Research Methods Flashcard Study Guide

Comprehensive vocabulary flashcard set covering foundational research methods, causation, research designs, measurement, experiments, sampling, and survey design.

Function of theory in research

Guides what to study, explains why patterns happen, organizes findings, and generates testable hypotheses.

Traditional science model

Cycle of theory → testable hypotheses → operationalization of concepts → empirical testing → improved theoretical understanding.

Hypothesis

A specific, testable, falsifiable (usually directional) prediction about what you should observe if a theory is right.

Null hypothesis

The claim that there is no relationship between variables; researchers test it and try to reject it.

Falsifiable

Could turn out to be wrong. If nothing you could observe would count against a claim, it isn't a hypothesis.

Proposition

A general claim about the relationship you expect between concepts.

Deductive method

Starts with a general theory and ends with a specific testable prediction.

Inductive method

Starts with specific observations, notices a pattern, and builds a general claim.

Pure sociology

Research aimed at advancing knowledge and theory for its own sake.

Applied sociology

Research aimed at solving practical problems or evaluating programs and policies.

Sociological research

Systematic, empirical study of society and social behavior rather than relying on personal experience or assumptions.

Empirical evidence

Observable, measurable evidence, as opposed to opinion or common sense.

Exploratory study

Used when a topic is new or poorly understood; asks "what's going on here?"

Descriptive study

Describes what something looks like (how much, how often, how many).

Explanatory study

Asks why something happens; the most difficult research purpose.

Idiographic explanation

Full understanding of one case, covering every detail that produced that outcome for those people.

Nomothetic explanation

A partial account of many cases, using a few variables to explain some of the pattern across a lot of people.

Complete causation

A cause explains every case with no exceptions; rare in social science.

Probabilistic explanation

A cause makes an outcome more or less likely rather than guaranteeing it.

Majority of cases

Social patterns describe what is true for most cases, not all.

Exceptional cases

Individual cases that don't fit the pattern; one exception doesn't break a probabilistic rule.

Suppressed evidence

Ignoring or leaving out evidence that contradicts a claim; closely related to selective observation.

Correlation

Two variables change together. Required for a causal claim, but does NOT prove causation.

Time order

The cause must come before the effect; a one-shot survey often can't establish this.

Nonspuriousness

The relationship isn't produced by a hidden third variable that causes both.

Necessary cause

Must be present for the effect to occur (you must be enrolled in a class to skip it).

Sufficient cause

When present, it guarantees the effect every time.

Independent variable

The presumed cause or explanatory variable.

Dependent variable

The outcome being explained.

Error of generalization (overgeneralization)

Assuming what is true for some cases is true for all; a conclusion problem.

Selective observation

Noticing only evidence that supports what you already believe; an attention problem.

Agreement reality

Something accepted as true because people collectively agree it is.

Experiential reality

Something believed true from personal experience.

Ecological fallacy

Drawing conclusions about individuals from group-level (aggregate) data.

Individualistic fallacy

Using individual exceptions to reject a group-level pattern ("my grandpa smoked and lived to 95").

Reductionist fallacy (reductionism)

Oversimplifying by explaining a complex phenomenon with a single narrow cause (e.g., sociobiology: genes explain all social behavior).

Aggregate vs. individual

Sociology identifies group-level patterns that don't necessarily predict any one person's behavior.

Unit of analysis

The thing being studied and described: individuals, groups, organizations, or social artifacts.

Social artifacts

Things people make, such as posts, articles, or songs.

Cross-sectional study

Data collected at one point in time; the cheapest design.

Trend study

Same general population measured repeatedly, with different people each time (freshmen every 5 years).

Cohort study

One subpopulation followed over time, with different samples from it each time (class of 2029 each year).

Panel study

The same people are resurveyed over time.

Quasi-experimental study

Compares groups to test causal ideas without random assignment, often using naturally occurring groups or events.

Conceptualization

Specifying exactly what you mean by an abstract concept.

Nominal definition

A definition assigned to a term; an agreed-upon meaning that is neither "true" nor "false."

