Political Science 050

Operationalization

Making an abstract concept measurable — defining exactly how a variable will be observed and recorded.

Hypothesis Creation

Forming a clear, testable statement predicting the relationship between an independent and dependent variable.

Null Hypothesis

The statement that there is no relationship between the variables, where any observed association is due to chance.

Independent Variables

The presumed cause, the factor you examine to explain change in the outcome.

Dependent Variables

The outcome being explained, which depends on the independent variable(s).

Control Variables

A third factor held constant so it can't distort the relationship between the IV and DV.

Dummy Variables

Binary variables coded as 0 and 1 to bring categorical data into analysis.

Deductive Reasoning

Reasoning from a general theory down to a specific, testable hypothesis (theory to data).

Inductive Reasoning

Reasoning from specific observations up toward a general theory (data to theory).

Spurious Relationships

Coincidence — an apparent relationship actually caused by a third factor, like ice cream sales "causing" murders, which shows why theory matters since it provides the reasoning behind your claim.

Types of Experimental Designs

True/classic, pre-experimental, and quasi-experimental designs, differing in whether they use random assignment and a control group.

Hawthorne Effect

People change their behavior simply because they know they're being observed.

Ethics

The reason political scientists mostly don't run experiments on people — you can't cause harm, and subjects must be informed for consent.

3 Levels of Measurement

Nominal (unordered categories), Ordinal (ranked), and Interval (equal numeric intervals).

Scales

Tools that combine multiple items to measure a single concept, such as a Likert scale.

Measure Manipulation

Smoothing or adjusting data so measures "talk to each other," or accounting for missing data (e.g., a missing year in a time study) — allowed as long as the researcher can reasonably justify it.

Survey Data

Data collected by asking a sample standardized questions, where exam questions may ask which forms of data are or aren't legitimate.

Survey Structure/Errors

Built on randomness, meaning everyone in the population has an equal chance of selection, with key errors including Sample Error, Wording Error, the Halo Effect/Agreement Bias, Sequence Error, and Specification Error.

Different Types of Data

Researcher observation, aggregate data (usually third-sourced), content analysis, and survey data.

Farming

Using one source's reference list to find new references.

Descriptive Statistics

Self-evident math — addition, subtraction, division, averages — that requires no interpretation, since 2+2 always equals 4.

Inferential Statistics

Requires interpretation and isn't self-evident, so an r value must be read in the context of the data to make predictions, generalizations, and conclusions, with its basis being probability and standard deviation.

Mean, Mode, Median, Sum

Mean is the average, mode is the most frequent value, median is the middle value, and sum is the total.

Probability

The likelihood that an event occurs, forming the foundation of inferential statistics.

Standard Deviation

A measure of how spread out values are around the mean.

Normal Curve

The symmetric, bell-shaped distribution at the center of statistical inference.

Standard Error

The estimated standard deviation of a sample statistic, which shrinks as sample size grows.

Statistical Significance

The probability that a result isn't just due to chance, commonly set at p < .05.

Type I/Type II Errors

False positive (rejecting a true null) versus false negative (failing to reject a false null), and can be applied to practically any data.

Interpreting Charts and Tables

Reading graphical and tabular data and drawing the correct conclusions from it.

Regression Analysis

A technique estimating how independent variables predict a dependent variable.

Multiple Regression

Regression using two or more independent variables to predict one dependent variable.

Intercept

The predicted value of the dependent variable when all independent variables equal zero.

Correlation Coefficient (r)

Measures the strength and direction of a linear relationship, ranging from –1 to +1.

Coefficient of Determination (r²)

The proportion of variance in the dependent variable explained by the model.

Positive/Negative Correlations

Positive means variables move together, while negative means one rises as the other falls.

Curvilinear

Nonlinear relationships, where the line bends rather than running straight.

Types of Curves

The different shapes a relationship can take, such as linear, curvilinear, and exponential.

Multicollinearity

When independent variables are highly correlated with each other, distorting the regression estimates.

Autocorrelation

When your dependent variable is correlated with itself over time.

Heteroskedasticity

When your errors are inconsistent — as the mean increases, the range of the errors also increases, whereas under homoskedasticity the error range would stay constant as the mean changes.

Game Theory (Nash Equilibrium)

A decision-making model that relies on path analysis rather than OLS, flawed because it assumes "perfect information" — you need to know the outcome before using it as an analytical tool, which you often don't. A Nash equilibrium is where no player gains by changing strategy alone.

Sucker Principle

Essentially "FOMO" or an aversion to fraud — the discomfort of someone else getting something you're not.

Lit Review

The steps are Research Question, Keyword Search, Farming, Synthesizing the Literature, Describing the Phenomena, and Revising the Research Question