Fishery biologists spend considerable effort over multiple years collecting data on fish population and community status using a particular sampling method or set of methods. However, new (and often more effective) sampling methods and technologies are continuously being developed. To incorporate these new sampling techniques, fishery biologists need a means for converting sample data collected using old methods so they can be compared with data collected using new methods. Similarly, fishery biologists often need a means to compare fish sample data collected using the same method over time (e.g., from year to year) and space (e.g., between sample sites). If fish abundance, species presence, or richness are estimated using an unbiased statistical estimator, the estimates can be validly compared, even if the fish sample data were collected with different methods. However, if unbiased statistical estimators were not used, biologists need methods for adjusting fish sampling data collected using different methods or using the same method collected under different sampling conditions. In this chapter, we describe and provide examples of statistical techniques for converting nonstandard fish sampling data to standardized data and for making comparisons of fish sampling data collected at different times or at different locations. We define standard fish sampling data as data collected using the standardized fish sampling methods described throughout this book. Any other sampling methods and associated data are thus defined as nonstandard. Before delving into the details of the statistical modeling techniques, we describe the nature of fish sample data, their uses, and their limitations.

Catch-effort measures, such as relative abundance and catch per unit effort (CPUE), are more formally described as indices. Here, we define an index as any measure or count of a species or community (e.g., species richness) based on direct observation without an estimate of the ability to count individuals or species. Indices have some very desirable characteristics for use in fisheries research and management. In general (but not always), indices require less effort to collect and are usually more precise than unbiased population estimators (e.g., CPUE versus capture–recapture estimates of abundance). The proper use of indices for assessment of fish populations or communities, however, requires that the relationship between an index and the true value (e.g., fish density, species richness) is relatively constant (1) across the observable range of true values, (2) through time when evaluating trends at a single location, and (3) across space when making comparisons among locations.