by Barry Krusch
To the next decision point: we have gathered all the available evidence, and have stipulated that the evidence for the prosecution is credible, and when we examine that evidence we find that the evidence for the prosecution taken together clearly supports the charge, so therefore we have successfully answered the question “is the evidence for the prosecution sufficient?”, which means that the prosecution has presented enough positive evidence to move to the next step.
Our final decision point is “is the evidence related to the charge consistent?” Since we have all the relevant evidence, and all the relevant evidence for the prosecution passes the “smell test,” and the evidence presented by the prosecution is sufficient to demonstrate the charge, we now subject the defense evidence to the same credibility tests, and once we determine it passes, we then need to see if all the individual pieces of evidence (not just that presented by the prosecution) are consistent with each other (that is to say, that each individual piece of evidence provided by prosecution and defense tells the same story). And in this example, we find out that for the most part it does, although we do have one small piece of evidence (provided by the defense) that does not tell the same story, which is that 35 years before the crime was committed, Jim became an Eagle Scout, and if we assume that all Eagle Scouts are generally of sound moral character, this evidence is somewhat inconsistent.
However, this minor inconsistency is easily dealt with since the inconsistent occurrence occurred 3.5 decades prior to the commission of the crime, and the character of a person can easily change over that period of time. If Jim were 18 years old, the inconsistency of this evidence would present more problems, although that inconsistency would ultimately be defeated by the overwhelming evidence for the charge, and the melancholy recognition that not every Eagle Scout is necessarily 100% pure.
Having run our evidence through these four filters, we find that the evidence makes it through for the most part with ease.
So, now we have the basis for assigning a confidence level, which, according to our flowchart, will be greater than 95%. But what precise number do we assign? Most people, when asked, would now say that it is 100% certain that Jim is guilty, but even if this is not the case, we can certainly say that the confidence level at the high end is above 99%, approaching Six Sigma levels. We have all the evidence we need, every item of evidence is legitimate, every item of evidence provided by the prosecution (taken as a whole) is sufficient to prove the charge, and there is no evidence provided by the defense (or prosecution) which substantively contradicts the charge.
Of course, there is always room for a little doubt, especially given Jim’s Eagle Scout achievement, so we can reduce our rating somewhat at the lower end of our range. Given this, let’s assign a confidence level range with a minimum value of 99, and a maximum value of 99.99 (it is hard to be 100% sure of anything!), which results in the following table:
We could average these numbers to give us one number to work with, but for now, let’s just stick with the range — a prosecutor’s dream. Our confidence level for the charge is solidly above 95%, therefore Jim is “guilty beyond a reasonable doubt.”
Now let’s take a different scenario, an example of an entirely different sort which will give us a low confidence level.
THE CHARGE
Jim has been charged with the crime of printing counterfeit bills. The prosecution and defense have introduced the following evidence into the record:
EVIDENCE FOR THE CHARGE
Prosecution Evidence
Jim tried to pay for a gaming system with counterfeit notes.
Defense Evidence
Jim became an Eagle Scout at the age of 15.
That’s it. That’s the evidence. So we look at our first criteria, and then ask ourselves the first question which follows from that criteria, “is the evidence related to the charge comprehensive?” We will assume that it is all the evidence that could reasonably be gathered (meager though it may be), and proceed to the next decision point.
Our next question is “is the evidence for the prosecution credible?” And the answer is “yes.” Jim actually did pass the notes, and he was an Eagle Scout.
Our next question is “is the evidence for the prosecution sufficient?” And at this point the red light flashes: the answer clearly is “no!” Jim has been charged with the crime of printing counterfeit bills, not passing counterfeit bills! And, in fact, no other evidence has been produced for the charge; no printing press, no videotape security camera record, no confession, no nothing. A skimpy record that screams “inadequate.”
Notice that because the evidence for the prosecution (who has the burden of proof) is not sufficient, we do not even need to move to our final decision point. If the evidence for the prosecution is not sufficient as it stands, we obviously do not need to worry about whether Jim actually was an Eagle Scout, nor whether that evidence is consistent with Jim’s Eagle Scout achievement (or any other evidence).
At this point, we now are shuttled down to the end point in the flowchart that says that we should assign a confidence level lower than 95% to this charge. But what specific confidence level is that? Here, there is more room for disagreement. Some might go as low as 1%, others might go as high as 15%, but your author’s assessment is as follows:
If you disagree with this assessment, that is fine, and to be expected. Note, however, that relative to the end goal your assessment is probably not all that far away from mine, and I think it’s safe to say that both of our assessments, even though they might be somewhat divergent, are still well short of the 95% threshold required; accordingly, even though we disagree as to the exact numbers, we both agree that Jim must be found “not guilty” of this charge, which from the due process perspective is all that counts.
