Wisdom of Crowds: who should be interviewed?
Diversity is the design principle
The most important sampling principle in BECCA is diversity. A good respondent pool should include people with different levels of experience, different ages, different gears, different fishing areas, different roles, and different relationships to the fishery.
Older and more experienced fishers are essential because they extend the dataset backwards. They may remember conditions before formal monitoring began, before major habitat change, before market expansion, before new regulations, or before a fishery became commercialised or recreationally important. Their knowledge helps recover historical baselines that would otherwise disappear.
Younger and newer fishers are equally essential because they anchor the present. They provide information on current catches, current effort, current fishing areas, and recent change. They also prevent the dataset from becoming overly weighted toward the past. If a BECCA only interviews older experts, it may produce a rich historical story but a weak picture of current fishing. If it only interviews current active fishers, it may describe the present well but miss the scale of historical change. A strong BECCA needs both.
This is the central Wisdom of Crowds logic. The aim is not to identify one perfect expert. It is to combine many partial views into a stronger collective estimate. In the bonefish optimisation work, small but diverse respondent groups performed as well as, and in some cases better than, homogeneous groups of highly experienced fishers. This supports a practical rule: prioritise diversity before prestige 14.
Why not only interview experts?
Expert fishers matter, but “expert-only” sampling can introduce bias. Highly experienced fishers may cluster in older age groups, use particular gears, fish particular places, or represent a specific social group. They may remember the past extremely well, but they may not represent current fishing patterns. Conversely, newer fishers may not know what the fishery was like decades ago, but they may be the best source of information on current catch rates, present-day effort, and recent changes in behaviour.
A BECCA should therefore seek a respondent pool that spans as many years as possible. The ideal dataset includes people who began fishing in different decades. Someone who started fishing in 1975 helps reconstruct the 1970s and 1980s. Someone who started in 1995 helps reconstruct the 1990s and 2000s. Someone who started in 2020 helps anchor the current baseline. Together, these respondents create a time series that no single fisher can provide alone.
Practical sampling guidance
There is no universal sample size that works for every fishery. A small village fishery may only have 20 active harvesters. A regional recreational fishery may have hundreds or thousands of participants. The practical goal is to interview enough people to capture the diversity of the fishery and enough people within each important area to avoid one or two voices dominating the pattern.
For a very small fishery, the best strategy may be to interview as many active and former fishers as possible. For a small community fishery, 20 to 40 respondents may provide a useful first assessment. For a medium-sized fishery, 50 to 75 respondents is a stronger target. For a large regional fishery, 75 or more respondents is preferable, especially if the fishery is divided across multiple regions. Where spatial comparisons are important, a useful working target is at least 20 to 25 respondents per important region, while recognising that this may not always be possible.
Recruitment can combine open community invitations, fisher association networks, landing site interviews, guide networks, women’s harvesting groups, market contacts, local partner recommendations, snowball referrals, and public survey links. The final report should always explain how respondents were recruited and which groups may be underrepresented.
As a practical guide:
Very small fishery: interview as many active and former fishers as possible.
Small community fishery: aim for 20 to 40 respondents if possible.
Medium fishery: aim for 50 to 75 respondents.
Large regional fishery: aim for at least 75 respondents, and more where possible.
Spatially divided fishery: aim for at least 20 to 25 respondents per important area, where feasible.