Reasonable Doubt About Reasonable Doubt Instructions (2/5/21)

I have written below about some of the problems with the venerable “beyond a reasonable doubt” for convictions for crimes.  E.g., Inspired by the Manafort Trial, On the Beyond a Reasonable Doubt Standard (8/15/18; 8/17/18), here (collecting some earlier posts).  See also my earlier pdf book the latest offering being for 2013 after discontinuing because of my chapter in the Saltzman book.   Townsend, John A., Federal Tax Crimes, 2013 (February 5, 2013). Available at SSRN: https://ssrn.com/abstract=2212771.  The key discussion starts on p. 680 [p. 717 of the pdf] beginning here.  I think the cited discussion is still a fair even though now over 7 years old.

Readers interested in the meaning and problems in the criminal standard “beyond a reasonable doubt” might be interested in a law review article I read yesterday.  Michael D. Cicchini, Reasonable Doubt and Relativity, 76 Wash. & Lee L. Rev. 1443 (2019), here.  The key discussion for present purposes starts with the outline  II.B. B. Reasonable Doubt is Not Self-Defining  1455 (pdf p. 14), here.

Key points 

1. After explaining problems in the approaches of not explaining to the jury what reasonable doubt means or attempting some explanation, the author notes from pp. 1455-1456 (pdf pp.. 14-15):  

With so many pitfalls awaiting the trial judge who attempts to define or explain reasonable doubt, many courts have determined  “that the better practice is not to attempt the definition.” Their justification is this: reasonable doubt is already “self-defining,” and, therefore, jurors require no further explanation to understand it. However, this assumption has now been thoroughly tested and debunked.

2. The author presents (pp. 1456-1460 (pdf pp 15-19)) some studies that indicate that jurors really do not reach different outcomes under the reasonable doubt standard as compared to the lesser burdens in civil cases – preponderance of the evidence and clear and convincing evidence.

3. The author discusses the so-called “60/65” rule (pp, 1460-1462 (pdf pp. 19-21), footnotes omitted):

In addition to the above studies showing no significant differences in acquittal rates under the three burdens of proof, researchers have also been testing the impact of reasonable doubt jury instructions in a different way: they seek to determine the subjective confidence level that jurors require before they are willing to convict. “This research has consistently shown that the jurors in criminal cases will often be satisfied with much less certainty than is conventionally assumed.”

When judges and others trained in the law are asked to equate the “proof beyond a reasonable doubt” standard with a corresponding confidence level in the defendant’s guilt, most reply that it should be somewhere near 90 percent. In fact, in a large survey of federal judges that produced 171 respondents, 74 percent (or 126 judges) set the corresponding number at “90 percent or higher.” Similarly, in a real-life example, the Supreme Court of Nevada reversed a conviction because the trial judge, when instructing the jury, had equated proof beyond a reasonable doubt with “anything more than a 75% chance” that the defendant was guilty. On appeal, the court held that setting such a low threshold, coupled with another mistake, “constituted prejudicial error.”

Contrary to the beliefs and desires of most judges, however, jurors equate our highest burden of proof with a much lower confidence level in guilt. In a 1996 study, when asked to quantify proof beyond a reasonable doubt, test participants set an average threshold that ranged from a mere 54 percent to 70 percent, depending on the particular version of the reasonable doubt instruction they received.  Similarly, in a 2007 study, test participants set the conviction threshold at a mere 63 percent chance that the defendant was guilty. And in 2014, researchers  found that laypersons were willing to convict at a 68 percent probability of guilt.

The 2019 study discussed earlier also tested the threshold for conviction and found that 90 percent of mock jurors followed what the researchers called “the 60/65 rule.” That is, regardless of the burden of proof instruction they received, when participants believed that less than 60 percent of the evidence favored the state, they voted to acquit; but when they believed that more than 65 percent of the evidence favored the state, they voted to convict. Or, as another researcher put it, “[r]ather than having to move jurors from 0% to 90% certainty, all prosecutors need do is move the needle on the scale from 50% to perhaps 65% certainty.”

 3.  The author concludes as follows (p. 1494-1495 (pdf pp. 53-54), footnotes omitted):

VII. Conclusion

The Constitution requires the jury to acquit a criminal defendant unless the state proves guilt beyond a reasonable doubt. However, this constitutional protection is only as strong as the trial court’s burden of proof instruction to the jury. Instructing jurors on reasonable doubt is risky business. Many attempts at a definition have created confusion or, worse yet, have lowered or even shifted the burden of proof.However, the solution is not to leave the concept of reasonable doubt unexplained. Empirical research demonstrates that, when left unexplained, the reasonable doubt standard offers defendants no greater protection than the two lower, civil burdens of proof. Further, jurors are willing to convict with only a 60 percent to 65 percent confidence level in the defendant’s guilt—a threshold far lower than what is expected under the proof beyond a reasonable doubt standard.

To adequately convey the state’s high burden of proof, courts should instruct jurors on a relative basis. By explaining proof beyond a reasonable doubt as being higher than the preponderance of evidence standard, and higher even than clear and convincing evidence, jurors will have the necessary reference points to appreciate how high the standard actually is. This relative, context-based approach is supported by a fundamental principle of psychology (contrast effects), the existing jury instructions of some states, sound logical argument, the recommendations of other researchers, and now direct empirical evidence. This Article presented the results of my controlled experiment where mock jurors read the same case summary and were then randomly assigned to two groups, each of which received a different instruction on reasonable doubt. Group A received a simple, undefined instruction and acquitted at the rate of 32.6 percent; Group B received a relative, context-based instruction and acquitted at the higher rate of 42.4 percent. 

This 30 percent increase was statistically significant at p < .05, with an exact p-value of 0.0496. Further, mock jurors that received the relative, context-based instruction required a higher subjective confidence level in the defendant’s guilt before the majority of them were first willing to convict, and the instruction also provided the defendant with more protection throughout multiple participant confidence levels.

Based on these findings, and the findings from other empirical research, this Article presents a comprehensive jury instruction on the presumption of innocence and the burden of proof. The instruction is designed to fulfill the Constitution’s promise that every accused person remains free of conviction unless there is proof of guilt beyond a reasonable doubt.