Okay, enough with the elementary examples, prototypes of easy decisions to illustrate how the breakout flow paradigm works. Now let’s graduate to a higher level of analysis and muddy the waters somewhat with a more complex example.
THE CHARGE
Jim has been charged with the crime of printing counterfeit bills. The prosecution and defense have introduced the following evidence into the record:
EVIDENCE FOR THE CHARGE
Prosecution Evidence
Jim tried to pay for a gaming system with counterfeit notes.
The police found a printing press in Jim’s house, with $7,500 in freshly minted bills (with identical serial numbers) next to the printing press.
Jim’s fingerprints were on the printing press. And the money.
Defense Evidence
A security camera in Jim’s house videotaped a person (unknown to Jim, according to Jim’s testimony) bringing the printing press into Jim’s house while Jim was on vacation, and also videotaped that person printing $7,500 worth of counterfeit bills. Jim is seen on the videotape, upon returning from vacation, apparently surprised that the printing press is in his house. He touches the printing press, and when he sees the money next to the printing press, picks the money up, and then puts some of it into his wallet.
Jim’s explains to the police after he is arrested for passing counterfeit bills that he had hired a housesitting service to watch his house while he was on vacation, and, unknown to Jim, the person hired to do the housesitting used Jim’s home to set up a criminal enterprise.
More complex, but once run through the four decision points, a clear answer should still emerge. Let us assume, for the purposes of this discussion, that the first two decision points are answered in the affirmative; this is all the evidence, and the prosecution’s evidence is credible. But now we have to pause at our third decision point. Here, we again need only focus on the prosecution’s evidence, and ask the question, if this was the only evidence offered in this case, would we be 95% confident that the charge was true? If you answer “no,” you are done, but if you answer “yes,” you still have one more hurdle, and it’s a big one. Let’s answer “yes,” and see why.
We analyze our defense eviden
ce, determine that it is credible, and now ask, “Is the evidence related to the charge consistent?” We find out that the answer is no — which changes everything!
Our confidence level for this charge, which might have been 97% or 98% after analyzing just the first three items, now has to be reduced because of this contradictory video evidence, and, based on the impact of this evidence (which positively refutes the charge), reduced to a very great extent. Again, there is going to be some room for disagreement as to how much the confidence level of the charge has been reduced, but that’s what jurors are for. If you were a juror, what would you say? Feel free to fill in the blanks!
To my way of thinking, the video is a sledgehammer that shatters the prosecution’s case. Therefore, my own personal view (relevant were I a prospective juror) is that, based on the evidence offered by the prosecution and defense (assuming both are equally credible), my minimum confidence level for the charge would be 0%, and my maximum confidence level for the charge would be 5% (based on the exceptionally slight possibility that the video had been tampered with). Contrast that evaluation with yours.
I assume that some readers would have put a maximum confidence level of 25%, versus my 5%. But again, recall, even with a disagreement as profound as this, the higher confidence level of 25% is still well short of the 95% threshold required; accordingly, Jim must be found “not guilty” of this charge (though possibly guilty of the related change of passing counterfeit bills).
So far, our linear model of the flowchart has worked fairly well. But these have been comparatively simple examples. Our next hypothetical is far more complicated — somewhat like the Kennedy case, though by comparison much simpler — and will present extraordinary challenges for our linear model:
THE CHARGE
Jim, age 50, has been charged with the crime of printing counterfeit bills. The prosecution and defense have introduced the following evidence into the record:
THE EVIDENCE FOR THE CHARGE
Prosecution Evidence
Jim tried to pay for a gaming system with counterfeit notes.
The police found a printing press in his house, with $7,500 in freshly minted bills next to the printing press.
The police also found the following book in Jim’s library: “How To Forge A New Identity.” Jim claims the book was given to him as a gift.
Jim’s fingerprints were on the printing press. And the money.
A security camera in Jim’s house videotaped Jim printing the money just two days before he was arrested, according to the testimony of a detective.
An email sent through his Gmail account from Jim to his friend Jack quotes Jim as writing “how easy the counterfeiting racket is.”
When shown the evidence, Jim admitted to the charge, and explained that he “got into the counterfeiting game” because he “needed the money.”
Defense Evidence
A week before, Jim had advertised his 20-year-old car in a classified ad in the local paper, and 3 days before Jim sold that car to a man named Bill, who paid Jim in cash. Jim used this cash to purchase the gaming system. The ad was definitely placed by Jim, and entered into evidence by the defense. Bill, however, cannot be located by the police force. Still, the title was recorded by the DMV as having been transferred the day before the notes were passed to a person named “Wendy.”
Jim claims the printing press was used to print flyers for his church. Jim denies that he had $7,500 in his room, and says that if there was any money there, it was planted by someone. A police log showed that $7,500 was found. A detective testified that he personally only found $2500, and is not sure why the log says $7,500. There is no other testimony by any other policeman that verifies the amount on the log.
