Fishery management takes place in a larger world. Thus, it’s hardly surprising that ideas and
opinions that arise outside the fisheries context are relevant to management
issues.
I was reminded of that the other day, when I was reading an
excerpt from The Harvard Gazette, in which a professor reminded readers
about the difference between anecdote and data, and noted that
“Too many leaders and influencers, including politicians, journalists,
intellectuals, and academics, surrender to the cognitive bias of assessing the
world through anecdotes and images rather than data and facts.”
Anyone who has ever attended a fisheries meeting understands
why that sentence rang a bell, because there is probably no other venue where
so many people try to pass off so many anecdotes—mere stories—as facts.
Such anecdotes take many forms. Some are stories told by people who oppose
further restrictions on harvest, and are trying to grasp onto any tale that
might stave off additional regulations.
Some are observations told by fishermen who don’t understand fisheries
science, and are merely reporting what they think that they see. And others reflect honest and accurate
observations of local abundance and local catchability, which might not reflect
the overall condition of the stock.
Gulf of Maine cod provide an example of one such situation.
A few years ago, it became apparent to most that the cod
were in serious trouble. A new,
peer-reviewed stock assessment revealed the stock was badly overfished, and
that abundance had fallen to well below 10% of the target level. Yet, as reported in
the blog TalkingFish.org, fishermen argued that
“Cod are everywhere.”
The Commonwealth of Massachusetts agreed to conduct its own
statistically-valid survey, to see whether it could develop data that would
support the fishermen’s claims. Just the
opposite happened. The survey results effectively
confirmed the findings in the stock assessment:
the cod stock was in bad shape.
Yet the fishermen rejected the survey’s data. According to TalkingFish, one outspoken
industry representative told the Boston Globe that
“The state survey literally does zero to improve our confidence. You can’t just sample anywhere. You have to go where the cod are supposed to
be.”
Such statement says a lot about too many fishermen’s chronic misunderstanding of survey and assessment techniques. As fishermen, their job is catching fish, and
if they are in the business for any length of time, they get pretty good at it. They learn where fish will be under various
conditions, and know how to chase down and find even small concentrations of
fish, located in very specific areas, in order to bring home a good catch.
They ignore all of the places where they don’t expect fish
to be, because it’s pointless to fish there.
Thus, when they drop down their hooks, deploy their gill
nets or tow their trawls, they usually end up catching a fair amount of fish. They tend to discount the fact that the
places that the fish still inhabit are surrounded by vast expanses of
once-productive, and now largely empty, bottom.
They know other fishermen who are catching fish, too. All of their success stories, considered together, provide a very
selective, and very distorted, overall view of fish abundance, composed of
individual anecdotes, and not of data.
But to a fishermen who wants to believe that all is well with the stock,
they serve as “proof” that the science is wrong.
Although that example is drawn from the commercial fishery,
we frequently see similar examples emerge from the recreational fleet. They are very common when dealing with migratory species
such as striped bass and bluefish, which can be very abundant at certain times
and places as they move along the coast.
This spring, for example, many fishermen who targeted the
big, pre-spawn female striped bass that were staging in Raritan Bay and lower
New York Harbor before moving into the Hudson River to breed rejected a
recent stock assessment’s conclusion that the striped bass was both overfished
and subject to overfishing.
After all, how could they be
filling coolers with bass (often legally, sometimes not) if there weren’t many bass to be found? Suggestions that striped bass need further
protection are often met with scorn and outright hostility, as anglers focused
on local, transient abundance fail to see the bigger picture that was captured
by the stock assessment. Few ever
stopped to question whether, in a time of relative scarcity, it made any sense
to kill large, fecund females on their way to the spawning grounds.
Once again, too many fishermen took reports out of context,
and allowed tales of abundance in one particular place, at one particular time,
to trump a detailed, data-driven and peer-reviewed assessment of the health of
the striped bass stock.
Similarly, stories of striped bass abundance in federal
waters more than three miles from shore, where fishing for bass is illegal, are
being used by some folks who are trying to impeach the findings of the latest
stock assessment.
