Thursday, April 01, 2010

A Consideration of the Criminal Investigation Process — Part One

El Inglés returns to the topic of law enforcement in the UK with the first in a two-part series of essays.

British police

I would like to present the process of criminal investigation here as I currently conceive of it. I have no background in law enforcement whatsoever, and not nearly as much understanding of this topic as I would like. Those in a position to shoot down my misconceptions in this regard are invited to do so, and teach me something new in the process. This ringing self-endorsement out of the way, let us proceed.

A Consideration of the Criminal Investigation Process — Part One
by El Inglés

The Two Sweeps

Criminal investigation consists of two non-overlapping though mutually-reinforcing processes that will, ideally, meet in the middle to deliver the perpetrators of crime into the waiting hands of the police. These approaches are what I refer to as the bottom-up sweep and the top-down sweep. Let us consider each in turn.

The bottom-up sweep is that in which evidence from crime scenes and other sources, typically consists of physical evidence, eyewitness testimony, and/or video evidence, is stitched together to provide clues as to the identities of the perpetrators. Police officers investigating crime scenes, using latex gloves to pick up empty shell casings, asking bystanders crowding round a corpse on the pavement if anyone saw anything, checking CCTV cameras to try to retrieve a mugshot of the gentleman in the top hat who held up the liquor store: these are all examples activities constituting a bottom-up sweep.

Note that some amount of evidence obtained in this manner will be required to secure a conviction in a case, unless the culprit, identified in whatever fashion, can be induced to confess, and, further, that the chances of said confession taking place in the absence of any such evidence may be rather slim. Also note that much evidence obtained via the bottom-up sweep will be confirmatory in nature. Obtaining fingerprints of the culprit in a given case will be of no value in tracking him down (unless their fingerprints are already on a fingerprint database, a possibility I ignore here to make the point), but of great value in convincing a jury of their guilt if they can be tracked down by some other means. Similarly, finding a shell casing from the firearm used in a homicide will not lead the police to the killer, but may help them secure a conviction against that killer if he can be identified by some other means and arrested with the firearm in question located conveniently in his bedside cabinet. Not for nothing do such types tend to dispose of weapons after use.

In contrast, the top-down sweep is that in which the police attempt to gain insight into who committed a crime by determining those who might have had reason and/or opportunity to do so. It is less physical and more human than the bottom-up approach, relying as it does on such things as familiarity with known criminals in a given area and the details of their criminal ways, interviews with the loved ones of murder victims to try to elicit information as to the likely killers, and using informers to gain access to the word on the street after a bank robbery. Thus, the top-down approach scans in the opposite direction from the bottom-up approach, sifting the entire population under suspicion in a different fashion. It does not directly generate evidence of anything; rather, it helps focus the bottom-up sweep on those investigative areas and/or targets where its efforts will be best spent, and allows bottom-up evidence such as fingerprint evidence to be checked against specific suspects.

The Relationship Between the Two Sweeps

When the police begin to investigate a crime, what they are actually doing is conducting preliminary bottom-up and top-down sweeps. These sweeps inform and refine each other, leading, in theory, to the eventual arrest and conviction of the perpetrator(s). Let us see how this might work out in the context of a hypothetical jewellery store heist, in which a gang of armed robbers have made off with millions of pounds worth of precious stones. This is, of course, a major crime, and one to which the police would undoubtedly be able and prepared to commit significant resources.

Once the crime has been reported to the police, one group of officers/specialists will be responsible for: a) checking any available CCTV footage; b) searching for forensic evidence such as fingerprints, hairs, blood, shoeprints, or articles of clothing left at the scene; c) ascertaining the exact modus operandi of the perpetrators, and d) obtaining witness statements from store personnel and any other witnesses to the crime. At the same time, a second group will busy itself with: a) drawing up a list of armed robbers/jewel thieves known to be at large in the area in question in recent years; b) asking paid informants whether they have heard anything about the crime 'on the grapevine'; and c) asking those being arrested for other crimes whether they have any pertinent information, perhaps offering to reduce sentences if cooperation is forthcoming.