Operational definition

Spells out exactly how a concept will be measured.

Operationalization

Turning an abstract concept into something measurable.

Progression of measurement steps

Conceptualization → nominal definition → operational definition → measurements in the real world.

Interchangeability of indicators

Different valid indicators of the same concept should substitute for each other and give similar results.

Single vs. multiple indicators

One measure of a concept versus several measures combined.

Nominal level of measurement

Categories only, no order (on campus / near campus / commuter).

Ordinal level of measurement

Ordered categories without fixed distances between steps (never / rarely / sometimes / often).

Interval level of measurement

Ordered with equal distances but no true zero (clock time of earliest class).

Ratio level of measurement

Equal distances with a true zero, so ratios make sense (classes skipped, hours worked).

Reliability

The measure gives consistent results.

Validity

The measure actually measures what it is supposed to measure.

Reliability assessment

Checking consistency using methods such as test-retest or internal consistency.

Test-retest reliability

The same measure given to the same people at two times gives similar results.

Internal consistency

Items intended to measure the same concept correlate with each other.

Validity assessment

Checking whether a measure captures the intended concept (face, content, criterion-related, construct).

Face validity

The measure looks like it measures the concept, based on common sense.

Content validity

The measure covers the full range of meanings of the concept.

Criterion-related validity

The measure is checked against an external criterion.

Predictive validity

A type of criterion-related validity: the measure predicts a future outcome (SAT predicting college GPA).

Construct validity

The measure relates to other variables in the way theory says it should.

Precision vs. accuracy

A precise measure isn't necessarily accurate; a scale always 5 lb too high is reliable but not valid.

Measurement validity (Pao et al.)

Do response options actually capture the concept? Open-ended "other" options can reveal categories fixed choices miss.

Closed-ended question

Researcher provides predetermined answer categories.

Open-ended question

Respondents answer in their own words.

Coding / recoding

Converting responses into categories that can be analyzed.

Classical experiment

Experimental group: pretest → treatment (X) → posttest. Control group: pretest → no treatment → posttest.

Experimental group

The group that receives the treatment.

Control group

The comparison group that does not receive the treatment.

Pretest / posttest

Measurement of the dependent variable before and after the treatment.

Effect

The change in the dependent variable attributable to the treatment.

Randomization (random assignment)

Chance decides group membership, making groups alike on measured and unmeasured traits.

Matching

Pairing similar subjects on key traits, then splitting each pair between groups.

Double-blind experiment

Neither the subjects nor the people assessing outcomes know who received which treatment.

Preexperimental designs

Designs lacking the safeguards of a true experiment (one-shot case study, one-group pretest-posttest, static-group comparison).

One-shot case study

Treat one group, then measure; nothing to compare to.

One-group pretest-posttest

Measure, treat, measure one group; anything else could explain the change.

Static-group comparison

Compare a treated group to an untreated one without random assignment.

Posttest-only control group design

Random assignment with a posttest only and no pretest; works with large randomized samples.

Internal validity

Whether the experimental stimulus really caused the change in the dependent variable.

External validity (generalizability)

Whether findings can be applied beyond the study sample or setting.

Statistical regression

Extreme scores tend to move toward the average when remeasured (99th percentile GRE dropping to 90th).

Testing effect

Taking a pretest influences posttest performance.

Reactivity / Hawthorne effect

Knowing you're being watched changes your behavior.

Natural experiment

An existing process (e.g., a lottery) determines who receives a treatment; researchers compare outcomes.

Web-based experiment

A website randomly shows visitors different versions and compares outcomes.

Survey experiment

Random versions of a question are shown within a survey (wording splits, vignettes).

Vignette

A short scenario shown to respondents, with details varied across versions.

Deception

Hiding a study's true purpose from participants.

Debriefing

Explaining the study to participants afterward and letting them withdraw their data.

Population

The entire group the researcher wants to understand.

Study population

The part of the population actually available for selection.

Sampling frame

The list from which the sample is drawn.

Element

One unit in the population (one undergraduate).

Sample

The smaller group actually studied.