45 Cornell L. Q. 322, 327 (1960). Professor Zeisel testified that, to ensure a random sample, he used actual jurors called to service at the Daley Center on the days that the surveys were administered. Transcript at 19, lines 10-13. Further, according to the uncontradicted testimony of attorney James Bailinson, the survey respondents were grouped by jury officials in the same manner as those jurors sent to courtrooms for actual jury service. Upon arriving at the Daley Center, the potential juror handed in her summons and pulled from a basket a folded up piece of paper which contained a panel number. Transcript at 1045, lines 9-12 (statement of James Bailinson). Once the juror received her panel number, she waited in Room 1700 of the Daley Center until her panel number was called. Transcript at 1045, lines 16-21 (statement of James Bailinson). Zeisel and his assistants "were given jurors from the assignment room just as the judges would get in batches from there." Transcript at 23, lines 20-22 (statement of Prof. Zeisel). Such a random sampling procedure is not only representative, but appears to be as good a probability sample of potential Cook County jurors that could be designed. See Transcript at 23, lines 23-24 (statement of Prof. Zeisel).
The law of large numbers states that the larger the sample size, the more probable that the sample mean will be close to the population mean. Fredrick J. Gravetter & Larry B. Wallnau, supra, at 202. After removing those jurors who did not meet the requirements of Witherspoon v. Illinois, 391 U.S. 510, 88 S. Ct. 1770, 20 L. Ed. 2d 776 (1968), the April 1990 survey sample consisted of 96 potential jurors, and the January 1992 survey sample included 95. According to Professor Zeisel, any concern that the sample size was too small to be representative is allayed not only by computation of a confidence interval for each question, but also by the results of a split-half reliability analysis. Transcript at 36, lines 10-23. Zeisel essentially divided the samples from each survey into two halves as they arrived in seriatim and treated them as independent samples. As a comparison of the answering patterns for the two half samples of each survey revealed that they were nearly identical,
we can conclude, and respondents have not rebutted, that the sample sizes were sufficiently large to ensure that the samples were representative of the population. See Transcript at 37, lines 5-15 (statement of Prof. Zeisel); see also Transcript at 727, lines 4-5 ("I feel that the sample size is adequate to infer about the population at large.") (statement of Prof. Rossi).
2. Generalization to DuPage County Jurors
Based on demographic differences between the survey samples, the Free jury and the populations of Cook and DuPage Counties as to gender, age, education level and race, Professor Lavrakas opined that the results of the Zeisel surveys cannot be extrapolated to the Free jury, a group falling outside the surveyed population. However, assuming that such demographic differences in fact exist, Free has presented compelling evidence that these variables are irrelevant to jurors' ability to comprehend the jury instructions.
Respecting differences in educational level, Professor Hans testified that the survey data demonstrate "fairly high levels of miscomprehension of the judicial instructions. And the kinds of educational differences that the researcher have discovered I don't think could really influence that conclusion." Transcript at 168, lines 2-6. Finding a uniform pattern of consistent failure to answer the questions correctly and fairly even dispersion across the types of answers with no systematic subgroup scoring better or worse, Professor Rossi concluded that the Zeisel data can be applied to the Free jury with high levels of confidence. Transcript at 645-651. The opinions of Professors Hans and Rossi are bolstered by the results of an independent study conducted by Professor Diamond and Jonathan Casper. The Diamond-Casper study in part tested juror comprehension of the IPI instructions, finding a high rate of incorrect answers to questions concerning aggravating and mitigating factors.
Significantly, the study accounted for differences in scores by college and noncollege graduates, showing that college graduates maintained a high rate of miscomprehension and that the margin of improvement associated with a college education was relatively small. Transcript at 308-310 (testimony of Prof. Diamond).
Likewise, Free presented substantial evidence that gender plays no role in jurors' comprehension of the instructions. See Transcript at 191, lines 6-10 ("So if our sample contained more women than were in Cook County or DuPage County, whatever, I would say, well, here is an instance where that's not going to influence materially the results. Because there isn't that link in advance between gender and comprehension levels.") (statement of Prof. Hans); Transcript at 1125, lines 7-10 ("In terms of their probability of getting an answer correct or incorrect, there is no evidence from these data that would lead you to conclude that there is a difference [between men and women] in getting these questions right.") (statement of Prof. Diamond). Further, based on his tabulation of the survey results by gender, Lavrakas conceded that there is "no consistent pattern in women answering differently than men." Transcript at 953, lines 6-7.
