Chapter 8 | Bivariate Correlational Research

What is a bivariate correlation?

A bivariate correlation is an association between exactly two variables. Both variables are measured, not manipulated.

What does it mean to interrogate association claims?

It means to evaluate how well the research supports the claim, focusing on construct validity, statistical validity, internal validity, and external validity.

What is the effect size, and what are Cohen’s criteria?

Effect size measures the strength of a relationship.

• Cohen’s criteria:

  • 2% = small

  • 5% = medium

  • 8% = large effect

What are outliers, and how can they affect correlations?

Outliers are data points that are far from others.

Online outlier: falls along the general trend

Offline outlier: goes against the trend
Outliers can inflate or weaken correlations. Detect using scatterplots; consider removing if they’re errors or unduly influence results.

What is restriction of range, and what is its impact?

Restriction of range occurs when scores on one or both variables are limited, making the correlation appear weaker than it truly is.

What is a curvilinear association? Name two types

A curvilinear association is when the relationship between two variables is not a straight line (linear).

U-shaped: increases, then decreases (or vice versa)

Inverted U-shaped: increases up to a point, then decreases

What are the three criteria for causation?

Covariance: variables are related

Temporal precedence: cause comes before effect

Internal validity: no confounding variables

What is a moderator variable in correlational research?

A moderator variable changes the strength or direction of the relationship between two other variables (e.g., relationship is stronger for one group than another).