Sampling Technique Comparison for Estimating Fishermen Population Parameters in Bayelsa State

Selecting an appropriate sampling strategy for estimating population parameters among artisanal fishermen in Bayelsa State, South South Nigeria, is a methodological challenge with direct implications for the reliability of fisheries management statistics in the Niger Delta region. This study compares the performance of three sampling techniques for estimating parameters of the artisanal fishermen population in Bayelsa State: simple random sampling, stratified random sampling by fishing community type, and cluster sampling by creek network. A simulation study was designed using a known population of 1,800 registered fishermen from the Bayelsa State Fisheries Department register. Samples of size n=150 were drawn under each design and repeated 500 times. Estimates of mean monthly catch, income variance, and proportion accessing credit were computed under each design and compared on bias, variance, and mean squared error. Stratified random sampling provided the lowest mean squared error for mean monthly catch estimation (MSE = 38.4) compared to simple random sampling (MSE = 67.2) and cluster sampling (MSE = 112.6). Cluster sampling showed the highest variance in all estimates, reflecting within-cluster homogeneity. Stratified sampling required the lowest fieldwork cost per unit of precision achieved. The study recommends stratified random sampling as the statistically optimal design for Bayelsa State fisheries population surveys and provides stratum weights for three community types for use in future survey design. Keywords: sampling techniques, stratified sampling, fishermen population, Bayelsa State, survey design