Methods for Fish Biology

Chapter 1: Research Methods: Concept and Design

William E. Waters and Don C. Erman

doi: https://doi.org/10.47886/9780913235584.ch1

The process of research has been variously defined to include chance discovery, pure reasoning, application of the scientific method, and the use of statistical concepts and methods. All of these elements may be included in a research investigation. For purposes of this book, we define research as a systematic and orderly process by which new knowledge or information is obtained in accordance with specified objectives. This definition implies a utilitarian purpose and, indeed, utility is a basic motivation for research in fish biology just as in other fields of natural resource science (utility does not necessarily mean direct or immediate application of the new information). The definition excludes sole reliance on casual observation and chance discovery, simple reasoning, random trial and error, and bibliographic searches. Our emphasis is on process.

The qualifier new may mean (1) a more complete description or explanation of a biological pattern, process, or relationship, (2) an extension of the time or space dimensions of existing information, (3) modification of current explanations or theories, (4) an entirely new explanation that provides a different base of knowledge with new vistas for future research, or (5) a new way of combining information for improved prediction and decision making. A research project may be conceived and designed to achieve one or any combination of these goals. The research process, then, can be depicted in a very general way as.

The intended use of the new information is a basic factor that determines the concept and design of every study or experiment. It also determines the approach to be taken in statistical analysis. Description and explanation draw on methods of statistical inference. Prediction (related to specific actions) and decision making utilize procedures of statistical decision making. For some situations, this dichotomy may not be entirely clear. It may be contended, for example, that a good description of a pattern of occurrence or an explanation for a change in population size is sufficient for a management decision. This frequently occurs because expediency and necessity override risk and uncertainty. However, if one understands at the start what kind of data and form of analysis are required for a predictive model or other decision tool, one should see the need for a different kind of study that extends research beyond the descriptive-explanatory stage. The intended use of the new information is a basic factor that determines the concept and design of every study or experiment. It also determines the approach to be taken in statistical analysis. Description and explanation draw on methods of statistical inference. Prediction (related to specific actions) and decision making utilize procedures of statistical decision making. For some situations, this dichotomy may not be entirely clear. It may be contended, for example, that a good description of a pattern of occurrence or an explanation for a change in population size is sufficient for a management decision. This frequently occurs because expediency and necessity override risk and uncertainty. However, if one understands at the start what kind of data and form of analysis are required for a predictive model or other decision tool, one should see the need for a different kind of study that extends research beyond the descriptive-explanatory stage. Barnett (1973) gave a historical perspective of the distinction between these dualBarnett (1973) gave a historical perspective of the distinction between these dual functions of statistics and some basic considerations on the requisites and techniques of each.