In keeping with the tradition of providing the world with education and enlightenment, always presented as a family oriented format and with strict accordance with proper boundaries of good taste and propriety, CooperToons is proud to present

A Most Merry and Illustrated History of

(with a discussion of offender profiling, predicting the future, the Iranian Revolution, JFK, and the Boston Strangler thrown in)

 

Our Favorite Pastime

We all know the favorite pastime throughout the world. No, no, no, we don't mean that, for heaven's sake. We mean being entertained with stories of murder, mayhem, butchery, plunder, spttin', belchin', cussin', and other not very nice things. Often condemned by the sanctimonious defenders of the 'Merkin Way, violent entertainment has a long and honorable history. It even stretches back to early Greek drama.

But in Greek theater the violence was (literally) off stage, and the characters mostly stood around and talked. Of course, the Roman theater had more action, and the Romans liked gladiatorial combat as well. But that's not quite the same thing.

No, our modern voyeurism with violence actually started in 1841 with Edgar Allen Poe's story, the "Murders in the Rue Morgue". This tale has pretty much what we've all come to love and enjoy in family entertainment: murder, mutilation, and lurid crime scene descriptions. We have bodies stuffed up chimneys, severed heads rolling off bodies, and even a villain that ...

Well, we won't give the ending away. It is, if you can weather Poe's style, a pretty good story.

But Edgar's story also introduced other staples for the modern mystery writer. Most of all we learned that the typical police force is made up of a bunch of blundering dunderheads that couldn't solve a crime if the criminal's name was projected from the lens of their bulls-eye lanterns. No, the people who really solve the crimes are news reporters (who are sometimes strange visitors from another planet and dressed in leotards), crime scene photographers, neighborhood teenagers, younger kids, or even dogs, for crying out loud!

But what Edgar really gave us was the "crime consultant". That was C. Auguste Dupin, who steps in and who by his flashy reasoning and deductions shows the Parisian police really how to solve crimes. Dupin was the direct inspiration for Sir Arthur Conan Doyle when he was writing his Sherlock Holmes stories. Later fictional consultants were Solar Pons (created by American writer August Derleth), the Hardy Boys, Nancy Drew, Ironside, and Scooby-Doo. But they all go back to M. Dupin.

Of course, in today's techno-era the consultants may have some official connection with the police department. But rarely are they actually policemen. Instead they might be laboratory geeks who can further humiliate the official force by using the razzle-dazzle techniques of modern forensic science.

So it is a bit of a surprise when you learn that in 2009, the National Academy of Sciences issued a blistering blast against the field. The report pointed out that even though forensic tests can mean (literally) the difference between life and death, virtually no tests - that's virtually no tests - had ever been properly tested or validated. Oh sure, the tests might generate reams of numbers and spreadsheets full of data. But the actual conclusions relied too much on the opinion and intuition of the "experts" - experts for whom there was no actual training or way to assess their competence. A forensic specialist can sit on a witness stand and make definitive claims which are totally unjustified. In some cases, people have wrongly been sent to prison based on things like blood splatter patterns, ballistic markings, teeth imprints, handwriting analysis, and even fingerprints. It's bad enough if the innocent go to jail. But it also means the police have stopped looking for the real crook, a crook who is out and happily committing more crimes.

Yes, we said the criticism even applied to fingerprints. After all, it's one thing to compare inked prints on cards taken at a police station with each other. But it's something else to take a smudged partial print lifted from a real world crime scene and say it was from the defendant. Although you wouldn't know it from listening to Joe Friday talking about ten-point confirmation, identifying a fingerprint sometimes involves judgement calls of surprising shakiness. More shocking is that fingerprint evidence can sometimes be manipulated by unscrupulous forensic scientists (yes, there are some unscrupulous forensic scientists) to send innocent people to jail. In fact, this topic was dealt with in surprising accuracy on one of the better Law and Order episodes.

 

Now A Profile, Please!

If you think the sacred cow of traditional forensics - fingerprinting - could come under fire, then consider something like one of today's hot topics which requires even more judgement, data sifting, and conclusion drawing. We mean, of course, criminal profiling. Criminal profiling not only gets flack from malcontents like defense attorneys trying to show that their client is innocent (or at least trying to get their client off), but sometimes from law officers themselves.

But just what is criminal, or more properly, offender profiling? In essence, offender profiling is correlating specifics of a crime with the characteristics of individuals who have committed such crimes in the past. By including the proper variables, the reasoning goes, you can narrow the possibilities of who committed the new (and unsolved) crime. Sometimes "psychological factors" are also often included in the profile to narrow down the choices even further.

Advocates say profiling is particularly useful in prioritizing suspects. So it leads to earlier apprehension of the perpetrators. True, its effectiveness depends on the skill and experience of the profiler, and it remains as much art as science.

Prioritizing suspects?

Which means, the critics snort, that no one really knows how to do it or whether it even works. In fact, some critics of offender profiling get downright discourteous. Profiling, they say, is simply based on the same tricks used by party game fortune tellers and is nothing more than gussied-up astrology. Predictions may seem precise and specific, but are really cleverly vague. Profilers make such a large number of predictions, often couched in psychobabbelistic double talk, that, sure, they make the occasional correct call simply by chance. Success stories are simply "anecdotal" and profiling failures, by far the more common outcome, are just ignored. Profiling, we hear, is a pseudoscience, worse than useless, and often sends police down the wrong path. It shouldn't be used even as an investigative tool.

Au contraire, as the profilers and M. Duplin would say. Like any newly created discipline, offender profiling is improving. It is becoming more "empirical" and "evidence based". Predictions are increasingly founded on hard data - the "bottom-up" approach favored in the UK as compared to the "top-down" method developed in the US. Certainly profiling isn't perfect. But as far as demonstrating its usefulness, just turn to the case inevitably cited as the Gold Standard of Profiling Success. This was the famous "Mad Bomber" of New York City.

On November 18, 1940, police in New York City were called to the offices of the electrical behemoth Consolidated Edison. On a box in a window sill was a pipe bomb. And there was a note taped on the bomb: "Con Edison Crooks - This is for you".

