Source: Research Methods, Statistics, and Applications, Ch. 1–2
Tags: research methods, correlational research, experimental research, correlation coefficient, cross-sectional design, longitudinal design, third variable problem, psychology research designs
Difficulty: Introductory | Prerequisites: None
This material covers the foundational research designs used across all of psychology. Before you can evaluate any study's claims, you need to know how researchers collect data and what each method can (and cannot) tell you. These two chapters set up the vocabulary and logic you will use for the rest of the course. If you are coming in cold, this is the right place to start.
Psychologists use correlational methods to spot relationships between variables and experimental methods to test whether one variable causes a change in another. The type of design chosen determines what conclusions you can draw. Understanding validity, bias, and proper controls is what separates a strong study from a weak one.
Correlational research
A method designed to discover relationships (associations) among variables, without manipulating any of them. In simple terms, this means you are looking at whether two things tend to move together, not whether one causes the other.
Experimental research
A method aimed at establishing causal relationships between variables by deliberately manipulating one and measuring the effect on another. Think of it as the gold standard for answering "does X cause Y?"
Socially desirable responding
The tendency of participants to answer questions in a way that makes them look good rather than answering truthfully. In simple terms, people fudge their answers to seem more likeable, moral, or competent.
Correlation coefficient (r)
A numerical measure of the degree and direction of the relationship between two variables. It ranges from -1.00 to +1.00. The sign tells you the direction; the absolute value tells you the strength. Think of it as a single number that summarises how tightly two variables track each other.
Positive sign: as one variable increases, the other increases.
Negative sign: as one variable increases, the other decreases.
Third variable problem
A situation where an unmeasured variable accounts for the apparent relationship between two other variables. In simple terms, you think A and B are linked, but really some hidden variable C is driving both.
Cross-sectional design
A research design in which variables are measured at a single point in time. Useful when other methods would be unethical or impractical. Think of it as a snapshot: you measure everything once and compare.
Longitudinal design
A research design that measures the same variables of interest in multiple waves over time. Think of it as a time-lapse: you keep coming back to the same participants to see what has changed.
Measures the association between two or more variables without manipulating them.
Uses the correlation coefficient (r) to quantify the relationship.
r close to +1.00 or -1.00 = strong relationship.
r close to 0 = weak or no relationship.
Cannot establish causation. Two variables may be correlated because of a third, unmeasured variable (the third variable problem).
Common pitfall: assuming that because two things are correlated, one must cause the other.
Cross-sectional design: Data collected at one point in time. Quick and practical, but you cannot see how things change over time. Particularly useful when longitudinal or experimental methods would be unethical.
Longitudinal design: The same participants are measured repeatedly over weeks, months, or years. Lets you track change, but it is expensive and participants may drop out.
The researcher manipulates one or more variables (independent variables) and measures the effect on another variable (dependent variable).
Random assignment is used to place participants into groups by chance, so that any pre-existing differences are spread evenly.
Experimental group: Receives the treatment or change the independent variable represents.
Control group: Does not receive the treatment. Serves as the baseline for comparison.
Confederate: A person given a scripted role in the study so the social context can be controlled.
Resembles an experiment, but participants are assigned to groups on the basis of non-random criteria (e.g., soldiers who were in combat vs. those who were not).
Because assignment is not random, you cannot rule out pre-existing group differences. This weakens causal claims.
Correlational research is the backbone of epidemiology: this is how researchers first linked smoking to lung cancer (they could not ethically assign people to smoke). Experimental designs underpin drug trials, where random assignment and control groups let regulators decide whether a treatment works.
Students often think correlation implies causation. It does not. A correlation tells you two variables move together; it says nothing about which one (if either) drives the other.
Students confuse cross-sectional and longitudinal designs. Remember: cross-sectional is one snapshot; longitudinal is repeated measurement over time.
Students sometimes believe a quasi-experiment is just a badly run experiment. It is a deliberate design choice used when random assignment is impossible or unethical.
Students mix up the independent and dependent variables. The independent variable is what the researcher manipulates; the dependent variable is what gets measured as the outcome.
Be ready to identify the type of research design from a scenario description. The exam will likely give you a short study summary and ask you to classify it.
Know the difference between correlational and experimental research, and be able to explain why correlational findings cannot establish causation.
Expect questions on cross-sectional vs. longitudinal designs, including their strengths and limitations.
Understand why random assignment matters and what happens to causal claims when it is absent (quasi-experimental design).
True or False: A correlation coefficient of -0.85 indicates a weak relationship. (False, it is strong; the sign indicates direction, not strength.)
Fill in the blank: In a(n) ______ design, the same participants are measured at multiple points over time. (Longitudinal)
True or False: Random assignment is used in quasi-experimental designs. (False, quasi-experimental designs use non-random assignment.)
Fill in the blank: The ______ variable is the one the researcher manipulates. (Independent)
True or False: Cross-sectional designs are useful when other methods would be unethical. (True)
Q: A researcher finds that ice cream sales and drowning deaths are positively correlated. Can the researcher conclude that eating ice cream causes drowning? Why or why not?
A: No. Correlation does not establish causation. A third variable, such as hot weather, likely drives both ice cream sales and swimming (and therefore drowning risk) upward at the same time.
Q: What is the key difference between a true experiment and a quasi-experiment?
A: In a true experiment, participants are randomly assigned to conditions. In a quasi-experiment, group membership is determined by pre-existing criteria (e.g., combat veterans vs. non-combat veterans), so the researcher cannot rule out that the groups differed before the study began.
Q: A study measures anxiety levels and social media use in 500 university students at one point in time. What type of design is this, and what is its main limitation?
A: This is a cross-sectional, correlational design. Its main limitation is that it cannot establish causation or track how the variables change over time.
Q: Why is random assignment important in experimental research?
A: Random assignment distributes pre-existing participant differences evenly across conditions, which helps ensure that any observed effect on the dependent variable is due to the manipulation of the independent variable rather than to confounding factors.
This material connects directly to validity and bias (covered in the second document in this set), because the design you choose determines how vulnerable your study is to threats like confounds and experimenter bias. It also lays the groundwork for understanding statistics later in the course: the correlation coefficient is one of the first inferential statistics you will encounter.
Correlation, causation, correlation vs. causation, r value, positive correlation, negative correlation, third variable, lurking variable, confound, cross-sectional study, longitudinal study, experiment, quasi-experiment, natural experiment, IV, DV, independent variable, dependent variable, random assignment, control group, experimental group, confederate, research design, psychology research methods