Rose and Smith, 1992

Rose, K. A., and E. P. Smith. 1992. Experimental design: The neglected aspect of environmental monitoring. Environmental Management 16:691-700.

Environmental monitoring is increasing due to regulatory mandates and public concerns over the health of the environment. Much of the data appear to be being collected without explicit statement of hypothesis and with little regard for principles of experimental design. Mismatches between the monitoring design and the hypothesis of interest arise when inappropriate designs are used, hypotheses of interest arise when inappropriate designs are used, hypotheses are poorly stated or change over time, and when data collected for one purpose are used for other purposes (i.e. data are used to evaluate hypotheses which the data were never intended to address). Two long-term time series on historical changes in dissolved oxygen (DO) in Chesapeake Bay are analyzed to illustrate how a mismatch between monitoring design and the hypotheses of interest can lead to analyses of low power and even contradictory conclusions. Neither time series was collected to evaluate long-term trends in DO; regression analysis results in one-time series showing a downward trend in DO, whereas the other time series shows no temporal trend. The importance of experiemental design considerations will increases in the future as monitoring used to not only identify and define environmental problems but also to quantify the effectiveness of management and remedial actions taken in response to identified problems. To ensure the usefulness of the large amounts of environmental monitoring data accumulating, we need to ask more questions that relate how data are collected (the experimental design) to why data are collected (the hypotheses of interest).