But why would a bomber - whom we would call a terrorist today - put a note on the bomb itself? That made no sense, and the police suspected that the crudely made bomb was not intended to go off at all. Instead the planting might have been just a "threatening gesture".

The police naturally wondered if they were dealing with a disgruntled employee. But a check of the records uncovered nothing, and Con Edison management told the police that the records before 1940 had been destroyed. So they couldn't check for past employees with a beef. In any case, the incident was considered so trivial that it didn't even make the newspapers. Times have indeed changed.

A year later another bomb, which also didn't go off, was found on 19th Street. This was a few blocks from a Con Edison office. But this time there was no note, and the timer wasn't set. On the other hand, the bomb was built like the other one, and so the police suspected the two bombs were planted by the same person. Possibly the bomber had gotten cold feet and just thrown the device away. Again, the incident was considered not worthy of a newspaper story.

Soon the police received a letter - with the words cut from magazines and newspapers - where the writer said he was going to make Con Edison pay for their "dastardly deeds". Other notes soon followed but without bombs. Some of the letters were handwritten and signed F. P. (which we later learn stood for "Fair Play"). Then after Pearl Harbor, the Bomber sent a note that he would not - as a patriotic gesture - plant any bombs for the duration. In fact, the bomber didn't plant another bomb for over nine years.

Then on March 29, 1951, a third bomb was found in a sand box in an oyster bar in Grand Central Station. It's not clear from the newspapers if this went off or not. In any case no one was hurt. But clearly the bomber was in earnest, and by 1954 he had planted a total of seventeen bombs. Some of these did explode, and along with them the "gesture" theory.

The bombs kept coming. By the end of 1956, the Bomber had constructed and planted 31 bombs, and 23 had exploded. Most were left in public places although one was sent by mail to the Con Edison offices.

Fortunately no one was ever killed, but a man was injured when one of the bombs exploded in Pennsylvania Station. Then when a bomb exploded in Radio City Music Hall, six people were hurt, one quite seriously. A particular danger was not just that the bombs were small and easy to hide, but they didn't even look like bombs. Once a workman found one in a phone booth and just thought it was a piece of pipe. He gave it to a friend to take home for a plumbing project. The bomb did explode, but again no one was hurt.

Even in those days when they thought they had a nutcase, the police might call in psychiatrists to make a profile. So they turned to one of the leading criminal psychiatrists of the day. That was Dr. James Brussel, then the assistant commissioner of the New York State Department of Mental Hygiene.

Although Dr. Brussel had previously created profiles for the FBI and for military intelligence, he wasn't too sure he could tell the police much that they didn't know already. But he agreed to give it a shot. Then we read on both the Fount of All Knowledge (the Internet) and in those non-electronic devices with white flappy things in the middle (books) that Dr. Brussel wrote a profile that hit the bomber spot on.

 

Correlating the Bomber

The Bomber, said the Good Doctor, would be a male (since most bombers were male). He would have a well-proportioned average build (in agreement with physiques of hospitalized mental patients) and would be neither fat nor skinny. Middle aged, he would also be neat and tidy.

A former or present employee of Con Edison, his work would be exemplary. He would arrive on time, be well-behaved and courteous and friendly. From the formal tone and old-fashioned phrasing of the letters, which sounded as if they had been written or thought out in a foreign language and then translated into English, the bomber would probably be foreign-born. Or at least he would be living in a community of the foreign-born. His ancestry was probably Slavic, since when resorting to terrorism, they preferred to use bombs. Being of Slavic extractions, he would also be a Catholic.

By the standards of the day, his education would be good. That is, he probably had some high school, but may not have graduated. It was unlikely he had attended college.

Dr. Brussel also noted that the Bomber wrote using rounded letter "W's" which resembled ladies' breasts. From this characteristic and the slashing and stuffing of theater seats, Dr. Brussel deduced the Bomber was troubled by something about sex. Possibly he had an Oedipus complex. In other words, he loved his mother and hated his father.

The bomber himself, though, would have little interest in women and would possibly never have had a physical encounter. So he would be unmarried, but living with an older female relative who reminded him of his mother. He would be a loner with no friends.

There were also indications the Bomber was living in Connecticut. For once thing, Connecticut has a high concentration of residents of Slavic descent. Also many of the bomber's letters were posted in West Chester County, New York, midway between Connecticut and New York City. A man with the Bomber's intelligence would not send the letters from his own town. Instead, he would post them some distance off, to and from his forays into New York City.

Finally Dr. Brussel told the detectives that when they found the Bomber, he would be wearing a double breasted suit - with the buttons buttoned.

Sure enough, when the detectives arrested George Metesky, a son of Lithuanian immigrants, they found an unmarried, mild-mannered, and courteous averaged sized Connecticut resident who was living with two older sisters. A former mechanic at Con Edison, he had attended high school but had not graduated. A man with no real friends of either sex, he was 53 years old and had been denied workman's compensation for an accident which caused him to contract tuberculosis and ultimately lose his job.

And when he was hauled to the station, what was he wearing? A double breasted suit with all the buttons buttoned!

Certainly all this can't be coincidence, can it? And that bit about the double breasted suit certainly clinches the matter. That couldn't be just a good guess, could it? So the success of profiling seems pretty clear. For 16 years, the police couldn't catch the Bomber. Then once they had the profile, they nabbed him. So even the most dense must admit that at least in this case, the profiling was 1) amazingly accurate and 2) helped the cops catch the crook.

 

Now for a Pause ...

Weeeeeeeelllllllll, hold on there, Pilgrim. We need to stop and take, if not the pause that refreshes, then a pause that lets us see what we have to do to determine if profiling - as the commercials say - really, really works.

The first place in our pause is to make sure we understand that there is a big difference between correlating known data compared to making a true prediction. Correlation is finding relations that already exist. When you see plots, graphs, or lists showing how some characteristic relates to others, this is correlation. However, if we can then (correctly) deduce something that's not actually in our data sets that is prediction.

Correlation or Prediction?

We can see the difference between correlation and prediction if we turn to the time when the British scientist, Edmund Halley, sat down and looked at some astronomical data. Edmund had noted that in the years 1531, 1607, and 1682 a comet was visible in the sky. In other words, he correlated the past appearances of a comet with a time interval - 75 to 76 years. Next he theorized we were seeing the same comet. Finally Edmund predicted that in 1758 the comet would show up again. Indeed it did, and Halley's comet - as it's now called - has done so with regularity up to (and so far) 1986.

Now just because you make a good correlation does not mean a good prediction will follow. It isn't often appreciated, but some methods which produce high correlations are worthless as predictive tools. So applied mathematicians have come up with all sorts of fancy pants ways to rate predictability as something separate from correlation. Sadly, we must avoid the technical details as they do depart a bit from middle school math. On the other hand we can cite at least some bare minimal guidelines for evaluating true predictions, and these are straightforward and simple guidelines that anyone can understand.

 

When is a Prediction Not a Prediction?

Many attempts at prediction do not involve numbers per se. This is often the case in offender profiling. But correlation can also occur without numbers as well. In fact, one of the simplest examples for illustrating the difference between the correlation and prediction is to look at some famous writings which have a reputation for predicting or prophesying the future and seeing how they are actually used.

Inevitably what we see is correlation not prediction. For instance, in the mid-1970's a young lady sat down and began her daily study of - well, we'll just call it "a book" - which contained the following passage: "Ephraim is enjoying wind and is pursuing an east wind. All the day lying and spoiling, he multiplies, and they make a covenant with Asshur, and oil is carried to Egypt."

This young lady suddenly remembered that with the recently signed Egyptian-Israeli peace treaty, the oilfields of the Sinai, which had been captured by Israel in the 1967 Six-Day War, had just been returned to Egypt. And the passage says "Oil will be carried into Egypt." Amazing, she thought. This book, written thousands of years ago, once more accurately predicted the future. Only the most dense, she believed, could doubt the veracity, wisdom, and prophetic power of the writings and the men who wrote them.

But note that what is really being done is relating - that is correlating - a phrase (and one that is taken completely out of context) with an event that has already happened. True, it was recent history, but even though a past event is recent, it is not in the future. What we have here is correlation not prediction or prophecy. (Note: In this case, there is also a mistranslation. The Hebrew word שֶׁמֶן means vegetable oil or animal fat, not the petroleum derived product, one word being נֵפט.)

In this case, to test the true predictive power of this process, the reader would have to find a passage and then predict 1) a specific event that hasn't happened, 2) something that will not happen anyway (for instance, saying you predict that there will be wars or earthquakes isn't really predicting anything), and 3) with a higher frequency of success than you get by just making a lot of guesses. Now from time to time - and fairly recently - people have tried making such true prophecies and without, to say the least, much success and usually with considerable embarrassment to the prophet who says he will try again later.

 

The Basic Predictables

Of course, when dealing with a criminal or psychological profile we are not dealing with past and future events à la classical prophecy. Instead, we make the correlations using previously investigated crimes with the known criminals. Then from the characteristics of a new crime scene that was not used in the original correlation, we then try to deduce (that is, predict) the characteristics of a completely unknown perpetrator. But when testing how effective or profiling is, we have the same issues.

First all predictions need to be clearly specified and must be properties that can be unambiguously verified. In other words, we can say something like "The perpetrator will be a male between 25 and 30 years old". But we have to avoid saying things like, "The criminal is an individual with a highly developed alternative egoistic projection which he tempers with an idiosyncratic sense of self-pity". Although this type of prediction may wow the more naive students in Psychology 101, it really isn't something we can test or will help a cop catch a crook.

For our second criterion we must avoid liberal use of words like "probably", "possibly", "maybe", and "perhaps". Although these words make a profile appear fair, honest, and undogmatic, they also let us wiggle out of profiles that are wrong.

How so? Well, let's say a profiler says a perpetrator "may have dark curly hair, possibly with some streaks of grey likely at the temples but perhaps on the top and back". But consider what happens if the offender had wavy gray hair. Then you could say, well, I didn't exactly say he had dark curly hair. Also wavy and curly are pretty much the same. So if he has gray hair now, then at one time he must have had dark curly hair possibly with some streaks of gray, likely at the temples but perhaps on the top and back". What is actually an incorrect prediction is now claimed to be right!

Finally and most importantly, only predictions that are clearly documented before the profiler knew the answer can be used.

This last point can't be stressed enough. There is a tendency for predictions to get more accurate when they are "remembered" after the fact. At the same time (and lest we draw the wrath of predicters everywhere), we emphasize this is not necessarily deliberate deception. What is called "memory deterioration" and "hindsight bias" are normal psychological phenomena that happen to everyone.

Some simple examples can amplify the danger of after-the-fact predictions. But since we are already discussing criminal profiling, we'll look at some other predictions where you have the same pitfalls.

One thing you must always be alert for are predictions which are actually incorrect (or nonexistent) suddenly becoming amazingly accurate once the event has happened. For instance, many people were shocked! shocked! when the longtime ally of the United States, Mohammad Reza Pahlavi, the Shah of Iran, fell from power in 1979. But a famous newspaper columnist boasted that, by golly, he had indeed predicted the Iranian Revolution, the fall of the Shah, and with it the rise of the Ayatollah. Impressive? No?

Well, actually no. The claim was made well after the Shah was out and the Ayatollah was in. Worse (for the columnist), when other reporters checked the gentleman's back columns, they found nothing about predicting a revolution in Iran or anything about about a Middle Eastern ruler rising to power.

So here is one thing we must remember. If you want to really confirm a prediction you must go back and look for what predictions were documented before the event occured. If all you have are predictions - quote - "remembered" - unquote - after the event, you cannot accept the prediction as validated.

Another example of an oft cited prophetic success is from the 1960's. This was when another newspaper columnist predicted President Kennedy's assassination. We even read that on November 22, 1963, the columnist was sitting in a local establishment and told people that very day something terrible would happen to the president. Then someone rushed in and announced that shots had been fired at the president. However, no one knew if he had even been hit. But the columnist - who did double duty as a psychic - said that they would soon learn the president was actually dead.

Amazing! Nicht wahr?

