Research Methods -- Part 1

Dependent Variable (DV)

The variable that is measured or observed in an experiment to assess the effect of the independent variable. It is the presumed effect or outcome in a cause-effect relationship. Changes in the DV are assumed to result from manipulation of the IV.

Extraneous Variable

Any variable other than the independent variable that could potentially influence the dependent variable. These must be controlled or held constant to ensure that observed effects can be attributed solely to the IV. Also called confounding variables when not controlled.

Independent Variable (IV)

The variable that is systematically manipulated by the researcher in an experiment to observe its effect on the dependent variable. It represents the presumed cause in a cause-and-effect relationship and must have at least two levels or conditions to allow for meaningful comparison.

Dependent Variable (DV)

The variable that is measured or observed in an experiment to assess the effect of the independent variable. It represents the presumed effect or outcome in a cause-and-effect relationship. Researchers must operationally define how the DV will be measured to ensure reliability and validity.

Operational Definition

A precise specification of how a variable or construct will be measured or manipulated in a particular study. It translates abstract theoretical concepts into concrete, observable, and measurable terms. Essential for replication and allows other researchers to understand exactly what was studied.

Internal Validity

The degree to which a study establishes a causal relationship between variables by ruling out alternative explanations. High internal validity means changes in the DV can be confidently attributed to manipulation of the IV rather than confounding variables. Critical for experiments but can be threatened by history, maturation, selection, mortality, testing, instrumentation, and regression to the mean.

External Validity

The extent to which research findings can be generalized beyond the specific conditions of the study to other populations, settings, and times. Involves both population validity (generalizability to other people) and ecological validity (generalizability to other settings and conditions). Often trades off with internal validity.

Reliability

The consistency and stability of measurement over time, across items, and between raters. Includes test-retest reliability (consistency over time), internal consistency (agreement among items), interrater reliability (agreement between observers), and split-half reliability. High reliability is necessary but not sufficient for validity.

Construct Validity

The degree to which a measurement instrument or operational definition accurately measures the theoretical construct it is intended to measure. Established through convergent validity (correlates with related measures), discriminant validity (does not correlate with unrelated measures), and factor analysis. Essential for ensuring that research truly examines the intended psychological concepts.

Random Assignment

A procedure in which all participants have an equal chance of being assigned to any condition or group in an experiment. This control technique minimizes systematic differences between groups and is the hallmark of true experiments. Helps equate groups on both known and unknown variables, thereby increasing internal validity and enabling causal inferences.

Random Sampling

A procedure in which every member of the population has an equal chance of being selected for the sample. Enhances external validity and representativeness, allowing researchers to generalize findings to the broader population. Distinguished from random assignment, which relates to internal validity. Includes techniques like simple random sampling and stratified random sampling.

Counterbalancing

A control technique used in within-subjects or repeated measures designs to control for order effects. Involves systematically varying the order in which participants experience different conditions. Can use complete counterbalancing (all possible orders) or partial/Latin square counterbalancing. Helps eliminate practice effects, fatigue effects, and carryover effects.

Experimental Research

Research method that involves manipulation of an independent variable, random assignment of participants to conditions, and control of extraneous variables. The only method that can definitively establish causation. Features include experimental and control groups, pre- and post-tests, and systematic control procedures. High internal validity but may sacrifice external validity.

Quasi-Experimental Research

Research that resembles experimental research but lacks random assignment to conditions. Used when random assignment is impossible or unethical (e.g., comparing naturally occurring groups). Includes nonequivalent control group designs and time-series designs. Weaker internal validity than true experiments but greater practical utility for real-world settings.

Correlational Research

Nonexperimental research method that examines relationships between two or more variables without manipulation. Measures degree and direction of relationships using correlation coefficients ranging from -1.00 to +1.00. Cannot establish causation due to directionality problem and third-variable problem. Useful for prediction and identifying variables for experimental study.

Informed Consent

Ethical principle requiring that research participants be fully informed about the nature of the research, procedures, potential risks and benefits, and their right to withdraw at any time before agreeing to participate. Must be voluntary, competent, and based on adequate information. Exceptions include minimal risk studies and deception research with debriefing.

Pseudoscience

Claims or practices that appear scientific but lack empirical support and do not follow scientific methods. Characteristics include reliance on anecdotal evidence, unfalsifiable claims, lack of self-correction, overuse of ad hoc hypotheses, and avoidance of peer review. Examples include astrology, ESP claims without evidence, and facilitated communication. Contrasts with genuine science's reliance on systematic observation and falsifiability.

APA Ethics Code

The American Psychological Association's ethical principles that guide research with human and animal subjects. Key principles include: respect for persons and autonomy, beneficence (do no harm), justice (fair treatment), informed consent, confidentiality, and debriefing. Researchers must obtain IRB approval.

Informed Consent

An ethical requirement that participants voluntarily agree to participate in research after being fully informed about the study's purpose, procedures, risks, benefits, and their right to withdraw at any time without penalty. Participants must be competent to give consent and must not be coerced.

Pseudoscience

Claims or practices that appear scientific but lack empirical support, testability, or adherence to the scientific method. Examples include astrology and ESP. Distinguishing features: unfalsifiable claims, reliance on anecdotal evidence, lack of peer review, and resistance to revision based on evidence.

Theory

An organized system of ideas that explains observations and generates testable hypotheses. Good theories are falsifiable, parsimonious (simple), and have broad explanatory power. Theories are supported or refuted by empirical evidence and guide future research.

Hypothesis

A testable prediction about the relationship between two or more variables. It is derived from theory and states what the researcher expects to find. Hypotheses can be directional (predicting the direction of an effect) or non-directional (simply predicting that an effect exists).

Variable

Any characteristic, attribute, or condition that can take on different values or levels. Variables can be independent (manipulated by the researcher), dependent (measured outcome), or extraneous (potential confounds that need to be controlled).

Confounding Variable

An extraneous variable that systematically varies along with the independent variable, making it impossible to determine which variable is actually causing changes in the dependent variable. Confounds threaten internal validity by providing alternative explanations for research results.