Ace the ClinPsy RMCQ Challenge 2026 – Elevate Your Clinical Psychology Game!

A large sample size tends to improve the reliability of regression estimates because it reduces sampling error and makes the estimated slope and intercept more precise. So, having a big dataset is not a reason for regression to be unreliable; it actually helps.

Outliers can skew the results, pulling the regression line toward them and inflating residuals, which can distort the estimated relationship, especially when the sample is small. If the data contain non-representative extreme values, the model may mislead about what the typical relationship looks like.

Non-linear data pose another issue because simple linear regression assumes a straight-line relationship. If the true pattern is curved, the linear model fits poorly, producing biased predictions and misleading inferences.

A small range of values, or range restriction, reduces variability in the predictor and/or outcome. This makes it harder to detect a true relationship, inflates uncertainty around the estimates, and can lead to a conclusion of no effect even when one exists.