Amphibian Count Database


Patuxent Wildlife Research Center	Managers' Monitoring Manual

Managers' Monitoring Manual > CV Tools > Amphibian Count CV Tool

Coefficients of Variation Tool for Amphibian Count Database

This tool is designed to help land managers and others set sample size for population monitoring programs. Understanding how variable a particular taxonomic group is (along with other variables such as geographic area, sex/age class, etc.) can help determine the number of sampling points needed to monitor that taxonomic group. Use the search criteria below to find coefficients of variation at various taxonomic scales.

(Once you have selected a taxonomic level at left, more specific search criteria will appear below)

See the Amphibian Count CV Summary Tables.

See the list of references these counts came from.

See how CV's are calculated.

Sample size is an important determinant of whether a monitoring program can detect population trends (should they exist) over time. Variance in counts of animals over space and time is the major determinant of how many samples are needed for a program to be effective. Obtaining a good estimate of variance for a particular species at a given place under a given sampling regime requires several years of pilot study data. Absent that, data from studies of similar species under similar circumstances might provide some guidelines. That's the idea behind this database: to organize what we already know from the literature.

Our goal for this database has been straightforward: compile counts of amphibians over time from published studies and calculate descriptive statistics such as mean, standard deviation, and coefficient of variation (CV).* To date, a total of 47 studies has yielded 212 separate count series for 54 amphibian species.

ACD Summary Tables: Compare and contrast CV's across taxonomic group, survey method, and other variables. (Note also that median and mean CV's for each subset of counts generated by searches on these pages are calculated dynamically and reported along with search output).

As always, we look forward to feedback on all aspects of this site. Please drop us a line if you know of studies we missed, methods relating to count data from which we might benefit, or if you would consider contributing data to our project.

*A coefficient of variation (CV) is calculated by dividing the standard deviation by the mean. To account for the effects of trend over time, which elevates variation in counts, we used the standard deviation of the residuals from a regression of the counts run against time. See how the CV is calculated using example data. Associated with this method are inherent assumptions which you may find disagreeable due to sticky statistical issues like model misspecification, autocorrelation, and effect of survey length on variance. For this reason, we provide the raw counts so you can analyze the data using your preferred method. Note that many of these counts were eyeballed off published graphs, and will vary from the true counts to some degree.