You are here:
Cumulative distribution functions and their use in monitoring programs
Citation:
KINCAID, T. AND A. R. OLSEN. Cumulative distribution functions and their use in monitoring programs. Chapter 14, Gitzen, Robert A., Joshua J. Millspaugh, Andrew b. Cooper, Daniel S. Licht (ed.), Design and Analysis of Long-Term Ecological Monitoring Studies. Cambridge University Press, Cambridge, Uk, , 313-324, (2012).
Impact/Purpose:
Ecological resource monitoring programs typically have estimating the status and change in status as an objective.
Description:
Ecological resource monitoring programs typically have estimating the status and change in status as an objective. A well designed and skillfully implemented survey design will produce an accurate representation of the status of the resource at the time the survey was conducted. Typically, status is determined using multiple indicators that are based on measurements taken at sites selected. Although estimating the average, or mean, of the indicator scores may be of interest, a more informative summary of status is the distribution of the indictor scores. The simplest distribution summary that can be use for categorical and continuous data is the estimated cumulative distribution function (CDF). For complex survey designs that employ stratification or unequal probability sampling, survey design weights must be incorporated into the CDF estimate. In this chapter, the process for including design weights in the CDF estimate is discussed and illustrated using the spsurvey library for R Statistical Software. In addition, the chapter discusses (1) the potential bias that can arise when indicator values include measurement error and (2) the comparison and testing of two CDFs either between two subpopulations or two time periods.