Let us assume that our thieves are competent professionals, and carry out their raid with sufficient skill as to ensure that: a) CCTV footage shows only four medium-build white men in ski masks, arriving and escaping in a stolen car which is later found burnt out three miles from the crime scene; b) no forensic evidence of note is left at the scene; c) the MO of the crime has no particular distinguishing features; d) witness statements provide no clues as to the identity of the perpetrators. This degree of professionalism places the onus squarely on those performing a top-down sweep, who have more luck, being able to ascertain that: a) a British armed robber of note, whom we shall call Mister A, was released from prison in France six months previously and known to be back in the UK; b) the word on the street is that said armed robber had been putting together a crew for some sort of 'job'; c) according to a burglar arrested for an unrelated crime, who knows a man who knows a man, Mister A has already starting dipping his toe in the waters of illegality. On the basis of this information, and having no other leads at all, the police decide to focus on Mister A as the target of the investigation.
- - - - - - - - -
So far then, the police have a complete failure on the part if the bottom-up sweep, but some progress with the top-down sweep. Pursuing this sweep further, they locate Mister A and place him under physical surveillance for several weeks. This allows them to piece together a picture of his activities and contacts. Further investigation reveals that several of the people so identified have previously come to the attention of the police for one reason or another. Three in particular have convictions for crimes of violence and larceny, and appear strong candidates for being Mister A's accomplices during the robbery.

Continuing physical surveillance of Mister A and his possible accomplices, Messrs. B, C, and D, reveal that Mister D is less careful than the others in keeping a low profile. He is observed to be spending a great deal of money, despite having no job and no known income from other sources, and the police obtain permission to place him under electronic surveillance, tapping his home phone and his mobile phone. Mister D continues to dig his own grave by discussing the possibility of unloading some quality merchandise with a man already known to the police as a fence for jewellery and other upmarket items.

Further surveillance allows the police to ascertain the time and place of the meet between Mister D and the fence, upon which they swoop. Mister D, the fence, and their companions are arrested, and merchandise stolen in the original jewel heist recovered from the possession of Mister D, who confesses to his role in that crime. A warrant is issued for Mister A's arrest and he is informed, untruthfully, that Mister D has fingered him, Mister B, and Mister C for the initial heist. He is also shown the jewellery recovered from Mister D, upon which he admits his part in the crime and testifies against Mister B and Mister C, something which Mister D refused to do. All four men are sentenced to long jail sentences.

I cannot vouch for the realism of this scenario in every detail, but the objective is not to be accurate in every regard. Rather, the objective is to highlight the interaction between the top-down and bottom-up sweeps. Consider a country with a population of one million in which a murder has just been committed, and the population to be a series of one million black circles laid out in a line. The bottom-up approach is akin to a robot running along the underside of the line flagging individuals consistent with the specific evidence it has gathered, while the top-down approach is a robot running along the top of the line flagging individuals considered plausible on the basis of its investigations. The objective of the process is to get at least one match but as few as possible. These matches are candidates in the investigation. The actual culprit may not be included among them, but that is simply one of the difficulties the police face.

Note that each sweep, as it makes progress, provides fodder for the other to refine and focus itself. The physical evidence that Mister D took part in the heist (the stolen jewellery in his possession) did not fall into the lap of the police from on high. Rather, an opportunity to discover it was created through careful investigative work based on the initial top-down sweep. The initial bottom-up sweep generated no hits at all, but the initial top-down sweep narrowed down the list of all possible culprits to a strong candidate. A fresh, tightly-focused, and patient bottom-up sweep (the raid on the illicit jewelry sale) was then initiated with the desired results. Numerous examples could be given of mutual feedback between the sweeps.

The Efficacy of the Sweeps

It is worth pointing out that different types of crime, and different types of criminal, will prove more amenable to being solved and identified respectively through different combinations of the sweeps. Burglary is a crime which is notoriously unresponsive to police investigation. Often one hears stories of the UK police barely taking the crime seriously at all. Frustrated homeowners ask attending officers (when those officers bother to turn up at all) what the chances of catching the burglar are, only to be met with shrugs of indifference. Though I sympathize with the rage people manifest in response to what they see as the indifference of officialdom, let us consider the problem in the context of the top-down and bottom-up scanning mechanisms.

The bottom-up scanning mechanism is greatly enfeebled by the relative lack of importance attached to burglary as a crime, which means the police tend not to sweep for fingerprints, DNA evidence, or the like. Indeed, they cannot, or police budgets would be consumed on that alone. The only type of bottom-up evidence likely to emerge should the burglar not be caught in the act is eyewitness testimony, which may not exist in any given case and may not be useful even if it does. A bottom-up sweep of the pool of all potential suspects is exceptionally poor even for generating confirmatory evidence for a crime like burglary unless the criminal makes a howling error of a type easy to avoid (e.g. dropping a signed photograph at the scene of the crime) or is extremely unlucky.

The top-down scanning mechanism is potentially more useful for burglary cases, in that most areas will have a number of usual suspects known to engage in certain types of crime repeatedly. Locking up a single one of these offenders will prevent a large number of burglaries while they are incarcerated. However, given that new burglars are presumably coming into existence all the time, even these top-down sweeps will often wrongly rule out actual culprits. Worse, unless the number of potential culprits can be greatly reduced in this fashion, it is not obvious that the time of the police is well used in questioning known recidivists in the hope that they let something slip. Thus it is that the clearance rate for burglary is largely a result of individuals asking for large numbers of former offences to be taken into account when the long arm of the law finally catches up with them selling off stolen goods or sneaking out of someone's house.