Well, we have to say "nicht" and certainly not "wahr". This is a perfect example of predictive revisionism plus not checking the original documented prediction. In 1956 (yes, 1956), what the "psychic" (and we must use quotes) actually "predicted" (again in quotes) was that the president elected in 1960 would die in office or be assassinated. The "prediction" said nothing who the president would be or the manner of death.

But the real problem is this prediction isn't even a real prediction. All the columnist (i. .e, the "psychic") was doing was nothing more than rehashing that old chestnut, the "Curse of the Presidents". The Curse of the Presidents, if you haven't heard of it, says that presidents elected in a year ending with a zero (1840, 1860, etc.) will die in office or be assassinated.

And when and where did we first learn of the "Curse of the Presidents"?

It was in Robert Ripley's Believe It or Not.

In 1934.

Of course, the "Curse of the Presidents" still gets quite a bit of hooplah today. After all, the next two "curse-year presidents" after Robert Ripley's column - Franklin Roosevelt and John Kennedy - did indeed die in office. And yes, since these deaths happened after Robert's prediction, they can indeed be counted as legitimate tests of predictive power.

The Curse of the Presidents

Unfortunately (or rather fortunately), the curse seems to have stopped once presidential security got better and medical care improved. None of the two presidents after Kennedy elected on the curse years of 1980 and 2000 died in office.

Now, some people point out that the assassination attempt on President Reagan - elected in 1980 - should be counted since it was only with modern medical treatment that the president survived. And remember one of the airplanes of the 911 terrorist attacks was targeting the White House and George W. Bush - elected in 2000. Shouldn't this also count as fulfillment of the Curse of the Presidents?

Weeeeeeeelllllllll, not really. And to understand why, we have to go back to the difference between correlation and prediction.

Now it is perfectly acceptable to revise theories if the first theory isn't quite correct. But then you have to go back to square one to test whether the revised theory is predictive. In other words, if you change the Curse from "die in office" to "nearly die in office" or "could have died in office", your old theory is out the window.

Here the difference in correlation and prediction becomes important. The new data did not confirm the old theory that presidents elected in zero-ending years will die in office. So you modified your theory. But now what you intended to be new data to test the old theory has now become old data that you are now correlating with the new observations regarding Presidents Reagan and Bush. So you have to wait for new presidents to be elected in the zero-ending years - starting with the 2020 election - before you can test your new theory - that presidents elected in zero-ending years will die in office, nearly die in office, or could have died in office.

Also note how this Revised Curse has become difficult, if not impossible, to test. In the Old Curse if someone died in office, well, that was pretty easy to determine. But whether someone had "nearly died in office" or "could have died in office" is not so clear cut.

For instance, George Washington - elected in 1789 and 1793 - at one point developed a dangerous infection on his leg that required surgery which in the pre-anesthetic and pre-antibiotic era "could have" proven fatal. Woodrow Wilson suffered a debilitating stroke while in office, the seriousness of which had to be covered up. Two presidents contracted cancer during office, Grover Clevland (oral) and Lyndon Johnson (skin), both of which requires surgery which was covered up. And of course if you want to include Reagan or G. W. Bush in the curse because of the attempted assassinations, you should also count Harry Truman and Gerald Ford, neither of whom were elected in zero-ending years (in fact, Gerald Ford, the 38th US President, was never elected President).

Another lesson the Curse teaches us is that many bogus predictions "succeed" because the probability of a chance occurrence is greater than you think. For instance, you might think the curse means that only presidents elected in zero-ending years die in office. It says nothing of the sort. Instead, presidents can die at any other time as well. Sure enough, Zachary Taylor was elected in 1848 and died before his term was over. Also the president does not have to die in the actual term that ended with the zero. Lincoln was assassinated after his 1864 election and McKinley's first election was in 1896 Finally Franklin Roosevelt, first elected in 1932, actually died during his 1944 term. The "curse" year of 1940 was well-sandwiched within his more than 12 years as President. So it seems the curse not only "predicts" who dies in office, but who gets re-elected.

Another Count Against the Curse is that presidents elected in 1800 (Thomas Jefferson) and 1820 (James Monroe) left the presidency hale and well. So to counter this objection, some people prefer to call the "Curse of the Presidents" the "Curse of Tecumseh". The story, then, is that the Indian chief, Tecumseh, put a curse on William Henry Harrison (elected 1840) for his activities as an Indian fighter. Unfortunately, you really can't cite something that never happened to prove a prediction.

In the end, the "Curse of the Presidents" is - as is its equivalent 1956 "prophecy" of JFK's assasination - an example of a selectively worded "golf ball in the fairway" fallacy. This is where you consider a huge number of low low-probability events, one at least which must occur. Then when one does indeed occur, you say this was something amazing.

Alas, it isn't. Instead the Curse We See is simply one out of the billions and billions of patterns (actually a total of 1 X 10⁶⁵ patterns), at least one of which must have happened when you have eight US presidents check in while in office. So what we think is a pre-ordained pattern is nothing but one of the necessary and random occurrences. It's just as likely - or unlikely - as any of the other patterns.

But wait a minute, you say. Remember that the psychic had the premonition and said that the president was dead before it was even officially confirmed. What about that?

Well, that, mes amis, is the whole point. That story that wasn't told until 1965, almost two years after the event. So it can't be used as a "proof" of a prediction.

 

Back to the Bomber

All right, how does a "Believe it Or Not" column that turned into an American myth relate to the Mad Bomber? Well, there is a wee bit of a problem with the amazingly accurate profile about the Bomber we cited above, and which is the one that you inevitably read in articles, books, and on the Fount of All Knowledge. They were written after the Bomber was caught.

So how can we then validate Dr. Brussel's profile? Well, we must find the profile that was actually documented before the Mad bomber was arrested. So all we need to do is just go to the police records and retrieve Dr. Brussel's original report. Simple, no?

Well, no. So far no original pre-arrest document has come to light. Instead if we believe Dr. Brussel's own memoirs, we read - with verbatim quotes down to the expletives the detectives used - he gave his incredibly accurate profile verbally.

So what to do?

Fortunately, all is not lost. On Christmas Day, December, 1956 the New York Times published a lengthy and detailed account about the Mad Bomber. This story was written before the identity of the Bomber was known. So we can use a profile from this article if there is one.

