Binary Data

Environmental data often occur in discrete form, such as the number of individuals living near a waste site, or whether or not a subject develops cancer after exposure to an environmental agent. This latter example is one of the simplest forms of discrete data: a dichotomous indicator of whether or not an event has occurred, i.e. a ‘success’ or a ‘failure’. Such an outcome can be coded simply as either 1 or 0 – corresponding to success or failure – and we refer to this as a case of binary data. If the data on each subject or unit in the study are coded in this binary fashion, and if each subject contributes one and only one such binary observation to the experimental data, we call this a case of ungrouped binary data. Alternatively, when the subjects have some underlying or coordinating feature that relates them, it is common to group the like subjects and record the proportion of subjects responding; i.e. take the proportion Y/N, where Y is the number of observed successes out of the N observed subjects in the common group. This is grouped binary data.