Note that, all other things being equal, the seriousness of the offence is a key factor in determining how likely the culprits are to be brought to justice. Though the careful criminal can take steps to reduce the amount of evidence left at the scene of the crime (or anywhere else), reducing the ripples caused by the entire process of planning, committing, and trying to get away with a crime will be far harder. If my house were burgled, the police would not be prepared to commit great resources to investigating the crime. However, if the President of the USA were murdered in my back yard by criminals a hundred times more professional, they would still be more likely to be caught, as the US federal government would leave no turn unturned in its hunt for the killer.

It is hard to overstate the importance of the following point: rendering either of the two sweeps ineffective kills the investigation. If the bottom-up sweep comes up empty, no top-down sweep can secure a conviction, however accurately it may narrow down the candidates for investigation (Note: this ignores the effects of interrogation, which will be discussed in the second installment of this essay). Of course it is hugely to the disadvantage of a criminal to be identified in this fashion if he hopes to continue a life of crime; he will be under close scrutiny thenceforth, and fresh evidence gained in the future may be used to secure convictions for crimes he thought he had got away with. Nonetheless, the fundamental point remains that, at least for the time being, they are in the clear. Conviction on the part of the police that they know who committed a certain crime does not constitute evidence that will stand up in court.

Similarly, ignoring that small subset of all crimes that can be solved purely by bottom-up sweeps (e.g. a rape committed by someone whose DNA is already on file and from whose victim a semen sample can be retrieved), rendering the top-down sweep ineffective leaves the investigators nowhere to turn. As already stated, evidence obtained in bottom-up sweeps tends to be confirmatory rather than suggestive of the identity of the perpetrator. When evidence thus obtained is strongly suggestive, the crime in question will almost certainly have been committed by a very unlucky individual, or a very stupid individual. Subsequent to a recent murder in the UK, one of the murderers took the victim's mobile phone, inserted his own SIM card, and started using it as his own, within minutes of the murder itself, allowing the police to identify him through the phone. A similar rank stupidity was committed by another murderer, who used his victim's registered Oyster card to travel home on the London Underground after committing the crime, thereby allowing the police to track him via CCTV footage taken from the station he disembarked at.

We must give thanks that these dangerous criminals were so careless in their post-crime behaviour. But it remains the case that even a modicum of intelligent behaviour and anything other than the most terrible luck will render it impossible for the police to identify criminals through bottom-up sweeps alone. Accordingly, effective top-down sweeps will tend to be required to bring to justice those who are not cognitively challenged. Not for nothing are the police so keen to try and determine who might have had a motive to commit, say, a murder. Without a plausible suspect, where do they begin their investigation? If it transpires that the person in question had just defrauded someone, or was having an affair with someone else's wife, then their job suddenly looks a lot easier, because their top-down sweep immediately identifies the defrauded business partner or the cuckolded husband as the prime suspect.

Of course, the real suspect may be someone else, but given that most people who are murdered are murdered by someone they know, the police are on strong ground now, statistically speaking. Assuming their prime suspect is in fact guilty, the investigation will now turn to whether or not the police can marshal enough evidence on the basis of a bottom-up sweep to get their man. Will the one eye-witness to the crime pick the suspect out of a line-up? Does the suspect have a credible alibi? Can the police find the murder weapon and establish that it was in the possession of the suspect? Did the suspect leave any physical or DNA evidence on the victim or at the scene of the crime?

The Primacy of the Top-Down Sweep

It appears to be the case that, on net balance (and we must generalize here to say anything useful at all), the top-down sweep is not only more important than the bottom-up sweep in criminal investigation, it is very significantly more important than the bottom-up sweep. Decades of unrealistic police dramas notwithstanding, police work does not tend to consist of Sherlock Holmes-style deduction, where the slightest clues are incorporated into a brilliantly accurate and irrefutable account of how the crime took place. The ability of the police to identify strong suspects for a crime (i.e. conduct a top-down sweep that throws up legitimate targets of inquiry) and then assemble as strong a case as they can out of the evidence they have (push the case as far as they can with the newly-focused bottom-up sweeps and interrogation based thereupon) is largely determinative of whether or not they can bring charges against careful criminals. Fools who blunder through open tins of paint at the scene of their crimes and then stroll the twenty yards home may not be covered by this generalization, but I am not especially concerned with them. The criminal with a sense of professional pride and concern for the craft is the central concern of this essay.