And of course there is (or we wouldn't be bothering with this essay). Not just a profile, but one that is specifically attributed to Dr. Brussel by name and is given in quotes. So verbatim down to the punctuation and in toto is Dr. Brussel's original profile of the Mad Bomber of New York:

"Single man, between 40 and 50 years old, introvert. Unsocial but not anti-social. Skilled mechanic. Cunning. Neat with tools. Egotistical of mechanical skill. Contemptuous of other people. Resentful of criticism of his work but probably conceals resentment. Moral. Honest. Interested in women. High school graduate. Expert in civil or military ordnance. Religious. Might flare up violently at work when criticized. Possible motive: discharge or reprimand. Feels superior to critics. Resentment keeps growing. Present or former Consolidated Edison worker. Probably case of progressive paranoia."

Hm. Note how many of the amazing incredible and unbelievably accurate predictions have vanished, flip-flopped, or transformed to something vague and unprovable. At least one of the predictions - the Bomber being a past or present employee of Con Edison - had been already guessed by the police.

Here we need to elaborate on us making sure that profiling characteristics are actually testable. For one thing, we have to avoid predictions that are simply "correct by definition". After all, the Mad Bomber avoided being caught for 16 years. So "by definition" he's cunning. But we could have also said he was "lucky" and also been correct.

There are also predictions we claim as true because we are using our own personal private definition. That's the only way you can say someone is "moral" and "honest" when he goes around secretly planting bombs. And some characteristics are "qualified to neutrality". For instance, how can you tell if someone is "resentful of criticism" if he "probably" conceals his resentment? So if we find he's sullen and uncooperative after his arrest, we can claim he's resentful. But if he is smiling, helpful, and tells all, then we claim he's concealing his resentment. No matter what, we can claim we were right!

You have a problem akin to those with private definitions when you use words which at first glance are specific but really are so broad as to be without meaning. For instance, the Bomber we read about in the earlier profile is "religious". So our earlier Bomber could be Catholic. But he also could have been Hindu, Muslim, Jewish, Shinto, Buddhist, or Druid. He could be a regular church goer or perhaps not. Maybe he's someone who just follows their own personal philosophy of life. Once the famous folk singer, Woody Guthrie, told an interviewer he was a religious man. But he said, he had no favorites. "I kind of like them all," he added. So we see that the word "religious", although we think of it as having a definite meaning, from a profiling standpoint is horribly vague.

Then, of course, some of the predictions changed or were just flat out wrong. In the Times profile, the Bomber is said to be at least a high school graduate with expertise in civil or military explosives and weapons. The real Bomber was neither. And now we read he's not well-behaved, courteous, and friendly after all, but contemptuous of other people and might flare up at work. There is also nothing about the Bomber living with a female relative.

And in a complete 180 degree turn, the Mad Bomber in 1956 is interested in women rather than having little interest in women - which we read in the following decade.

Other writers have consulted other pre-arrest newspapers and mentioned that before he was arrested, the original Bomber was thought to be German, not Slavic, and would be living in White Plains, New York, not Connecticut. He would be a veteran of World War II where he learned his bomb making skills (the first bombs were crude and didn't go off, and the later bombs were more sophisticated and worked). In fairness, though, we're not positive these predictions were a direct product of Dr. Brussel's psychoanalysis. In any case, none of them were correct.

But above all, there is nothing - that's nothing, nicht, keins - about a double breasted suit, much less how many buttons were buttoned.

So what can you say? Well, possibly Dr. Brussel just didn't give his whole profile to the papers or he was still working on it. Perhaps the police didn't want to tip their hand and so removed the part about the double-breasted suits and living in Connecticut with his sisters. Unfortunately, that brings up another thing that is verboten in validating predictions. You can't make up facts to make your prediction true.

The more charitable critics can again cite "memory deterioration" coupled with "hindsight bias" in explaining the discrepancies. Once we know the answer, vague and even incorrect predictions adapt to fit what we learn later. Also you can rationalize you were pretty close, even if you are actually wrong. So if the Bomber is 53 years old, we soften our earlier precise quantitative (and incorrect) prediction of 40 - 50 years old to a more qualitative "middle aged". If he is of Baltic ancestry (i. e., Lithuanian), then we can claim that's not too far from being Slavic. Also if we predict a man was planting bombs because he was dismissed or reprimanded when it was really because he was denied workman's compensation, we can say we were at least sort of correct as long as he was in fact dismissed sometime after his injury.

And of course, if we now say a man is not interested in women and we earlier said he was interested in women, shoot, we can predict anything. Or maybe this was the first not-joke in history.

The first "NOT" joke?

Although some may want to argue that the newspaper account may be in error ("Hey, maybe the reporter just forgot to write 'not' before the 'interested in women'") that violates the main criterion for validation tests. No "predictions" - and we again must use quotes - written after the event can ever be used as a gauge of predictive power. And all unambiguous, objective, and documented predictions made before the event must be used - and as they were written. No cherry picking is allowed, החברים שלי. And you certainly cannot assume something was an error just because it proves your profile is wrong.

As a final aside, our exercise into profiling also shows us the necessity of retrieving the original sources when trying to document what is actually predicted. An article about the Bomber that appears on a popular informational website quotes Dr. Brussel's New York Times profile.

And it's accurate in its quotes.

Except for one.

Where the original newspaper story has Dr. Brussel saying the Bomber was interested in women, the website mysteriously morphs the newspaper quote as Dr. Brussel saying the Bomber is not interested in women. So even today Dr. Brussel can get a little help from his memory-deteriorating and hindsight-biased friends.

So what about that double breasted suit with the buttons buttoned? Where did that story come from? Well, in the first photographs after the Bomber was in custody, he was indeed wearing a double breasted suit with all the buttons buttoned. But once more we have nothing written before the photographs were taken to indicate that anyone thought this was his preferred garb. So unless someone can find mention of a double breasted suit that was actually written in 1956 (or at least before the police caught the bomber) we once more have to sadly relegate the "prediction" to a case of - quote - "memory deterioration and hindsight bias" - unquote.

Besides, George wasn't wearing a double breasted suit when arrested.

He was in his pajamas.

It's worthwhile at this point to summarize the before- and after- profiles in a table together with George's real profile. For succinctness, we are listing the characteristics that are most easily tested and those that are definitely ascribable to Dr. Brussels. Again we conclude that before the Bomber was caught, the profile was so-so at best.

* Prediction reasonably close; given benefit of the doubt.

So we see that if we use the profile stated after the Bomber was caught, we have 100 % agreement. But sticking to the profile before we knew who the bomber was, the profile is correct 30 % of the time - and this is giving the benefit of the doubt in questionable cases.

In conclusion, then, if we follow the proper guidelines for testing true predictive power, suddenly the cream of offender profiling success stories, that of the famous Mad Bomber of New York, fizzles into nothing more than what the cops already suspected as well as some so-so guesswork - some right, some wrong, and some too vague to test. But this point is buried beneath the mound of websites and books that inevitably quote the "predictions" from the 1960's.

 

Not Banned in Boston

Dr. Brussel continued to work in law enforcement and was asked, along with other psychiatrists, to profile the Boston Strangler. As you can guess by now, you can find articles on the Fount of all Knowledge that tell us that the Good Doctor's profile aided the police. For instance, Dr. Brussel was emphatic that the murders were the work of one man. And other psychiatrists disagreed. So it was Dr. Brussel's more accurate profile, according to the Fount, that led the police to Albert DeSalvo, a habitual con man, burglar, and serial sex offender whom we can now, after recent DNA tests, accept as the Strangler.

The truth, though, is taken in its entirety Dr. Brussel's profile wasn't that great. He said the Strangler would be single, have problems with his manly performance, and be "powerful and compact" like a ballet dancer. Now at the time of the stranglings, Albert was a bit stocky and in a few years pudged up quite a bit. However, in his earlier years, Albert had been a middleweight boxing champion, and some middle weight boxing champions are indeed built like some of the heftier ballet dancers. So we'll give that one to the Good Doctor. But otherwise the profile fell flat. Albert had absolutely no problem with masculine prowess (quite the contrary), and when arrested he had been married for a decade.

You would think that DeSalvo, as an ex-con and repeat offender, would have been already checked out by the police. But at that time, Albert's official conviction had been only for breaking and entering. The police, on the other hand, had guessed - "profiled" if you will - that the Strangler would be found somewhere among Boston's past sex-offenders. So Albert was not a serious suspect.

Then in 1964, a woman called the police that she had been tied up and assaulted in her home. Unlike in the inaccurate movie with Tony Curtis, the lady described her assailant with no trouble. A police artist made up a sketch and then one of the detectives, after seeing the assailant's distinctive pompadour and humongous schnoz, recognized DeSalvo. So Albert was arrested and identified in a line-up. Then when the story broke (with Albert's photo), other women came forward. Soon it was clear Albert had been involved in a large number of until-then unsolved sex crimes.

The extent of the new crimes were so serious and the evidence so strong that Albert was convinced (correctly as it turned out) that he was going to spend the rest of his life in jail. But more importantly, the new crimes put him back on the Strangler List. Then when the police found that he had been neither at work nor at home when every strangling had been committed, they felt they were onto something. But Albert still denied he was the strangler.

But only to the police. He not only told his attorney (who was bound by attorney-client privilege) but also a fellow prisoner, George Nassar. George, then told his own attorney who happened to be none other than the up-and-coming legal whiz, Francis Lee Bailey. Lee then arranged to meet with DeSalvo and took over as his lawyer.

But Lee wondered. Was Albert really the Strangler or just the type of blowhard you always find in jails? He didn't know. So he went to the investigators and got some information to test Albert's knowledge of the crime scenes. The answers convinced Lee that he was indeed representing the Strangler. By legal maneuvering, he managed, not just to save Albert from execution, but even prevented him from being tried for the murders at all.

So in the end the Strangler was identified by a combination of good police work, an enterprising attorney, Albert confessing, and finally in 2013, DNA evidence. The profiles of Dr. Brussel and the other psychiatrists had nothing to do with it.

 

Back in New York

And Dr. Brussel's profiling had nothing to do with catching the Mad Bomber either. The Bomber finally gave himself away. He not only wrote letters to the police but also to the newspapers. In one of his last communications he confirmed he had been an employee of Con Edison, injured at work, and was denied compensation.

At that point, George might as well have signed his name. The bosses at Con Edison decided to go back and asked a clerk, Alice Kelly, to recheck the files (Con Edison had in fact not destroyed the earlier records). She came across letters George had written seeking workman's compensation after an accident. But his claim had been denied on the technicality that he had filed for compensation too late. George had sent further letters, and Alice noted one particular comment where George had said that if necessary he would take events into his own hands. So Alice notified her supervisors, and they told the police.

George was arrested without incident and readily confessed. He was ruled hopelessly and incurably insane and sent to the Matteawan State Hospital for the Criminally Insane (today the Fishkill Correctional Facility) where he was finally released in 1973, completely cured and sane. He lived peaceably and without trouble until 1994 when he died at age 91. We wonder just how sick - physically or mentally - George really was.

The last point we need to address since it is a topic relevant when we talk about the Bomber, the Strangler, or even the notorious Jack the Ripper - also a popular candidate for profiling. Were these guys really nuts?

In the US, the definition of insanity - always a legal term - has varied from time to time and from state to state. In the good old days it had a rather flabby definition. A person could be ruled insane if he had a mental disease, didn't appreciate the significance of his actions, or suffered from an "irresistible impulse". But nowadays - if you whittle away the legalese - insanity is simply having a mental condition that prevents you from realizing the criminal act was illegal. Some states have even done away with the insanity plea altogether. George clearly knew that planting bombs was illegal, he was coherent and lucid, and he was capable of aiding lawyers for his defense. So had the Mad Bomber of New York gone to trial today he would certainly have been ruled sane and fully competent.

 

The Bottom-Up Approach

In recent years there have been attempts to objectively test offender profiling. But if you read the journals, the success rates don't look that great; 17 % - 20 % are typical values cited. So the question remains whether it helps the investigators or throws them off track. Even who writes the best profiles - professional law enforcement officers or psychologists - is disputed (particularly between law enforcement professionals and psychologists). One study found that the most accurate profiles did come from professional law officers with experience in profiling. But another study found the best profiles came from chemistry students!

Profilers, though, point out that such "lab" tests are not the real world and are not indicators of profiling's power. Fortunately, if we keep to the criteria we stated above - make sure the profile has definite testable characteristics and is clearly documented when first created - it will be possible to evaluate profiling effectiveness from actual police cases.

We mentioned there are efforts (particularly in the UK) to put profiling on a firm "evidence-based" foundation. What you have to do is find quantitative correlations that can be unambiguously linked from victims to the criminals and crime scene. For instance, there is a definite correlation (yes, correlation) between the age of homicide victims and the age of the criminal. You can see that by reviewing the raw data of previous cases, making a graph, doing the regression fit and calculating the confidence levels. You will end up with a graph like we see below, which, the statisticians tell us, shows there is a 99.999 % confidence that there is "real" trend.

But a warning is in order. A 99-plus % fit does not mean you can pick out the correct age with 99-plus % accuracy. That becomes evident if you actually plot, not just the line, but also the data (which in this particular graph we caution is "representational" rather than literal). But - and this is even harder to believe - you can deduce meaningful information from data even with such a shotgun scatter. Without going into details, we can show that the graph tells us that if we are investigating a homicide with a thirty year old victim , then there is a 95 % chance the criminal between was between 21 and 39 years old. But if you show such data in a court, you'd get laughed out of your briefs. So you can see why profiling, if used at all, is best as an investigative tool and should never be confused with actual evidence.

Still if you were a detective investigating the homicide of a 30-year old victim and you're stumped where to start, there's certainly nothing wrong with first checking out the victim's acquaintances between 21 to 39 years of age. Of course, the crime scene might have additional information that lets you narrow the search further. After all, if the victim was bludgeoned to death with a prune bottle, you might look for a more elderly criminal. But there's always - that's always - the possibility that the real criminal is outside the norm of past offenders. After all, some teenage gang members may like prunes. We repeat, a profile is not evidence.

 

Profiling Jack

So after this rather long and inconclusive foray in the case of the Mad Bomber, the Boston Strangler, JFK, and offender profiling, what have the various profilers told us about the most famous offender of all? We mean of course Jack the Ripper.

Student's of Jack the Ripper - called "Ripperologists" - know that profiling is by no means a new crime fighting tool. Almost as soon as the first victim was found, the police began to profile Jack. The basic principles were pretty much the same. Correlate the evidence on the crime seen with the type of criminals you know - or at least suspect - commit such crimes. The trouble, though, is at that time, serial murderers were a new phenomenon. What, then, has profiling told us about Jack?

Well, we learn he may have been between 25 and 45 years old. Or perhaps he was even older. He probably was not married or maybe he was. He was likely a loner but possibly liked to be around a lot of people. One profiler - with years of experience in law enforcement - said we could also narrow down Jack to someone who is "below or above average in height and / or weight". That really helps, doesn't it?

And of course, the profiles also tell us that Jack - who went around brutally murdering female inhabitants of London's East End - had problems relating to women.

Perhaps its not such a surprise that Jack was never caught. But in any case, profiling is here to stay. After all, it is a lot of fun whether it actually works or not.

Jack the Ripper - Never Caught

 

References

"16-Year Search for Madman", Phillip Meagher, New York Times, December 25, 1956. The author of CooperToons had always accepted the accuracy of the Mad Bomber profile which mentioned the double breasted suit with the buttons buttoned. So he was shocked! shocked! to read that if you want the amazingly accurate profile you need to go to what was written years after the bomber was caught. He was not, though, surprised to find that the "psychic" who predicted JFK's assassination had done no such thing.

The Mad Bomber of New York: The Extraordinary True Story of the Manhunt That Paralyzed a City, Michael M. Greenburg, Union Square Press, 2010. About as detailed an account about the Mad Bomber as you can get. Extremely well documented.

All in all the book is fairly positive about Dr. Brussel's profiling efforts and quotes considerably from his memoirs. Still the book mentions both sides of the profiling issue and points out that Dr. Brussel has been criticized for how his revised profile doesn't always jibe with the original. The New York Times article is also quoted as saying the Bomber was "[Not"] Interested in Women", and so at least honestly indicates a restored word, something the other article on the Fount of All Knowledge does not do. The book also points out that one detective thought that the profile was useless and so vaguely worded that it could have applied to virtually anyone, a common criticism of "top-down" psychological profiling.

Casebook of a Crime Psychiatrist James Brussel, Dell, 1968. Dr. Brussel's own account. Copies are rare and a bit more expensive than you might want to pay, although not ridiculously so. A good candidate to borrow on an interlibrary loan.

One of the problem's with Dr. Brussel's profile is it relies heavily on 1950's psychiatric theory. Even at that late date, psychiatry took a lot from Freud's basic ideas that aren't as popular - or at least aren't as believed - as they are now.

Psychiatrists probably get more static than any other medical professionals - and not just from the anti-psychiatric crowd. Once there were two psychiatrists and a neurologist sitting at a table (sounds like the beginning of a joke, doesn't it?) and the neurologist mentioned the difference between neurologists and psychiatrists. He said that psychiatrists are doctors trying to treat patients who may not have a real illness, and they use methods which no one knows if they work. Then, he said, the neurologists step in and find what the real problem is. One of the psychiatrists agreed!

"Freud's Not Dead; He's Just Really Hard to Find", Susan Whitbourne, Psychology Today, http://www.psychologytoday.com/blog/fulfillment-any-age/201205/freud-s-not-dead-he-s-just-really-hard-find. Before we start trashing Freud too much, we need point out that he was a 19th century physician practicing 19th century medicine. Still, his influence on later - and hopefully more correct - theories is considerable. This article points out that we should consider Freud more like Issac Newton. We still teach Newton in physics classes even though we also study later and more modern (and correct) theories like quantum mechanics and relativity. A Respectful CooperToons Opinion is this sentiment is correct, although he thinks we should consider Freud more like Aristotle.

