Bias

Definition

Bias is the systematic difference between a statistic's expected value and the true population parameter it estimates. An estimator is unbiased if, on average across repeated samples, it equals the parameter. Bias can arise from model misspecification, measurement error, sample selection, or omitted variables.

Why It Matters

Understanding bias is essential because even a precise estimator can be misleading if it consistently misses the true value. In econometrics, endogeneity is one of the most common sources of bias, producing coefficient estimates that do not reflect genuine causal relationships. Researchers must evaluate potential sources of bias at every stage of analysis, from study design through model specification to interpretation of results.

Example

Suppose a survey on household income only collects responses during weekday working hours. Working individuals are less likely to respond, so the sample over-represents retirees and under-represents employed people. The resulting income estimate is biased downward because the sampling procedure systematically excludes a high-earning subgroup.

Related Terms

• Endogeneity
• Standard Error
• Confidence Interval
• Ordinary Least Squares (OLS)

Software Notes

• SPSS: No single command; assess bias through residual analysis (Analyze > Regression > Linear > Plots) and compare models
• R: bias() from the metrics package, or compute manually as mean(estimate) - true_value
• Stata: estat bias after estimation; use ovtest for omitted-variable bias detection