Bias in Estimating Recreationists' Length of Stay from Sample Interviews

Publisher Website
Google Scholar
Abstract
Estimating recreational use of relatively unsupervised areas objectively depends upon sampling. Most land managers count visitors, and multiply by an estimate of length of stay to obtain man-days. Estimates of length of stay from on-site samples are biased upwards, however, becausethe probability of a visitor falling in the sample is a function of length of stay. Estimates of the attitudes or characteristics of recreationists based on interviews are also biased if the characteristics are associated with length of stay, as seems likely. The sample apparently cannot bedrawn to produce an unbiased estimate directly, but a simple weighting procedure gives the desired estimate. The procedure requires multiplying the number and total days stayed of the samples with a stay of X days by 1/X, and totaling the products. The sample must be drawn with replacement.If the weighted total of days stayed is multiplied by the inverse of the sampling rate, an estimate of total man-days is obtained. Dividing this estimate by the average length of stay gives an estimate of visits. This procedure is believed to be more efficient than starting with visits.

This publication has 0 references indexed in Scilit: