"Smart people (like smart lawyers) can come up with very good explanations for mistaken points of view."

- Richard P. Feynman, Physicist

"There is a danger in clarity, the danger of over looking the subtleties of truth."

-Alfred North Whitehead

March 19, 2011

The Problem(s) with Memory - Part Two

In my last entry, I discussed three of the seven “sins of memory” identified by Daniel L. Schacter, in his book, The Seven Sins of Memory, How the Mind Forgets and Remembers. I will now discuss two more sins—ones that threaten the certainty of recollections we hear coming from a witness stand.

The fourth “sin of memory” is misattribution. Misattribution includes believing we remember things, that in fact, never occurred, and believing we are imagining things that we are actually remembering.

Schacter writes: “.. misattributions are surprisingly common. Sometimes we remember events that never happened, misattributing speedy processing of incoming information or vivid images that spring to mind, to memories of past events that did not occur. Sometimes we recall correctly what happened, but misattribute it to the wrong time or place. And at other times misattribution operates in a different direction: we mistakenly credit a spontaneous image or thought to our own imagination, when in reality we are recalling it—without awareness—from something we read or heard.”[i]

Examples of misattribution are déjà vu and erroneous eye witness identifications. Schacter discusses numerous examples of individuals falsely accused of crimes based on faulty eye witness identifications. He discusses the research on “unconscious transference” where one unconsciously transfers memory of an individual from one context to another, and binding failures where we incorrectly glue together various pieces of memory to create a fabricated memory.

The fifth “sin of memory” is the sin of suggestibility.

Schacter writes: “Suggestibility in memory refers to an individual’s tendency to incorporate misleading information from external sources—other people, written materials or pictures, even the media—into personal recollections. Suggestibility is closely related to misattribution in the sense that the conversion of suggestions into inaccurate memories must involve misattribution. However, misattribution often occurs in the absence of overt suggestion, making suggestibility a distinct sin of memory.”[ii]

This sin of memory is also implicated in false eyewitness identifications through things such as show-ups or poorly conducted lineups. Further, Schacter reports research that casts considerable doubt on an eyewitness’s evaluation of their certainty of their identifications. Anyone who has defended someone on charges based on an eyewitness testimony knows its power on a jury, especially the eyewitness’s testimony that they are “certain” that the defendant was the person who committed the crime. The research Schacter references infers that a very uncertain identification can be transformed into a certain identification merely with the word “Okay” spoken by the officer administering the line-up.

As a defense attorney, I was involved in an attempted homicide case that involved an issue of whether or not a knot would slip. I claimed that the knot would slip—which would help exonerate my client. The State argued that it would not slip. An investigator with the DA’s office and I took the knot to a Coast Guard Officer in Milwaukee. I was watching the officer play with knot. Unfortunately, the investigator had left the room for a minute. While the investigator was playing with the knot, the knot slipped. When the investigator came back into the room I told the investigator what happened. The investigator said to the officer: “It didn’t slip did it?” The officer immediately said, “No”. I couldn’t believe it. We went to trial. The knot slipped at trial, and my client was acquitted. I do not believe that the officer was lying. I believe that the officer’s memory was changed by the investigator’s suggestion that the knot did not slip.

Suggestibility also is implicated in false confessions. Certainly many false confessions are the result of outright coercion. But, as Schacter notes, a subset of these false confessions involve the confessors believing that they committed crimes that they actually didn’t commit. In his book, Schacter discusses individual cases involving false confessions and suggestibility.

Schacter further reports on the research regarding the suggestibility of interviews with young children. I understand this research is controversial as child victim advocates sometimes view this research as antithetical to protecting children and getting child molesters off the street. As a prosecutor I was extremely interested in this issue. It’s the reason I bought Schacter’s book, and started requiring recorded interviews by trained forensic interviewers. Anyone who has been involved with interviewing children knows the power of suggestibility—in both directions—from allegations to recanting of allegations. Anyone making decisions about other’s lives based on memories should know the research about suggestibility and be on guard for those factors that may influence memories.

I will continue my discussion of memory a future entry. As I will be involved in another project for the next several weeks, I probably won’t be back for a week or so.



[i] Schacter, Daniel L. (2001) The Seven Sins of Memory, How the Mind Forgets and Remember,N.Y., N.Y. p. 90

[ii] Ibid. p. 113.

The views expressed in this blog are solely the views of the author(s) and do not represent the views of any other public official or organization.

March 14, 2011

The Problem(s) with Memory

Human memory is often central to decisions made in the court system. Witnesses testify based on their memories of their observations of various matters. How confident should one be in a human’s memory? How confident should one be in someone’s confidence of their memories? Let’s say that human memory is far from infallible.