Criminal Profiling: An Introduction to Behavioral Evidence Analysis, Brent Turvey. A textbook on profiling. New editions are a bit pricey, but inexpensive copies of earlier editions can be found. Oddly enough, in a major library system of one of the US's larger cities the only books about profiling are in the kid's section. Just say you're getting something for your nephew's social study class.

Using Homicide Data to Assist Murder Investigations, Brian Francis, Jon Barry Russell Bowater, Nicky Miller, Keith Soothill, Elizabeth Ackerley, Home Office Online Report 26/04. An example of how profiling is done in the UK using quantitative multivariate analysis. The methodology described used objective measurable characteristics and avoids using modern psychiatric theories which alter as society changes.

This report does mention there are problems with this "frequency" approach drawn from actual crime data. For instance, the data is rarely independently distributed over the variables, and the sample sizes tend to be small. So you might think you've found a link between ages of the victim and the criminal when you may really be looking at the gender relationship.

"Criminal Profiling: The Reality Behind the Myth", Lea Winerman, Monitor on Psychology, July/August 2004, Vol 35, No. 7, p. 66.

Introduction to Psychology, James Kalat, Wadsworth, 2008. There is a small sectin that summarizes the discrepancies in the pre- and post-arrest Bomber profile. However, it is not clear if some of the errors mentioned were actually from Dr. Brussel.

The Boston Strangler, Gerald Frank, New American Library, 1966. The definitive account although it omits certain tidbits such as when Peter Hurkos, the psychic who was called in to help find the Strangler, was arrested for posing as an FBI agent. It does, though, mention Peter picked the wrong man, something that Hurkos had done before.

This book disucsses the Mad Bomber profile with the double breasted suit and all the buttons buttoned. This was two years before Dr. Brussel published his memoirs. But it was still a decade after the Bomber was arrested.

Although this book was a best seller it has vanished from bookstores (the few that are left) and even major libraries. The author of CooperToons does remember when reading the book (at a relatively tender age) that he was a bit perplexed at how Dr. Brussel's profile of the Bomber could have been so accurate and yet that of the Strangler was so far off.

The Boston Strangler, Alan Rogers, (New England Remembers Series), Commonwealth Editions, 2006. A brief yet factual account of the Boston Strangler with mention of Dr. Brussel and the other psychiatrists. Also Dr. Brussel was not the only psychiatrist to think the crimes were all from one man. Dr. Phillip Solomon, head of psychiatry at Boston City Hospital, also agreed to write a profile for the Bomber and had formed the same opinion.

We need to point out that there were investigators that thought - and some still do - that the crime scenes were varied enough to indicate there was more than one murderer. And in fact there was at least one strangling in Boston at the time that was not by De Salvo. This crime, though, was solved fairly quickly and is now never included as one of the Boston Strangler. The DNA tests of 2013, though, have proven that DeSalvo was at the scene of the last victim and there was strong enough similarity in this crime and virtually all of the others that a Respectful CooperToons Opinion is DeSalvo was indeed the Strangler and guilty of all the crimes. The explanations for the apparent discrepancies are found in Albert's own confession.

From reading the books about the Strangler, you see how psychological profiling has an additional major pitfall. What is termed mental illness or personality disorders changes with time and depends upon the culture. In this case, some of the psychiatrists relied on mid-twentieth century definitions of mental illness and believed that "unstable homosexuals" were behind the stranglings. This is a classic case of how early psychiatric theory was often independent of facts. There was never any evidence then, nor now, that homosexuals are more prone to violence than anyone else nor that they target women.

Another point that this book brings out is that when you have many profiles created by many people, parts of some will be right, parts of others will be wrong, and you can pick and choose the ones you like. All in all, though, the real guess - "profile" - of the Strangler that turned out to help was the cops' thinking they were dealing with a man who was a previous sex offender.

The Boston Strangler case should also caution all data analysts that what appears to be definitive and meaningful trends can easily be due to chance. From the age of the first five victims the police thought the Strangler was specifically targeting elderly women. So the age of the victims was incorporated into the psychological profile. The fact the Strangler later switched to include younger women was also adopted into his hypothetical mental make-up. The truth is that Albert's selections were purely random.

This book also points out that Albert never recanted his confession which is contrary to what some other writers say. The confusion, though, comes from a statement Albert made after his famous escape from Bridgewater State Hospital in 1967. After he was recaptured, he was filmed saying "I never hurt anyone". So it's natural to think he's denying the original crimes. But before his escape he left a note to the warden saying he was escaping to draw attention to poor conditions at Bridgewater. He added that when he was out, he would not hurt anyone. So Albert's statement refers to his actions during the escape - that he had not hurt anyone - and he was not recanting his original confession.

Finally, here we do read about the story about Peter impersonating the FBI agent. When it came out that the Boston police had called in a psychic, the Attorney General, Edward Brooke, had to spend some time justifying what to many was a crazy idea and a waste of time. And indeed Peter was 100 % off. Though we read that a sketch of the Boston Strangler created from Peter's description did "resemble" DeSalvo, we also read it "resembled" the man Hurkos wrongly identified. This would lead the more critical reader to suspect Peter's description was rather vague.

"Terror in the Age of Eisenhower; Recalling the Mad Bomber, Whose Rampage Shook New York", Charles Delafuente, New York Times. September 10, 2004

"The Criminal Profiling Illusions: What's behind the smoke and mirrors?", B. Snook. C. Bennell, P. J. Taylor, P. Gendreau, Crimnal Justice and Behavior, Vol. 35 No. 10, October 2008, pp. 1257-1276

Ripley's Big Book: Believe it or Not, Robert Ripley, p. 296, Simon and Schuster, 1934.

"Loner, scarred face, speech problems': FBI dossier reveals chilling profile of Jack the Ripper... shame it was released 100 years too late", Sadie Whitelocks, Daily Mail Online, April 20, 2011. This is where we read Jack the Ripper was profiled as "below or above average in height and / or weight". No doubt.

"XXXXXXXXXXXXX",Website X on http://www.xxxxxxxxx.com. Lest the author of CooperToons appears overly critical and pedantic, he prefers not to specifically identify the article or the website that quoted the New York Times article as saying the Mad Bomber was not interested in women [sans brackets] when the actual Times article said he was interested in women. After all, we all make - quote - "mistakes" - unquote.