Should you trust unanimous decisions? - Derek Abbott

Prevodilac: Milenka Okuka Lektor: Mile Živković

Imagine a police lineup where ten witnesses

Zamislite paradu prepoznavanja gde je od deset svedoka

are asked to identify a bank robber they glimpsed fleeing the crime scene.

zatraženo da identifikuju pljačkaša banke

koga su letimično spazili kako beži s mesta zločina.

If six of them pick out the same person,

Ako šestoro njih izabere istu osobu,

there's a good chance that's the real culprit,

dobre su šanse da se radi o stvarnom krivcu,

a ako svih desetoro naprave isti izbor,

and if all ten make the same choice,

you might think the case is rock solid,

pomislili biste da se radi o čvrstom slučaju,

but you'd be wrong.

ali ne biste bili u pravu.

For most of us, this sounds pretty strange.

Za većinu nas ovo zvuči prilično čudno.

After all, much of our society relies on majority vote and consensus,

Konačno, naše se društvo uglavnom oslanja na glas većine i konsenzus,

whether it's politics,

bilo u politici,

business,

poslu

or entertainment.

ili zabavi,

So it's natural to think that more consensus is a good thing.

Pa je prirodno misliti da je više konsenzusa nešto dobro.

And up until a certain point, it usually is.

I do određene tačke, obično jeste.

But sometimes, the closer you start to get to total agreement,

Ali ponekad, što se više približavate potpunom slaganju,

the less reliable the result becomes.

rezultat je sve nepouzdaniji.

This is called the paradox of unanimity.

Ovo se naziva paradoksom jednoglasnosti.

The key to understanding this apparent paradox

Ključ za razumevanje ovog očiglednog paradoksa

is in considering the overall level of uncertainty

je u razmatranju sveukupnog stepena nesigurnosti

involved in the type of situation you're dealing with.

uključenog u situaciju kojom se bavite.

If we asked witnesses to identify the apple in this lineup, for example,

Ako bismo tražili od svedoka da prepoznaju jabuku u ovom prepoznavanju, na primer,

we shouldn't be surprised by a unanimous verdict.

ne bi nas iznenadila jednoglasna presuda.

But in cases where we have reason to expect some natural variance,

Ali u slučajevima gde postoji razlog da očekujemo neke prirodne varijacije,

we should also expect varied distribution.

takođe bi trebalo da očekujemo varijaciju u distribuciji.

If you toss a coin one hundred times,

Ako bacite novčić stotinu puta,

you would expect to get heads somewhere around 50% of the time.

očekivali biste da dobijete glavu negde oko 50% puta.

But if your results started to approach 100% heads,

Ali ako bi vaš rezultat bio sve viši rezultatu 100% glava,

you'd suspect that something was wrong,

posumnjali biste da nešto nije u redu,

not with your individual flips,

ne s vašim pojedinačnim bacanjima,

but with the coin itself.

već sa samim novčićem.

Of course, suspect identifications aren't as random as coin tosses,

Naravno da identifikacija osumnjičenih nije nasumična kao bacanje novčića,

but they're not as clear cut as telling apples from bananas, either.

ali nije ni tako precizna kao razlikovanje jabuka od banana.

In fact, a 1994 study found that up to 48% of witnesses

Zapravo, istraživanje iz 1994. je otkrilo da do 48% svedoka

tend to pick the wrong person out of a lineup,

bira pogrešnu osobu u paradi prepoznavanja,

even when many are confident in their choice.

čak i kad je većina njih sigurna u svoj izbor.

Memory based on short glimpses can be unreliable,

Sećanje zasnovano na letimičnom pogledu može da bude nepouzdano,

and we often overestimate our own accuracy.

a mi često precenjujemo sopstvenu preciznost.

Knowing all this,

Znajući sve ovo,

a unanimous identification starts to seem less like certain guilt,

jednoglasna identifikacija sve manje zvuči kao izvesna krivica

and more like a systemic error,

a sve više kao sistemska greška

or bias in the lineup.

ili predrasuda u paradi prepoznavanja.

And systemic errors don't just appear in matters of human judgement.

A sistemske greške se ne pojavljuju samo kod ljudskog prosuđivanja.

From 1993-2008,

Od 1993. do 2008,

the same female DNA was found in multiple crime scenes around Europe,

DNK iste žene je pronađen na više mesta zločina širom Evrope,

incriminating an elusive killer dubbed the Phantom of Heilbronn.

okrivljujući neuhvatljivog ubicu označenog kao Fantom iz Heilborna.

But the DNA evidence was so consistent precisely because it was wrong.

Ali DNK dokaz je bio toliko dosledan baš zato što je bio pogrešan.

It turned out that the cotton swabs used to collect the DNA samples

Ispostavilo se da je pamučne tampone, korišćene za sakupljanje DNK uzoraka,

had all been accidentally contaminated by a woman working in the swab factory.

sve slučajno kontaminirala žena koja je radila u fabrici tampona.

In other cases, systematic errors arise through deliberate fraud,

U drugim slučajevima, sistemska greška nastaje namernom prevarom,

like the presidential referendum held by Saddam Hussein in 2002,

poput predsedničkog referenduma koji je održao Sadam Husein 2002,

which claimed a turnout of 100% of voters with all 100% supposedly voting in favor

koda je tvrđeno da je odziv glasača bio 100% i svih 100% su navodno glasali

of another seven-year term.

za novi sedmogodišnji mandat Huseina.

When you look at it this way,

Posmatrano na ovaj način,

the paradox of unanimity isn't actually all that paradoxical.

paradoks jednoglasnosti nije uopšte toliko paradoksalan.

Unanimous agreement is still theoretically ideal,

Jednoglasni sporazum je i dalje teoretski idealan,

especially in cases when you'd expect very low odds of variability and uncertainty,

naročito u slučajevima gde ne očekujete značajnu promenljivost i neizvesnost,

but in practice,

ali u praksi,

achieving it in situations where perfect agreement is highly unlikely

njegovo dostizanje u situacijama gde je savršen sporazum skoro nemoguć

should tell us that there's probably some hidden factor affecting the system.

govori nam da verovatno imamo skriveni faktor koji utiče na sistem.

Although we may strive for harmony and consensus,

Iako težimo harmoniji i konsenzusu,

in many situations, error and disagreement should be naturally expected.

u mnogim situacijama bi trebalo prirodno da očekujemo greške i neslaganje.

And if a perfect result seems too good to be true,

Ako savršen rezultat izgleda suviše dobro da bi bio istina,

it probably is.

verovatno i jeste.