Using pattern recognition to avoid bad people

[Ken Fabian]

Inevitably we do invoke prejudices - the bearded, tattooed bike rider who is part of an organised crime gang will make us wary of the bearded, tattooed bike riders that raise money for charities and never hurt anyone. We can be okay, even pleased with harsh police actions - or vigilante actions - against bearded, tattooed bike riders in general, on the basis that all share responsibility and it will be a lesson to the ones who are criminal.

We can be okay with that - even getting vicarious satisfaction from knowing police are targeting bearded, tattooed bike riders - but if we are bearded tattooed bike riders who never hurt anyone it is prejudicial and unjust.

I think it is one of humankinds most serious flaws that we can feel satisfaction and even enjoy violence so long as we think the victim deserves it. Thinking they deserve it doesn't require investigation or weighing evidence; just being told that someone is bad can be good enough. Or having similar dress, appearance, ethnicity, religious or political affiliation to someone deemed "bad" can be good enough. When retribution for criminal acts can be applied to people who just look like someone who committed criminal acts then retribution becomes a crime all of it's own.

[MigL]

But if 10% of the general population are part of an organized crime gang ( my numbers may be skewed because a lot of my acquaintances are Italian  ), yet 50% bearded, tattooed bike riders are, why should this valuable piece of statistical information be ignored ?
Of course the 'targeting' must be followed up by rigorous investigation and the weighing of evidence.
Of course being told, or simply assuming, that someone is bad isn't good enough.

But we are supposed to be scientific thinkers; why would we ignore information because it doesn't suit our sensibilities
( see the above quip regarding Italians' crime connections )

[dimreepr]

8 minutes ago, MigL said:

But we are supposed to be scientific thinkers; why would we ignore information because it doesn't suit our sensibilities

It's akin to a physicist understanding every aspect of a car, other than how to build/fix one; but we're afraid to find out...

[J.C.MacSwell]

9 minutes ago, MigL said:

But if 10% of the general population are part of an organized crime gang ( my numbers may be skewed because a lot of my acquaintances are Italian  ), yet 50% bearded, tattooed bike riders are, why should this valuable piece of statistical information be ignored ?
Of course the 'targeting' must be followed up by rigorous investigation and the weighing of evidence.
Of course being told, or simply assuming, that someone is bad isn't good enough.

But we are supposed to be scientific thinkers; why would we ignore information because it doesn't suit our sensibilities
( see the above quip regarding Italians' crime connections )

Generally speaking, we simply don't have time to investigate all our biases in anything close to a scientific manner, never mind a rigorous one.

It's at best some degree of quality control sampling, then (hopefully) left to our good judgement gained from our experiences.

Obviously guarding against confirmation bias is a good idea.

Try to be accepting of people as much as possible...even if they're wearing a Toronto jersey....😄

[CharonY]

On 11/16/2020 at 9:10 AM, MigL said:

But if 10% of the general population are part of an organized crime gang ( my numbers may be skewed because a lot of my acquaintances are Italian  ), yet 50% bearded, tattooed bike riders are, why should this valuable piece of statistical information be ignored ?

Because numbers are generally not that high and folks misunderstand what it means. What are more common are low numbers, say 4% of the total are criminal and within a subgroup 8% are criminal. Many would then assume that folks falling into the subgroup are more criminal and hence should be avoided. Yet of course 92% are not criminal, meaning that that applying that particular logic would fail most of the time in predicting criminality. There will only be fairly few markers that would have any predictive value (say, a specific gang membership tattoo, though I am not really sure whether that is true, either). As most folks do not really go ahead and calculate positive predictive values for these kind of things and moreover, we have a psychological bias for positive correlations, folks are more likely to misinterpret things with at time rather ugly outcomes.

In biomarker for disease research we face a similar challenge where many biological markers are more likely associated with a given condition, but if we really try to apply them indiscriminately, our ability to detect a given disease is often abysmal. 

[MigL]

54 minutes ago, CharonY said:

many biological markers are more likely associated with a given condition

And you, as a scientist, don't ignore them, as they can certainly point you in a more favorable direction, and get proper diagnostics. but you certainly would not base the diagnostics solely on the markers. They are simply a tool.

[CharonY]

1 hour ago, MigL said:

And you, as a scientist, don't ignore them, as they can certainly point you in a more favorable direction, and get proper diagnostics. but you certainly would not base the diagnostics solely on the markers. They are simply a tool.

Indeed, but that is also why we do not just broadcasting it into the population without qualifying what the statistic mean. 

[Ken Fabian]

4 hours ago, MigL said:

And you, as a scientist, don't ignore them, as they can certainly point you in a more favorable direction, and get proper diagnostics. but you certainly would not base the diagnostics solely on the markers. They are simply a tool.

 

I think this fundamental human flaw - that we recognise danger and target our enemies by outward markers like dress codes, skin colour, religion or politics - makes it more important to base significant actions on actual evidence. I don't dispute that there can be advantage to danger/crime avoidance at personal level and possibly small advantages to crime prevention/detection at societal level in profiling by appearances but I think there are significant downsides, including the reinforcing of prejudices and existing hatreds. If our society is not homogenous - most developed democracies with rule of law are not - and seeks to be inclusive, with ideals of fair treatment under law - most developed  democracies do - it become important to be aware of that propensity to judge (wrongly) by outward appearance.

Being routinely treated with suspicion on the basis of surface appearances undermines a sense of belonging. Those who feel unfairly targeted and excluded are more likely to get angry - whilst being subject to that same flaw as well; it may engender distrust of police and authorities in general rather than of specific wrongful acts or perpetrators. It becomes more likely they will turn to those experiencing the same prejudice for solidarity and support. More likely to feel that lashing out indiscriminately is justified, that any unwitting victims of that "deserve it" - because of the group the victims are perceived to belong to. It can also lead to acceptance of criminal behavior - where it targets victims that are perceived to "deserve it"; if people are treated like a thief whether doing theft or not then I think some people will be more inclined to steal without guilt when opportunity arises. That is what I think this short circuit in human behavior can lead to - a tolerance for and even satisfaction in harsh treatment of those deemed to deserve it by virtue of the community they are perceived to belong to.

[MigL]

Unfortunately, Ken, that is an evolutionary trait we've developed.
So maybe it should be fortunately, because we're good at recognizing 'patterns' and 'markers' that indicate danger.

We instinctively swat at flying, buzzing insects because we recognize the danger of being stung, even though bees, flies and mosquitos are nowhere near as aggressive as wasps. We fear noises at night because we instinctively 'remember' that is when the carnivorous hunters come out to find prey. Many of our fears are instinctive, and we become 'fearful' when we recognize certain signs of danger ( even if it's not there ).

Similarly, it's ingrained in our consciousness that, if you are being followed by a black/hispanic gang member, at night, you are most likely afraid of being assaulted. This obviously is not evolutionary, but has become apparent during the last couple of hundred years in our society, where black people ( and other immigrants ) have been portrayed a evil people who can't control their urges/violence. This obviously is wrong.

But if you're alone at night and being followed by someone, wrong or not, I suggest being afraid ( heightened senses ) and getting to a safe place.