An evaluation of the punch card method of estimating salmon-steelhead sport catch

by Ronald Herbert Hicks

Institution: Oregon State University
Department: Statistics
Degree: MS
Year: 1964
Keywords: Sampling (Statistics)
Record ID: 1583078
Full text PDF: http://hdl.handle.net/1957/48604


A primary objective of the management practices of the Oregon State Game Commission is to maximize and sustain the annual harvest of salmon and steelhead by sportsmen fishing in Oregon. Assessment of these practices is dependent upon an accurate determination of total annual sports harvest throughout the state. Present total harvest figures are derived under a report card system which requires each salmon-steelhead angler fishing in Oregon to purchase and subsequently return to the Game Commission a punch card giving dates, rivers, and numbers of salmon and steelhead caught during the year. Unfortunately, only about 30 percent of the salmon-steelhead anglers return their cards. Lacking information on the harvest characteristics of anglers not returning punch cards, estimates of total annual catch, total catch by river, month, and type of angler have been made by simple expansion of punch card reported catch. Inherent in this method of estimation are the assumptions that the average catch per angler by month, by river, by type of angler, and for the year are the same for anglers returning and for anglers not returning punch cards. This paper reports on an evaluation of these assumptions and on the method for estimating total catch. Evaluation of the assumptions is based on comparisons of catch rate estimates from sample data (not punch cards) collected from both anglers returning and anglers not returning punch cards. Sample data was collected by means of double return postcard questionnaires in a sample survey during 1951. Stratified sampling was employed with allocations approximately proportional to the angler-fishing months in each stratum. Variances for the catch rate estimates made for both groups of anglers were estimated using replicated subsampling. Nearly all comparisons indicate that the catch rates for anglers not returning punch cards were significantly lower than the catch rates for anglers returning punch cards. On the basis of the observed differences in catch rates, a procedure using ratios of the 1961 catch rate estimates is suggested for future estimates of total catch.