The Analysis Data Model (ADaM) is an emerging CDISC standard that is specific to analysis data sets. ADaM data sets, in contrast to SDTM data, contain both raw and derived observations and variables and they are accompanied by clear metadata that describes both the derivation of observations and variables as well as how the data is used to produce the final statistical analysis. This presentation will describe the ADaM model and associated metadata and highlight the important concepts that a statistician needs to be aware when creating successful ADaM data sets.