PLSC

PLSC 3603 Final Exam Study Guide (Chapters 9–14)

Chapter 9: Quantitative Research Designs

Quantitative research designs use systematic empirical investigation of observable phenomena via statistical, mathematical, or computational techniques. They emphasize identifying causal relationships.

1. Randomized Experimental Designs

  • Posttest-Only Control Group Design: Participants are randomly assigned to either a treatment or control group. Only post-treatment data is collected, assuming randomization equates the groups. Used to establish causality with high internal validity.

  • Pretest-Posttest Control Group Design: Both groups are measured before and after treatment. Allows researchers to compare change over time and across groups.

  • Repeated-Measurement (Time-Series) Design: Multiple pre- and post-treatment measurements are taken. Enhances understanding of patterns and trends.

  • Multiple-Group Design: Includes more than one treatment group, enabling comparison across different interventions or dosage levels.

2. Field Experiments

  • Conducted in real-world environments, maintaining random assignment.

  • High external validity due to natural setting.

  • Example: Niven’s study on the effect of negative campaign mailers on voter turnout. Found that negatively toned messages lowered turnout.

3. Natural Experiments

  • Treatment assignment is beyond the researcher's control but resembles randomization.

  • Example: Susan Hyde's study on international election observers in Armenia. She found election monitors increased voter confidence and turnout.

  • Strength: High external validity.

  • Weakness: Lack of control over intervention specifics.

4. Quasi-Experimental Designs

  • No random assignment; groups are pre-existing.

  • Requires careful statistical controls for confounders.

  • Common in public policy evaluation where controlled experimentation is not feasible.

5. Observational Designs

  • Cross-Sectional Design: Observes variables at a single point in time.

  • Longitudinal (Time-Series) Design: Observes variables over time to detect trends or changes.

    • Intervention Analysis: Measures the impact of a specific event or policy.

    • Trend Analysis: Tracks long-term developments in variables like voter participation or public opinion.

Chapter 10: Survey Design and Content Analysis

Survey Design

  • Question Types:

    • Open-ended: Allow respondent to reply in their own words. Useful for exploratory research.

    • Closed-ended: Provide fixed responses. Easier to code and analyze.

  • Good vs. Bad Questions:

    • Double-barreled: Ask two things at once (e.g., "Do you think taxes and crime are too high?")

    • Ambiguous: Vague or unclear (e.g., "Do you support fairness?")

    • Leading: Suggests a correct answer (e.g., "Don’t you agree that…")

  • Wording Principles:

    • Use familiar language.

    • Avoid jargon.

    • Keep questions short and specific.

    • Avoid negative phrasing.

  • Ordering:

    • Start with easier, non-threatening questions.

    • Avoid priming or framing later responses with early questions.

Content Analysis

  • Used to systematically code textual, visual, or audio material.

  • Units of Analysis:

    • Recording Units: Words, sentences, paragraphs.

    • Context Units: Thematic elements or document sections.

  • Coding:

    • Manual or software-based (e.g., NVivo).

    • Requires clearly defined categories.

    • Inter-coder reliability is essential for validity.

  • Advantages:

    • Non-invasive.

    • Allows for longitudinal comparison.

  • Limitations:

    • Subjectivity in interpreting content.

    • Difficulty capturing nuance.

Chapter 11: Descriptive Statistics

Distributions and Frequency Tables

  • Organize data to show how often values occur.

Measures of Central Tendency:

  • Mean: Average; sensitive to outliers.

  • Median: Middle value; robust to outliers.

  • Mode: Most common value; useful for categorical data.

Measures of Dispersion:

  • Range: Highest value - lowest value.

  • Variance: Average squared deviation from the mean.

  • Standard Deviation (SD): Square root of variance; shows average distance from the mean.

Skewness:

  • Describes asymmetry of distribution.

    • Positive skew: Tail on the right.

    • Negative skew: Tail on the left.

Normal Distribution:

  • Bell-shaped, symmetrical curve.

  • Empirical Rule: 68% of data within 1 SD, 95% within 2 SDs, 99.7% within 3 SDs.

Central Limit Theorem:

  • Sampling distribution of the mean approaches normality as sample size increases.

Charts:

  • Bar Graphs: Categorical comparisons.

  • Histograms: Distribution of interval/ratio data.

  • Pie Charts: Proportions of a whole.

  • Line Graphs: Trends over time.

Chapter 12: Hypothesis Testing and Inference

Key Concepts:

  • Z-Score: Standardized score showing how many SDs a value is from the mean.

  • T-Test: Compares the means of two groups to see if they differ significantly.

  • Degrees of Freedom (df): Number of values that are free to vary.

Confidence Intervals (CI):

  • Range where a parameter is likely to fall.

  • CI = estimate ± (critical value * standard error).

Hypothesis Testing:

  • Null Hypothesis (H0): No relationship or difference.

  • Alternative Hypothesis (H1): There is a relationship/difference.

Errors:

  • Type I (False Positive): Rejecting a true H0.

  • Type II (False Negative): Failing to reject a false H0.

Significance Level (α):

  • Common threshold: 0.05 (5% chance of Type I error).

P-value:

  • Probability of observing the data if H0 is true.

  • If p < α, reject H0.

Critical Values:

  • Cutoffs that determine statistical significance.

Chapter 13: Measures of Association and Correlation

Relationships:

  • Direction: Positive (as X increases, Y increases), negative (as X increases, Y decreases).

  • Strength: Determined by correlation coefficients.

Pearson's r:

  • Measures strength and direction of linear relationships.

  • Range: -1 to +1

    • +1: Perfect positive correlation

    • -1: Perfect negative correlation

    • 0: No correlation

Chi-Square Test (χ2):

  • Tests relationship between two categorical variables.

  • Compares observed vs. expected frequencies.

Measures of Association:

  • Lambda: For nominal variables.

  • Gamma: For ordinal variables.

  • Cramér’s V: Adjusted chi-square for strength of association.

Interaction Effects:

  • Occur when the effect of one variable on an outcome depends on another variable.

Variance:

  • Amount of spread in data.

Chapter 14: Regression Analysis

Ordinary Least Squares (OLS) Regression:

  • Predicts value of DV based on IV(s).

  • Assumes linearity and normally distributed residuals.

Regression Line (Line of Best Fit):

  • Minimizes squared differences between observed and predicted values.

Key Concepts:

  • Regression Coefficient (b): Estimated change in DV for 1-unit change in IV.

  • R-Squared (R2): Proportion of variance in DV explained by IV.

  • Residual: Difference between actual and predicted values.

  • Residual Sum of Squares (RSS): Total of squared residuals.

OLS Assumptions:

  1. Linearity

  2. No multicollinearity

  3. No autocorrelation

  4. Homoscedasticity (equal variance of residuals)

  5. Normality of errors

  6. Correct model specification

  7. Independent observations

  8. Valid measurement of variables

  9. No influential outliers

  10. Continuous DV

Nonlinear Models:

  • Use Maximum Likelihood Estimation (MLE).

  • Handle binary or ordinal dependent variables (e.g., logistic regression).

  • Heteroscedasticity: Unequal residual variance, violating OLS assumption.

Test Tips:

  • Focus on identifying research designs and their strengths/weaknesses.

  • Know how to interpret descriptive and inferential statistics.

  • Be able to explain hypothesis testing logic.

  • Understand when and how to use regression analysis.