Bias

Definition

A systematic, non-random deviation of observations or estimates from the true value; in biological research, any factor that distorts sample representativeness, measurement accuracy, or inference validity. Unlike random error, bias creates consistent skew in a predictable direction and cannot be reduced by increasing sample size alone. In entomological and ecological contexts, bias arises from sampling methods (e.g., light traps attracting phototactic species while missing diurnal taxa), observer effects (e.g., collectors favoring larger, more conspicuous beetles), spatial or temporal autocorrelation, or analytical choices (e.g., pseudoreplication). Recognizing and minimizing bias is central to experimental design, biodiversity assessment, and population parameter estimation.

Etymology

Example

A Malaise trap samples bias toward flying Hymenoptera and underestimates ground-dwelling Carabidae; combining trap types reduces this taxonomic bias in arthropod community studies.

Synonyms

• systematic error
• distortion

Related Terms

• Accuracy
• Precision
• sampling error
• pseudoreplication
• detection probability
• selection bias
• observer bias
• trap bias

Usage Notes

Distinguished from imprecision (random scatter around the true value). Bias can be positive or negative depending on direction of skew. In mark-recapture studies, trap-happy or trap-shy individuals create behavioral bias in population estimates. Reporting bias (e.g., publishing only significant results) affects meta-analyses in entomology. Some biases are estimable and correctable (e.g., detection probability in occupancy modeling); others require design changes to prevent.