When I was a prosecutor, I handled most of the sexual assault cases. In many of those cases, the only direct evidence against a defendant was the victim’s memory of the assault. Many of the victims were children. I attempted to have the law enforcement agents who investigated these crimes consider the victim’s memory similar to a crime scene. I wanted to preserve the crime scene and prevent contamination of it until all of the evidence had been collected. I had trained interviewers complete videotaped forensic interviews of the victims. I insisted on taped statements from suspects and witnesses.

I also read a book by Daniel L. Schacter, a professor of psychology at Harvard, entitled The Seven Sins of Memory.[i] Although published ten years ago, it still offers much about the research into memory. The book is arranged in chapters around what Dr. Schacter has called the seven sins of memory. I will touch on each of the sins.

Schacter identifies the first sin of memory as transience which is forgetting things due to the passage of time. He writes: “Perhaps the most pervasive of memory’s sins, transience operates silently but continually: the past inexorably recedes with the occurrence of new experiences.” We forget things with time.

The second sin of memory that Schacter identifies is absentmindedness. Absentmindedness is “lapses of attention that results in failing to remember information that was either never encoded properly (if at all) or is available in memory but overlooked at the time we need to retrieve it.” Schacter believes absentmindedness is often the result of divided attention. You are thinking about something else so you are not remembering another thing. Anyone who has set his or her keys or glasses down, only to forget where, has experienced this sin.

Schacter states the third sin of memory is blocking. Blocking occurs when the information that you are attempting to recall has been encoded in memory, but you don’t have the ability to recall it when desired. We all have had the experience of having something on “the tip of our tongue” but can’t seem to come up with it.

These first three sins are mostly passive culprits in reducing the accuracy of the justice system. People forget, and therefore cannot tell us, what we want to know to make an accurate determination of past event. An honest statement from a witness of “I don’t remember” will often be considered as no evidence at all, providing no evidentiary weight in any direction. The sins that I address next week will be more pernicious in our quest for the truth.



[i] Schacter, Daniel L. 2001, The Seven Sins of Memory, How the mind forgets and remembers. Houghton Mifflin Co., N.Y., N.Y.

The views expressed in this blog are solely the views of the author(s) and do not represent the views of any other public official or organization.

March 7, 2011

Competing Ideas or Different Questions?


Two recent Sunday editions of the Milwaukee Journal Sentinel had opinion pieces that addressed the use of evidence-based strategies in addressing crime. The first article was written by Milwaukee County Sheriff, David A. Clarke, Jr.. See Let's treat criminals like...criminals .

It is safe to say that Sheriff Clarke is highly skeptical of using researched-based strategies when making decisions within the criminal justice system. He believes that those who conduct and use such studies are either “academic elites,” “criminal sympathizers”, or “criminal advocates” (or a combination of the three) engaged in an attempt to “indoctrinate the public.” He ultimately believes that the best way to address crime is to reduce the cost of incarceration so that longer sentences are more affordable. The sheriff advocates locking criminals up for as long as possible in cheap, private prisons. The sheriff argues that the value of incarceration in terms of reduced victimization is not being fully considered by those advocating being “smart on crime.”

Milwaukee County District Attorney John Chisholm, and two Milwaukee County Judges, Jeffrey Kremers and Richard Sankovitz, responded to the sheriff in the following Sunday paper. See Rely on the facts to fight crime . They argue that “Decisions about how to correct offenders should be based on research and actual experience, not myth or anecdote or mere intuition.” They further argue that “…we in the criminal justice system are stewards of limited resources; we are responsible for holding offenders accountable and protecting the community—but cost effectively, within what the community can afford.”

The authors of each of the articles struggle with the central issues of crime and punishment in our society. First, we all can agree that crime is an economic drag on our society. In an ideal society, everyone would voluntarily follow the laws of society. In such society, we would no longer have the costs of victimization, the costs of law enforcement and the criminal justice system, and the costs of incarceration. Those societal resources could be used for such things as education and health care, or returned to the citizens to spend in anyway they desire. But we don’t live in this ideal society.

Recently, we have heard much about Wisconsin having to be more competitive with other states. Can we be competitive with states, like Minnesota, that have similar crime rates but spend considerable less taxpayer money on incarceration? Are the people in Minnesota doing something that allows them to control crime at a lower cost than us? Or are Wisconsin people or Wisconsin communities more criminally inclined than Minnesota’s, so Wisconsin requires the expenditure of more resources to fight crime to obtain similar crime rates? If there are ways to provide the same level of community safety at lower costs, or more community safety at the same costs, shouldn’t decision makers explore those avenues?

The use of “evidence-based” strategies when making decisions within the criminal justice system is one such avenue. The sheriff favors long sentences because he correctly understands that incarceration, and therefore incapacitation, for the longest period of time possible is highly effective in reducing the risk that someone will re-offend over the period of time the person is incarcerated. Further, one can say with a high level of certainty that incapacitation will reduce recidivism for most everyone who is incarcerated. For all practical purposes, we don’t really have to be concerned about a population of offenders for which incapacitation doesn’t work. The implication is that the best strategy to maximize community safety is to incarcerate all offenders for the maximum period of time possible. The reduction in the probability that someone will re-offend reduces the expected costs of victimization which is the benefit of incarceration. I believe that is what the sheriff is arguing.

