For most habitat analyses, researchers typically collect and examine environmental data from the landscape scale (a few square kilometers to hundreds of square kilometers) all the way down to the scale of a microhabitat (tens of square meters). At the larger spatial extents, the data may be GIS-based such as spatially referenced land cover data. At smaller spatial scales, the data may be collected (variables measured) in the field at the study sites. Data for a habitat analysis are often based on randomly located and spatially delineated sampling or survey plots. The environmental data compose a set of a few to tens of predictor variables that are used in statistical tests for a relationship with the response variable that is typically species presence–absence, abundance (counts of individuals), or activity level. Depending on the spatial scale of analysis, predictor variables could represent different environmental variables such as vegetation structure, soil properties, and other characteristics of the substrate. Climate and weather variables are environmental, but they are not considered to be characteristics of the habitat. The formal habitat analysis consists of testing for a statistical relationship between the response variable and one or more environmental predictor variables so as to identify those variables that truly are habitat characteristics. A study of the habitat of the brown-throated sloth in Costa Rica is used to further explain the type of data used in characterizing the habitat of a species.