater reliability helps reduce this problem.
• Ethical concerns about privacy must also be addressed, especially when subjects don't know they're being observed.

Case Studies

A case study is an in-depth analysis of a single individual, group, or event. These are critical for studying rare phenomena like the case of Genie (language critical periods) or patients with unique brain injuries (like Phineas Gage).

• They produce rich qualitative data that can generate hypotheses and reveal complexities that large-scale studies miss.
• The main weakness is limited generalizability: findings from one unusual case may not apply to broader populations. Researcher subjectivity can also influence interpretation.

---

Gathering Self-Report Data

This method relies on participants reporting their own thoughts, feelings, and behaviors. The key challenge is that what people say may not match what they actually think or do.

Surveys and Questionnaires

Surveys collect data from large samples efficiently, reaching thousands of respondents through online, paper, or interview formats. They're especially useful for measuring attitudes, beliefs, and self-reported behaviors that can't be directly observed.

• Response biases are the biggest threat to validity:
  • Social desirability bias: answering in ways that make you look good (e.g., underreporting drug use)
  • Acquiescence bias: the tendency to agree with statements regardless of content
• Careful question design matters. Leading questions, confusing wording, or loaded terms can all skew results. The way a question is phrased (called framing) can significantly change how people respond.

---

Isolating Nature vs. Nurture

This method specifically addresses the heritability question: how much of a trait is attributable to genetics versus environment?

Twin Studies

Twin studies compare individuals with different degrees of genetic similarity to estimate genetic influence on traits.

• Identical (monozygotic) twins share 100% of their DNA, while fraternal (dizygotic) twins share roughly 50%. If identical twins are more similar on a trait (like intelligence or risk for schizophrenia) than fraternal twins are, genetics likely plays a role.
• Studies of twins raised apart are especially powerful because they help separate genetic influence from shared environment. These have been foundational for understanding the heritability of intelligence, personality, and psychological disorders.
• Assumptions can be problematic. The "equal environments assumption" holds that identical and fraternal twins experience equally similar environments. But identical twins may be treated more similarly than fraternal twins (dressed alike, placed in the same classes), which could inflate apparent genetic effects.

---

Synthesizing Existing Research

This method doesn't collect new data but instead combines and analyzes findings from multiple studies. The power comes from aggregating evidence to identify patterns that single studies might miss.

Meta-Analysis

A meta-analysis statistically combines results from many studies on the same research question. Think of it as a "study of studies."

• By pooling data, meta-analyses increase statistical power and provide a more reliable estimate of true effect sizes (how large or meaningful an effect actually is).
• They can resolve conflicting findings by weighing studies based on sample size and methodological quality. This makes them essential for evidence-based practice in clinical psychology.
• The main limitation: quality depends on the studies included. If the original studies are poorly designed, the meta-analysis inherits those flaws. Researchers must also watch for publication bias, since studies with significant results are more likely to be published than those without.

---

Quick Reference Table

---

Self-Check Questions

1. A researcher wants to determine whether sleep deprivation causes decreased memory performance. Which method should they use, and what feature of that method makes causal conclusions possible?

2. Compare and contrast longitudinal and cross-sectional studies: What question can each answer, and what is the primary limitation of each approach?

3. A study finds a strong positive correlation ($r = 0.72$) between hours of social media use and anxiety levels in teenagers. Why can't researchers conclude that social media causes anxiety? Name two alternative explanations.

4. Which two methods would be most appropriate for studying a rare condition like prosopagnosia (face blindness), and what are the tradeoffs of each?

5. An FRQ asks you to evaluate a study comparing stress levels in nurses versus accountants. Is this an experimental or quasi-experimental design? What confounding variables might explain any differences found?