However, the other issue that must be considered is the societal costs of the long periods of incarceration. We don’t live in a society where the costs don’t have to be considered. That is another ideal society that doesn’t exist—yet one we often like to pretend does. When a judge sentences someone to prison for ten years, in effect he or she is deciding to spend approximately $250,000 of taxpayers’ money with the hope that the society will gain at least $250,000 of community safety and other value from the expenditure. If ten people are sentenced to prison for ten years, the society will need to pay $2.5 million for incarceration costs. Sheriff Clarke understands that point, but argues that we should be trying to reduce the cost of incarceration—again a worthy goal.

But what if five of the ten people that we send to prison for ten years would never re-offend again regardless of the sentence? If that is the case, we are spending $1.25 million dollars and getting no additional return to community safety for the expenditure. We are also unnecessarily losing that defendant’s output as a worker and member of a family and community. Wouldn’t it be wise to be able to try to identify the five individuals who wouldn’t reoffend regardless of their punishment, and not spend the $1.25 million to incarcerate them, but spend it somewhere else where it could benefit us? This identification process, of matching defendants with an appropriate sentence, and only using treatment that has been proven effective, is what using “evidence-based” practices is about.

Everyone in the criminal justice system, including law enforcement officers, prosecutors, judges, and correctional professionals are required to make decisions about how best to handle individuals whom they believe have committed a crime. Currently, many of these decisions are based on folk-theory, custom, and erroneous intuitions. The use of empirically supported theories when making decisions has been shown, in fields such as medicine, to be a more effective strategy in obtaining desired goals than using the other, less effective strategies.

Sheriff Clarke was clearly, and rightfully, upset by a convicted bank robber robbing another bank after being released from prison on some type of early release program. He called this program a “failed experiment”. We don’t know if this convict’s release was based on an evidenced-based assessment or not. But does the failure of one convict who participated in a program mean that the program has failed? It doesn’t, but it does exemplify a weakness in any risk-assessment approach—evidence-based or otherwise. Risk assessments can be wrong.

Part of the “evidence-based” approach to crime uses scientifically validated risk-assessments of defendants to identify treatment needs and then match those needs with the sentence. For example, if a defendant has been identified, through the risk assessment, as a good candidate for probation and treatment, then he or she is placed on probation rather than sent to prison. If someone is identified as a continued threat to society, that person is sent to prison. The weakness is that the identification of high and low risk offenders by using a risk assessment instrument is far from accurate.

The research underlying any risk-assessment uses group averages with confidence intervals. The research allows one to say that for a group of individuals, one can be 95% confident that the proportion of individuals who will recidivate will fall within the upper and lower limits of the confidence interval. (See for example, Hart, Stephen D., Christine Mitchie, and David J. Cooke, (2007), “Precision of actuarial risk assessment instruments. Evaluating the ‘margins of error’ of group v. individual predictions of violence” British Journal of Psychiatry, 190 (suppl. 49) pp. s60-65.)

However, when evaluating a single person a different question is asked. For example, if a defendant takes Risk Assessment A and has a score of 10, one can say that we are 95% certain that people with a score of 10, the probability that such a person will recidivate is between an upper and lower limit. The problem is that these limits are quite wide, and much wider than the group confidence intervals. The prediction equation is nothing more than an average value one would observe over many replications, with a high level of risk of being wrong for any one individual.

Therefore, unlike a policy of incarcerating everyone, where we know incapacitation will without a doubt prevent recidivism, a risk assessment will on the average be correct, but will, at times, be wrong. The use of the risk assessment will help identify some of the people who would not recidivate regardless of sentence, but will also erroneously identify some defendants as non-recidivists when in fact they will recidivate and will erroneously identify some defendants as recidivists when they will not recidivate.

Does that make these evidence-based risk assessments useless? Negative. Currently, police officers, prosecutors, judges, probation agents, and correctional professionals make assessments regarding whether or not a particular defendant will re-offend. These assessments drive the decisions on whether to arrest and prosecute, as to what sentence to impose, or whether a defendant shall be released from prison early. The use of “evidence-based” information will help these professionals make better decisions. The research shows that using a risk-assessment instrument increases the accuracy of predictions over predictions based solely on unguided personal judgment. We all know that some of the decisions that will be made will be wrong, as they are now, but these tools, developed and used appropriately (I share the sheriff’s concerns about the system’s ability to separate solid research from hype) will help us become more accurate. The increased accuracy, in turn, will increase community safety, not decrease it, and also preserve taxpayer dollars.

The views expressed in this blog are solely the views of the author(s) and do not represent the views of any other public official or organization.