Estimating non-compliance among recreational fishers: Insights into factors affecting the usefulness of the randomized response and item count techniques. (September 2015)

• Record Type: Journal Article
• Authors: Thomas, Alyssa S.; Gavin, Michael C.; Milfont, Taciano L.
• Is Part Of: Biological conservation. Volume 189(2015)
• Page Start: 24
• Page End: 32
• Publication Date: 2015-09
• Subjects: Conservation management -- Illegal fishing -- Indirect questioning -- Methodological study -- Sensitive behaviours -- Violation rates; Conservation of natural resources -- Periodicals; Nature conservation -- Periodicals; Ecology -- Periodicals; Environment -- Periodicals; Environmental Pollution -- Periodicals; Electronic journals; 333.9516
• DOI: 10.1016/j.biocon.2014.09.048
• Languages: English
• ISSNs: 0006-3207

Abstract

Highlights: Non-compliance with the two main blue cod fishing regulations is alarmingly high. Item Count Technique failed to provide the highest estimate for any behaviour. Randomized Response Technique provided the highest estimate for only one behaviour. Minimum levels of behaviour sensitivity/frequency are needed for indirect methods. Using one sample population increases confidence results are due to the method used. Abstract: Non-compliance with fishing regulations has a critical influence on the success of the associated management regime. Yet, estimating the extent of non-compliance is challenging in part because of the sensitive nature of the subject and direct questioning is likely to result in low estimates. This study tested the effectiveness of two indirect methods, randomized response technique (RRT) and item count technique (ICT), in providing higher and more accurate estimates of recreational fishing non-compliance than traditional direct questioning (DQ) in the Marlborough Sounds blue cod fishery, New Zealand. Although RRT provided a higher non-compliance estimate than ICT and DQ for one of the three regulations (size limit), ICT failed to provide a higher estimate than DQ for any of the three regulations. We suggest that the online mode of the survey, behaviour frequency and question sensitivity (although not measured) had a strong influence on our findings. The version of RRT used, offering increased privacy protection for respondents, is also likely to have contributed to its performance against ICT. This study is the first to use the same sample population for all methods, eliminating potentially confounding socio-demographic factors and providing more confidence in attributing differences to the method used.