Exam 3 Research Methods

Study Guide

Correlation (Chapter 8)

Core Definition

  • Correlation = predictive relationship between two variables

  • Variables must be interval or ratio level

Correlation Coefficient (r)

  • Indicates direction + strength

  • Range: -1.00 to +1.00

Direction

  • Positive (+): variables move in same direction

  • Negative (−): variables move in opposite directions

Strength

  • 0.10 = very weak

  • 0.20 = small

  • 0.35 = moderate

  • 0.50+ = strong

Visualizing Data

  • Scatterplot = shows relationship between two variables

  • Pattern = direction + strength visually

Standardization

  • Z-scores = standardize values to compare across variables

  • ~95% of data falls within typical z-score range (normal distribution concept)

Degrees of Freedom

  • df = N − 1

  • Adjusts for sample estimation error

Statistical Significance

  • Significant if unlikely due to chance

  • p < .05 = statistically significant

  • Smaller sample → harder to reach significance (more sensitive to outliers)

Reliability

  • Test-retest reliability: stability over time

  • Internal consistency: how well items measure same construct

    • measured using Cronbach’s alpha

Chapter 9: Experimental Design & Causation

Causation Requirements

  1. Covariation (variables change together)

  2. Temporal precedence (cause happens first)

  3. No alternative explanations (control extraneous variables)

Core Experimental Variables

Independent Variable (IV)

  • manipulated variable

  • has conditions (groups)

Dependent Variable (DV)

  • measured outcome

  • not manipulated

Groups

  • Experimental group = receives treatment

  • Control group = baseline comparison

Manipulation check

  • confirms IV worked as intended

Design Types

Between-Subjects Design

  • different participants in each condition

  • uses random assignment

  • or matched groups

Within-Subjects Design

  • same participants in all conditions

  • participants serve as own control

  • higher statistical power

  • lower error variance

Order Effects (Within-Subjects Problem)

  • Carryover effects

  • Practice effects

  • Fatigue effects

Control methods

  • Counterbalancing

  • Block randomization

  • Balancing conditions

Confounds & Control

  • Confound = variable that varies with IV

  • Threatens internal validity

Control groups

  • Empty control = no treatment

  • Placebo = fake treatment

Statistical Tests Overview

t-test

  • compares 2 groups

  • independent = between-subjects

  • paired = within-subjects

ANOVA

  • used when 3+ groups

  • avoids multiple t-tests problem

Formula idea:

  • F = between-group variance / within-group variance

Degrees of Freedom (ANOVA)

  • df between = k − 1

  • df within = N − k

  • total = N − 1

Errors in Hypothesis Testing

  • Type I error: reject true null (false positive), α = .05

  • Type II error: fail to reject false null (false negative)

Post Hoc Tests

  • used only after significant ANOVA

  • identifies which groups differ

  • controls Type I error inflation

Factorial Designs

Definition

  • study with 2+ IVs (factors)

  • notation: 2 × 2 design

Effects

Main effect

  • effect of one IV ignoring others

Interaction effect

  • effect of one IV depends on another IV

With 2 IVs, ANOVA tests:

  • Main effect IV1

  • Main effect IV2

  • Interaction (IV1 × IV2)

Key Rule

  • Interaction can exist without main effects

  • Always check interaction first if present

Design Advantages

  • tests multiple IVs at once

  • identifies interactions

  • more efficient than separate studies

If you want, I can also:

  • turn this into a 1-page cheat sheet

  • make practice exam questions

  • or quiz you like your professor would

Just tell me 👍