The stories all take about the same format. Someone will get up at a meeting and declare something like “My pals Frankie and Joey were out in the canyon fishing
for tuna, and on their way back to the market they ran across acres—acres,
they said—of striped bass maybe 20 miles from shore. They told me that there were so many bass
that when they ran through the school, their depthfinder lit up and showed the
bottom 20 feet under the boat, when they were still in 150 feet of water.”
That story gets repeated by a few people a few times in a
few different forms, and maybe Frankie and Joey themselves are at the meeting
to tell it themselves, and everybody looks at one another with satisfied
expressions on their faces, nodding and fully convinced that they just “proved”
that the science is wrong and that there are plenty of striped bass offshore—something
all real fishermen know is the truth—but that the scientists who
count fish, again, just don’t know where to find them.
Such stories can even convince legislators who are weak on
the details of fisheries science.
“used flawed data that measures the Atlantic Striped Bass
stock based on the entire eastern seaboard, yet failed to account for the
Atlantic Striped Bass outside of the 3-mile fishing area, assuming that fish
abide by arbitrary bureaucratic boundaries.
Alternative data that shows the Striped Bass stock is in a
better place outside of the 3-mile was…thrown out by the [Atlantic States
Marine Fisheries] Commission… [emphasis added]”
Thus, we see exactly the problem that the piece in The
Harvard Gazette alluded to, which arises when a politician falls victim to
anecdotes, and confuses such anecdotes for data and for facts. For the stories told about striped bass being
encountered offshore does not rise to the level of data—not even to the level
of “alternative data” which, like “alternative facts,” tend
to distort what is true.
That doesn’t mean that the fishermen’s observations can’t
lead to real, statistically-valid data. But
to do that, a scientifically rigorous investigation must take place. Saying “Frankie and Joey know what they saw,”
just isn’t good enough.
If someone wanted to prove that there are striped bass
offshore that were never included in any of the surveys considered in the stock
assessment, they wouldn’t do that merely by providing an accounting of times
that such fish were seen more than three miles from shore—even if such list was
fairly detailed, and contained dates, defensible estimates of the number of
fish seen each time, etc. In the end, all that such list would contain are
stories. A lot of stories, to be sure, but still nothing more than
anecdotal reports.
To truly prove the assertion that there are uncounted bass offshore,
supporters of such claims would first have to create a “null hypothesis” that,
in this case, would assume that there are no such striped bass in federal waters, and then collect valid data to prove that the null hypothesis is
wrong.
They might, for example, get an exempted fishery permit that
allows them to catch and tag bass in federal waters, then see how many of those
tagged fish show up in the inshore surveys used to prepare the striped bass
stock assessment. If it turns out that
none of the tagged fish show up in inshore surveys after a reasonable period of
time has passed, then there would be at least some data that tends to contradict
the null hypothesis, and so supports their claim.
It turns out that Massachusetts
recently did something very much like that, placing acoustic tags in 125
striped bass, then waiting to see whether such fish were detected in inshore
waters (although by acoustic gates, and not in the surveys used for stock assessment
purposes). As it turns out, fully 95% of
the tagged fish returned to Massachusetts state waters, and fully 77% of them
returned seasonally for up to 2 years (as the tagging program ended in 2011
and the results of the study were published in 2014, it is possible that bass
continued to return seasonally for longer than that, but were not reported due
to the study’s conclusion).
That real data, which shows the vast majority
of “offshore” fish tagged soon returned inshore, is not completely
dispositive of the question, but it tends to strongly support the null hypothesis, and
thus does not support any “alternative data”—that is, fishermen’s stories—suggesting that there
are uncounted bass swimming somewhere in federal waters.
And that’s how fisheries management—and any science-based issue—is
supposed to work: Driven by facts, not
by anecdotes and compelling images. As The
Harvard Gazette’s piece concluded
“We need to make ‘factfulness’ an inherent part of the
culture of education, journalism, commentary, and politics. An awareness of the infirmity of unaided
human intuition should be part of the conventional wisdom of every educated
person. Guiding policy or activism by
conspicuous events, without reference to data, should come to be seen as risible
as guiding them by omens, dreams, or whether Jupiter is rising in Sagittarius.”
Because fish stories can be entertaining, but they are not data.
And they shouldn’t drive fisheries policy.
No comments:
Post a Comment