Research Methods in Social Sciences: Research Designs, Causality, Measurement, and Sampling

Comprehensive vocabulary flashcards covering research design types, causality criteria, measurement levels, validity and reliability types, and sampling fundamentals.

Cross-Sectional Research

A research design where findings are based on data collected at $1$ point in time (a "snapshot"); it provides quick, affordable insights into phenomena at any time, but cannot document change over time.

Longitudinal Research

A research design where data is collected at multiple points over an extended period of time to track how populations and social phenomena change; limitations include extensive time (decades), higher cost, and the necessity of keeping procedures consistent for comparisons.

Trend Study

A type of longitudinal design that collects the same data at different times from different samples of the same population (repeated cross-sectional studies or repeated snapshots).

Panel Study

A longitudinal study design in which data is collected from the exact same individuals over multiple points in time.

Cohort Study

A specific type of panel study focusing on a group of individuals who share a common characteristic or experience, such as a birth cohort, school cohort, or seniority cohort.

Quantitative Research

A research approach that measures variables numerically.

Qualitative Research

A research approach that examines variables in a natural setting and probes their underlying meanings through words and thematic exploration.

Unit of Analysis

The specific entity—such as an individual, group, community, or nation—that is being studied in research.

Ecological Fallacy

A logical error that occurs when a researcher draws conclusions about individuals based on aggregate data from groups.

Reductionist Fallacy

A logical error (also known as reductionism) that occurs when a researcher draws conclusions about groups based on observations of individuals.

Temporal Order

A criterion for establishing causality requiring that the cause (independent variable) must precede the effect (dependent variable) in time.

Correlation

An association or relationship between an independent variable and a dependent variable; a required condition for establishing causality.

Positive Correlation

A relationship where variables change in the same direction: an increase in one variable leads to an increase in the other, or a decrease in one variable leads to a decrease in the other.

Negative Correlation

A relationship where variables change in opposite directions: an increase in one variable leads to a decrease in the other, or a decrease leads to an increase (e.g., as age increases, participation in crime decreases).

Pearson Correlation Coefficient

A measure of the type and strength of correlation between variables, spanning from $-1$ (strong negative correlation) to $0$ (no correlation) to $+1$ (strong positive correlation).

Lack of Plausible Alternatives

A criterion for causality met when a researcher eliminates all other possible causes for change in the dependent variable (DV), establishing the independent variable (IV) as the sole cause.

Spuriousness

A relationship where two variables appear to have a cause-and-effect link when they actually do not.

Measurement

The process by which social scientists determine how to assess and evaluate the phenomena they want to study.

Conceptualization

The process of specifying definitions and establishing what the main concepts mean in a research study.

Operationalization

The process of detailing how the main concepts are measured in practice (e.g., recording screen time per app or counting interpersonal interactions).

Nominal Level of Measurement

A discrete level of measurement using names and labels that cannot be arranged in any mathematical order or sequence (e.g., sex, gender, religion, name of town, group affiliation).

Ordinal Level of Measurement

A discrete level of measurement with categories that can be rank-ordered, but without precise or equal distances between attributes (e.g., likelihood to reoffend, Likert scales, course grades, happiness).

Interval Level of Measurement

A continuous level of measurement where categories can be arranged in a logical order with equal distances between attributes, but without an absolute zero (e.g., SAT scales, personality scores, education in years).

Ratio Level of Measurement

The most precise, continuous level of measurement that contains labels, logical ordering, equal intervals, and a true absolute zero (e.g., income, number of times recidivates, number of responses).

Measurement Validity

The ability of an instrument to accurately measure the specific concept it claims and intends to measure.

Face Validity

A subtype of validity evaluating whether the items of a measurement logically make sense on the surface.

Criterion Validity

The extent to which a measure relates to another, more direct measure of the same concept (e.g., checking a self-report of alcohol consumption against a breathalyzer test).

Construct Validity

The extent to which a measure corresponds to other variables related to an underlying theoretical construct (e.g., whether light acts as a capable guardian in Routine Activities Theory).

Content Validity

A form of validity established when a measure includes all dimensions and facets of a given concept (e.g., including all aspects of alcoholism in a survey).

Reliability

The ability of a measure or study to consistently yield the same results when assessing a concept.

Test-Retest Reliability

A method of testing reliability by administering the same measure twice to the same group across two points in time to verify that the results are highly correlated.

Inter-Rater Reliability

A reliability procedure where multiple researchers independently measure a concept and compare observations (useful in field research, such as multiple observers assessing bullying behavior on a schoolyard).

Split-Halves Reliability

A method of testing reliability where an instrument's measures are split in half and administered to two groups to determine if they produce similar results.

Benchmarks and Baselines

The practice of using previous research or generally agreed-upon standards to test the reliability of a new measure (e.g., comparing a new IQ test's results against the original test).

Sampling

A scientific technique that allows a researcher to learn about an entire population by studying a subset of that population.

Central Limit Theorem

A statistical theorem stating that repeated samples drawn from the same population will be similar, producing a normal distribution where most samples fall in the middle.

Population

The entire set of individuals relevant to a researcher's study (e.g., all students).

Census

A complete enumeration collecting data from every single member of a population (e.g., all students at NAU).

Sample

A subset selected from a larger population to be studied directly (e.g., $500$ students).

Sampling Frame

The actual list of individuals in the population from which a researcher draws a sample (e.g., a registrar's list of NAU students).

Sample Units

The specific individuals or total count of participants in a study (e.g., $500$ NAU students).

Sampling Bias

A condition occurring when a sample is unrepresentative of the population it is intended to reflect.

Random Sampling Error

An unavoidable error always present in sample studies because researchers cannot observe every single individual in a population.

Selection Bias

A systematic flaw in sampling procedures that increases or decreases the likelihood that certain members of the population are chosen for the sample.

Level of Precision

A measure of a sample's representativeness, determined by the size of the sample relative to the population's size and internal variability.

Probability Sampling

A sampling technique relying on random selection, which allows researchers to generalize findings and predict patterns in the larger population.

Simple Random Sampling

A probability sampling method where researchers identify the sampling frame and units, then randomly select participants (e.g., using numbers out of a hat, coin flips, or number generators) so that every person has an equal chance of being selected.

construct validity

The extent to which a test or instrument measures the theoretical construct it is intended to measure, ensuring that the results accurately reflect the underlying concept.

face validity

The degree to which a test appears to measure what it is supposed to measure, based on subjective judgment rather than statistical analysis.

criterion validity

The extent to which a measure correlates with an outcome considered to be a gold standard or criterion, indicating how well the test predicts an individual's performance in a specific area.

content validity

The degree to which a test covers the entire content it is intended to measure, ensuring that all relevant dimensions of the construct are assessed comprehensively.

Test-retest reliability

The consistency of a test's results when administered at different times to the same group, indicating the stability of the measure over time.

inter-rater reliability