Suppose a predictor shows a significant effect only when another variable is held constant. This phenomenon is called:

Prepare for the Discovering Statistics Using IBM SPSS Statistics Test with detailed questions and thorough explanations. Enhance your statistical understanding and apply SPSS effectively. Get ready to excel in your assessment!

Multiple Choice

Suppose a predictor shows a significant effect only when another variable is held constant. This phenomenon is called:

Explanation:
A suppressor effect is at work when a predictor looks non-significant on its own but becomes significant once another variable is included and held constant in the model. This happens because the suppressor variable accounts for variance in the predictor that isn’t related to the outcome, purifying the predictor’s true relationship with the outcome. As a result, the unique association emerges more clearly, and the coefficient can even change sign as irrelevant variance is removed. This is different from confounding, where controlling for a third variable removes bias from a spurious association; and from mediation, where the effect operates through another variable. It’s also distinct from multicollinearity, which causes instability and inflated standard errors due to high predictor intercorrelations rather than producing a new significance pattern by itself.

A suppressor effect is at work when a predictor looks non-significant on its own but becomes significant once another variable is included and held constant in the model. This happens because the suppressor variable accounts for variance in the predictor that isn’t related to the outcome, purifying the predictor’s true relationship with the outcome. As a result, the unique association emerges more clearly, and the coefficient can even change sign as irrelevant variance is removed.

This is different from confounding, where controlling for a third variable removes bias from a spurious association; and from mediation, where the effect operates through another variable. It’s also distinct from multicollinearity, which causes instability and inflated standard errors due to high predictor intercorrelations rather than producing a new significance pattern by itself.

Subscribe

Get the latest from Examzify

You can unsubscribe at any time. Read our privacy policy