Jim claims a security videotape would prove the money was planted. Unfortunately for Jim, the police confiscated the tape, and originally refused to release it until a period of 75 years had passed. Eventually, under pressure by the defense, they did. When the tape was run, unfortunately, the defense found a 15 minute gap, and even more unfortunately, the gap was located precisely in the area that would have exonerated Jim. Police initially denied that they caused this gap, but later changed their story to this: a new hire at the police force had inadvertently erased that part of the tape! “No problem,” say the police. One of their detectives testified that he saw the tape before it was erased, and in that 15 minute segment, he saw Jim printing the money. The detective who made this allegation had been on the force for 20 years, had received numerous commendations, and had been a member of Jim’s church for 20 years. The detective also passed a lie detector test.
Jim was going to call five other witnesses for his defense. As luck would have it, though, one witness died of natural causes, two died in an accident before the trial, and the other two who had signed pre-trial affidavits confirming Jim’s story then changed their mind and refused to testify in his defense. This left Jim with no witnesses and two discredited affidavits.
With regard to the fingerprints on the printing press, Jim said that he handled the printing press in the course of printing the flyers. His fingerprints were on the money because after receiving it from Tom, he counted it.
Jim denies sending the e-mail. His friend Jack went on a long-term vacation overseas, and could not be personally contacted, but an affidavit provided by the prosecution, in Jack’s handwriting, confirms the receipt of an e-mail without specifically addressing the contents.
Jim said that he was coerced into making a confession, but failed a lie detector test administered by the police department. However, the defense found three witnesses to testify that the person administering the lie detector test had altered the data on three separate occasions. Two of these witnesses, as it so happened, had a prior forgery conviction.
Uh-oh!! A little different, huh? This is the type of devilishly complicated hypothetical situation law school professors provide as fodder for final exam essays, and one that provides a hint of the complexity offered by the Kennedy case. And it is extremely instructive in at least one way; notice, when you saw only the prosecution’s evidence, you were ready to hang Jim high; but now, this evidence presented by the defense changes everything. In addition, and more to the point being made here, you’ll notice that due to this complexity, this hypothetical makes the breakout flow model unworkable.
Suppose we ask our first question, “is the evidence related to the charge comprehensive?” When we analyze the data, we quickly discover that we can’t reliably answer this question! To take the most obvious example, a key piece of evidence — the tape — was obliterated at the most important point. The evidence could have completely exonerated the defendant! And yet, while we might want to conclude that negligence of this sort on the part of the authorities ought to play the same role in this case as the role the atom bomb played to Nagasaki, we do have the testimony of the detective that Jim was lying. The detective appears to be reliable, and we do not want to believe that a detective would commit perjury. Can this evidence be seen as a legitimate substitute? To ask it another way, since the tape was erased in the worst possible place, can we really say that the detective’s “fill-in” testimony — supposedly reliable though it may be — is an adequate substitute for the cold hard evidence the defense would have had had not the police negligently handled the evidence?
Now, you might be confused, and consequently undecided at this point, but unfortunately the flowchart really doesn’t allow for indecision of this nature. And simply moving forward won’t be of much help, because assessing the credibility of the prosecution will require a simultaneous assessment of the credibility of the defense.
At this stage, we can see that the linear model is going to present some difficulties, so let’s take a stab at a new model:
We can refer to this new approach as the dynamic input model, a model which allows for a more or less simultaneous analysis (i.e., it need not be analyzed in any particular order). What this model sacrifices in certainty, it gains in flexibilit
y. Different weights can be assigned to the parameters as warranted by the evidence, and in no particular order.
The center rectangle with broken lines represents a bar indicating confidence level, a bar to be raised or lowered depending on the quality and quantity of the inputs. For example, if the evidence is only 90% comprehensive, the bar can be reduced to a 90% level. If the evidence for the prosecution lacks credibility, the height of the bar will be lowered, even more and if the evidence for the defense lacks credibility, the height of the bar will be raised. If the evidence presented by the prosecution lacks consistency, the height of the bar can be raised or lowered depending on the nature of the inconsistency. And so forth.
Because this model is not linear, our questions change to reflect the reality that they can be asked in any order. Let’s try to answer these questions in relation to our latest hypothetical scenario, and see how the answers affect the height of the bar:
How comprehensive is the evidence?
The evidence is not nearly as comprehensive as we want it to be. The 15 minute gap is a very serious omission. It is true that the detective has provided substitute testimony, but at the expense of a substantial reduction in the quality of testimony; the detective’s testimony functions almost like hearsay evidence, an evidence of much lower quality, generally considered inadmissible.
How credible is the evidence?
As indicated above, there are serious credibility issues with both prosecution and defense evidence. However, since the burden of proof is on the prosecution, if its credibility issues are not successfully addressed, the credibility issues of the defense are not nearly as important.
