Political Analysis Midterm Study Guide

Vocabulary flashcards covering fundamental concepts, descriptive statistics, causal modeling, research design, experimental methods, and controlled comparisons from the Political Analysis Midterm Study Guide.

Multidimensional Concept

A concept that contains two or more distinct dimensions of concrete traits.

Conceptual Dimension

A set of concrete traits of a similar type that define an aspect of a concept.

Operational Definition

A specification of how a concept should be empirically measured in research.

Test-Retest Reliability

A reliability assessment method that applies the same measure to the same subjects at two different times.

Unit of Analysis

The specific entity that researchers describe and analyze in a study.

Pooled Dataset

A longitudinal dataset collected over time that includes different units of analysis at each measurement point.

Systematic Error

Measurement error that causes consistent, chronic distortion, such as when participants alter their behavior because they know they are being observed.

Random Error

Nonsystematic measurement error caused by chance factors, such as an interviewer recording an answer incorrectly or an assistant making data entry mistakes.

Panel Study

A longitudinal study design in which the exact same respondents are surveyed repeatedly at set intervals over time.

Codebook

A reference document that details information about a dataset, including data collection methods, reliability, variable descriptions, and value labels.

Hawthorne Effect

A phenomenon wherein study participants alter their behavior simply because they are aware they are being observed.

Ecological Fallacy

The logical error of deducing individual-level attitudes, characteristics, or behaviors from aggregate-level or state-level data.

Construct Validity

The degree to which an operational measure correlates with other variables in ways predicted by theory, such as evaluating an ideology scale by comparing it to party vote choice.

Face Validity

An assessment of whether an operational measure intuitively appears on its surface to capture the intended concept, such as evaluating whether welfare spending support reflects economic liberalism.

Aggregate-Level Unit of Analysis

A unit of analysis composed of collections of individuals or entities, such as using each state in an FBI crime report.

Individual Unit of Analysis

A unit of analysis based on individual entities, such as each single congressional bill or each individual presidential speech.

Percentile

A measure indicating the percentage of cases in a distribution that fall below a given value.

Kurtosis

A statistical measure that describes how a distribution's peakedness and tails differ from a standard normal bell curve.

Dispersion

The amount of spread or variation among the values of a variable.

Median

The middle value of an ordered dataset; the measure of central tendency most resistant to skew.

Mode

The most frequently occurring value in a dataset or distribution.

Nominal Variable

A categorical variable whose values represent distinct groups with no inherent numeric ranking, such as gender or state region.

Ordinal Variable

A categorical variable whose values can be logically ranked in an ascending or descending order, such as education levels ranging from high school diploma to advanced degree.

Interval-Level Variable

A variable whose values represent precise, identical numeric differences between units, such as age measured in years or temperature.

Cumulative Percentage

The running total percentage of cases that fall at or below a given value in a distribution.

Frequency Distribution

A table summarizing the number or proportion of observations that fall into each category of a variable.

Additive Index

A composite measure constructed by summing multiple ordinal variables, frequently utilized to create an interval-level scale.

Metadata

Information describing data characteristics, such as variable definitions, units of measurement, and value labels.

Value Label

A text description assigned to clarify the substantive, non-numeric meaning of numeric codes in a dataset.

Likert Scale

A measurement scale recording levels of agreement or disagreement, commonly debated because researchers frequently treat it as interval despite its ordinal nature.

Dummy Variable

A variable restricted to only two values, typically coded to represent the presence or absence of a trait (e.g., yes/no).

Network Node

An individual unit within a network that forms connections with other units.

Network Centrality

A metric reflecting the structural importance or prominence of a node within a network.

Automated Content Analysis

A computational technique that converts non-numeric textual information into numeric data, such as word cloud analysis.

Logarithmic Transformation

A mathematical adjustment used to normalize a strongly positively skewed distribution, which can be reversed by exponentiation.