Given the threat the top-down sweep poses to the discerning criminal, it is worth considering in a little more detail just how broad this sweep can be in a modern, technologically-advanced society. Consider the following case. Some months ago, a gang of three youths were convicted of carrying out three home invasions in the UK. According to the newspaper reports that accompanied their convictions, they had been, despite their youth, an unusually careful gang, conducting their raids at different locations across the country so as to avoid quickly coming to the attention of any one police force.

However, integration of police databases and crime reports on a nationwide level is a key concern not only for the British police but for police forces in general, and this case gives us some insight into why. Let us assume that the gang carried out its home invasions in a competent fashion, wearing masks at all times, not allowing the licence plates on its vehicle to be seen by anyone at the crime scene, and taking care not to leave physical evidence at the scene. One might naively think that, given this degree of care and their habit of targeting widely separated properties, the gang would have been virtually untouchable, but this was obviously not the case. Now for all this author knows, one of the gang could have drunk too much one night and carelessly told his girlfriend the source of his newfound wealth. But let us ignore that possibility here and return to the top-down sweep to illustrate the principle, which is important and becoming all the more so irrespective of whether it actually applied in this case.

Despite the best efforts of the criminals, the pattern their string of home invasions constituted would undoubtedly have been picked up on quickly by whatever computerized criminal database our lords and masters have been busy assembling over the last few years; three home invasions, almost certainly the same perpetrators. Let us pose a question. What are the chances of any given person or vehicle having been in close proximity to any two of the targeted houses at the times they were targeted if they had not been involved in any way with the crime? The answer is clearly very low, and would drop away effectively to zero for three crimes or more. If the police have available to them any means at all of retrospectively scanning, in an admittedly partial fashion, the environs of each crime scene at and around the time the crime was committed, they may well identify persons or vehicle who were present at least twice, and, if they were lucky, all three times. But would these scans take place?

We already have an automatic licence plate recognition system covering a substantial portion of the UK's road traffic (presumably tied up with the speed camera system and therefore operating across a significant fraction of major roads throughout the country). London in particular will have other CCTV systems, partly as a legacy of past IRA bombing campaigns and partly as a response to the emerging threat of Muslim terrorism. Efforts are currently ongoing to conduct a massive expansion of this system to cover virtually all UK road traffic. It should be becoming clear why conducting multiple home invasions while driving to one's targets in the same vehicle is a virtual guarantor of a long spell at Her Majesty's Leisure. The phenomenal power of the police's souped-up top-down sweep must be understood, and similar sweeps considered: the already-mentioned electronic Oyster card introduced for the London Underground system in recent years, credit and debit cards (whether checking for location or tickets purchased), mobile phone records which allow phone (and therefore user) locations at given times to be ascertained through proximity to different masts, and undoubtedly others too. The introduction of facial and even gait recognition systems and their application to vast databanks of CCTV camera footage cannot be far away, at least technologically. Telephone and email surveillance, both widely used, fall into a slightly different category as they possess content and are therefore probably better categorized as part of a bottom-up than a top-down sweep. Nonetheless, their importance in this respect should not be underestimated.

If I were interested in conducting a series of home invasions across the UK (which I am not), I would add certain things to my to-do list and, as a consequence, probably stay out of prison at least somewhat longer than our trio above. I would not use the same vehicle to travel to each target, and would not use my own vehicle or that of any conspirator under any circumstances at all. At the very least I would rent different cars on different occasions with different credit cards to make it impossible for the police to get an immediate match with their scan. Better still would be to use cars that I had borrowed on pretexts from friends or relatives not obviously connected to each other and not living in the same area. Best of all would be to steal a car and dump it at the end of the job, though that of course creates extra risks. I would also ensure that no one in my gang took any sort of mobile communication device with them on the job, and that they left all credit and debit cards at home to assure they were not inadvertently used, with all that that could later imply for the plausibility of alibis.

If these things seem a bit excessive, the reader should bear in mind the earlier point about how the resources the police bring to bear on a crime are roughly proportional to its seriousness. If you plan to be driving around the country, kicking people's doors in, pointing guns at their heads and threatening to shoot them if they refuse to give you the combination to their safe, you had better expect the police to start pulling out all the stops. And the longer you get away with it, the deeper they will dig, and the greater the chances of their souped up top-down sweeps bringing your name up on a screen.

Interestingly, any evidence gained in this fashion, or by other top-down sweeps, remains very strongly circumstantial unless and until the sheer coincidence it represents becomes such that a jury would not believe it was a coincidence. What if the new, focused, bottom-up sweeps still yield nothing, as they may well? Are the police then confounded, left to lick their wounds and wait for the next opportunity to put the criminal in question away? Not at all. Once the perpetrators have been correctly identified, they are in serious trouble even in the absence of conclusive evidence. Why this should be the case will have to be explained in the second and final part of this essay.