Representation

Representation Bias arises when data does not accurately reflect the target population, affecting the generalizability of findings. Causes vary across data types, with

• sampling errors, low response rates, or the under-coverage of certain demographic groups affecting representation in survey data,
• platform selection biases and the need to filter and (pre)process large volumes of data affecting digital behavioral data, and
• non-consent to the sharing of linkage information, and errors in the linking process affecting representation in linked datasets.