Linear Transformation

A mathematical operation that adds, subtracts, multiplies, or divides all values by a constant, maintaining both relative distances and rank order.

Null Hypothesis

A statement positing that no relationship or effect exists between the variables being examined.

Hypothesis

A testable proposition stating an empirical cause-and-effect relationship between an independent variable and a dependent variable.

Cross-Tabulation

A table showing the joint distribution of two categorical variables, displaying the dependent variable across categories of the independent variable.

Intervening Variable

A mediator variable that serves as the causal mechanism transmitting the effect of an independent variable to a dependent variable.

Fork (Causal Model)

A causal configuration where two variables share a common underlying cause but do not directly influence one another.

Chain (Causal Model)

A causal sequence in which an independent variable influences an intermediate factor, which subsequently impacts the dependent variable.

Fishbone Diagram

A causal visualization tool, also called an Ishikawa diagram, where the diagonal lines or 'bones' represent different categories of causes.

Tautology

A circular statement that is true by necessity of its own phrasing rather than by empirical evidence, such as saying 'good guitar players play guitar well.'

Inverse Relationship

A negative relationship where the dependent variable moves in the opposite direction of the independent variable.

Curvilinear Relationship

A relationship whose direction or rate of change varies across different intervals of the independent variable, such as rising, leveling off, and then falling.

Chartjunk

Unnecessary visual elements and decorative embellishments that clutter a graph and diminish its data-to-ink ratio.

Jittering

The technique of introducing small amounts of random noise to reposition overlapping data points on a scatterplot to reduce overplotting.

Field Experiment

An experimental study carried out within subjects' natural, real-world environment rather than a laboratory setting.

Laboratory Experiment

An experiment conducted in a controlled, artificial setting (such as a classroom theater) that affords high control over confounding influences.

Internal Validity

The extent to which a research design successfully isolates an independent variable's causal effect and rules out rival alternative explanations.

External Validity

The degree to which the findings of an empirical study can be generalized beyond the specific study sample to broader populations and real-world contexts.

Random Selection

A sampling process ensuring that every individual in a target population has an equal likelihood of being chosen for the sample.

Random Assignment

A procedure ensuring that research participants are equally likely to be assigned to the treatment or control group, which reduces selection bias.

Sampling Frame

The operational population list or source from which researchers draw their sample participants.

Stratified Random Sample

A sampling method where the target population is first divided into distinct categories or strata before cases are randomly drawn from each subgroup.

Cluster Sampling

A probability sampling technique applied when a population is distributed across defined geographic units or clusters.

Preregistration

The formal filing of study hypotheses and analysis protocols before data collection begins, aimed at eliminating post hoc theorizing.

Benford's Law

A mathematical law describing the frequency distribution of leading digits, often employed to identify fabricated or fraudulent data.

Feeling Thermometer

An operational measurement tool capturing a respondent's warmth or emotional favorability toward a target on a scale ranging from $0$ to $100$.

Controlled Effect

The observed relationship between an independent variable and a dependent variable evaluated within a single specific value of a control variable.

Partial Effect

The estimated effect of an independent variable on a dependent variable that persists after accounting for a rival control variable.

Zero-Order Relationship

The raw, overall bivariate relationship observed between two variables prior to controlling for rival variables or third factors.

Spurious Relationship

A coincidental or illusionary association between an independent variable and a dependent variable that disappears once a shared causal rival variable is controlled.

Additive Relationship

A relationship pattern where both the independent variable and the control variable exert independent effects on the dependent variable, appearing as parallel lines on a graph.

Interactive Relationship

A relationship where the impact of an independent variable on the dependent variable changes in magnitude or direction across different levels of a control variable, often appearing as crossing lines on a graph.

Difference-in-Differences Design

A quasi-experimental evaluation design that compares changes over time between treatment and control groups to control for the presence of unobserved variables.

Matching

An observational technique that pairs units with similar characteristics to replicate the balance of random assignment without true experimental manipulation.