Estimating Population Size with Correlated Sampling Unit Estimates

Quattro Pro Spreadsheet to perform the example calculations described in the manuscript can be copied from here.

This spreadsheet consists of 13 worksheets.  The first 6 worksheets generate the variance-covariance matrix of estimated population size per quadrat (worksheet named “VC NTotalHat”) from the program MARK output.  For a new problem, the user could enter the variance-covariance matrix of the estimated population sizes directly.

The variance-covariance of the population estimates by quadrat and these estimates are used to compute 3 different estimates of the total population size and their standard errors for 3 different auxiliary variables: area (worksheet “Ratio Area”), TIN Ratio – 1 (worksheet “Ratio TINRatio”) and the TIN Ratio – 1 raised to the 1.52 power (worksheet “Ratio TINRatio 1_52”).  For a new problem, the user could use one of these worksheets as a template to compute the population total by replacing the existing auxiliary variable values with their own.

The last worksheet demonstrates an example of computing power to detect a trend based on the Ganey (1999) problem used in the manuscript and this spreadsheet.