Lavrakas' speculation that age may affect comprehension is unsupported by the evidence adduced at the hearing. Indeed, Lavrakas "did not tabulate the survey data by age and agreed that he had no empirical evidence suggesting that older people would be better able to comprehend the jury instructions than younger people." Report and Recommendation at 16; Transcript at 954-55 (testimony of Prof. Lavrakas). To the contrary, Free's expert witnesses testified that based on the survey data, there is no correlation between age and comprehension of the jury instructions. See, e.g., Transcript at 1080, lines 5-16 ("out of the 17 questions, older people were less accurate on average on eleven of them, and they were more accurate on six of them. Overall I would say . . . there is no difference.") (testimony of Prof. Diamond).
Finally, regarding racial demographics, Professor Diamond discovered that, of white survey respondents, less than fifty percent answered correctly on twelve of seventeen questions, suggesting an extraordinarily high level of misunderstanding on each of the five measured issues. Transcript at 1081-83. As Free was tried by an all white jury, it follows that differences in racial composition are irrelevant to the issue of whether the Zeisel data may properly be applied to the Free jury. In any event, Professor Rossi testified that he could not detect a relationship between race and comprehension. Transcript at 1189-91. Further, Lavrakas "ultimately agreed that race as a variable was not important in determining comprehension level." Report and Recommendation at 17; Transcript at 1225-1227.
We are mindful of the possibility of a "Type II error," that is failing to reject a false hypothesis. However, as explained by Professor Diamond, because a Type II error assumes a difference in effect that goes undetected, we can never know the probability that such an error is being made in concluding that the demographic differences in question are irrelevant to comprehension. Transcript at 1125-26. Professor Diamond continued to state that although it is "hypothetically" possible, "there is nothing in the data to support the notion that a Type II error is being made." Transcript at 1126, lines 5-7. Accordingly, merely raising the possibility of a Type II error is insufficient on the part of respondents to rebut Free's showing that no correlation exists between the various demographic differences and juror comprehension. As such, we agree with Magistrate Judge Weisberg's conclusion that "if the same survey instruments were used with prospective jurors in the DuPage County Courthouse under conditions like those in the Daley Center surveys, it is reasonable to expect similar results showing substantial levels of misunderstanding of the IPI instructions and the capital sentencing instructions used in the Free case." Report and Recommendation at 21.
3. Mundane Realism
In support of their claim that the survey data may not be applied properly to the Free jury, respondents further argue that the Magistrate Judge "ignored Dr. Lavrakas' testimony on 'mundane realism,' that is, the extent to which the Zeisel survey conditions approximated the 'real world.'" Respondents' Objections to Report and Recommendation at 18. We begin with the observation that Professor Lavrakas admittedly is not an expert in juror comprehension, nor has he read any of the literature concerning jury research. See Transcript at 808-823 (voir dire of Prof. Lavrakas). As such, it is ironic that respondents challenge the substance of the testimony of Free's experts solely on the basis of Lavrakas' speculation regarding issues of juror dynamics and comprehension. Nonetheless, Magistrate Judge Weisberg recognized that the Free jury differed from the Zeisel survey respondents in three important ways:
 They made an actual life or death decision,  they deliberated as a group instead of separately answering written questionnaires and  they made their decision after a trial and sentencing hearing in which they heard aggravating and mitigating evidence and arguments by the prosecutor and Free's attorney which discussed how the jury should decide on the sentence.
Report and Recommendation at 21. Accordingly, it is incumbent upon this court to consider the likely impact of these differences in conditions upon juror comprehension despite the lack of any competent testimony offered by respondents on the matter.
The gravamen of respondents' first assertion is that because the Free jurors made an actual life or death decision they were more "motivated" than the survey respondents who were merely taking a pencil and paper test. Indeed, there is no dispute that the Free jurors likely were more "motivated" in determining if death was the most appropriate sentence. In other words, jurors asked to pass ultimate judgment over another's life will no doubt participate in an inner conflict requiring them to balance personal notions of morality and their perceived obligations to the legal system. Such emotional turmoil clearly was not present in the Zeisel studies. As pointed out by Professor Zeisel, however, whether the jurors reached the "correct" result as a matter of personal morality and whether the jurors comprehended and followed the judicial instructions are two different questions, Transcript at 22-23 -- the relevant issue for our purposes, and the issue designed to be studied by the Zeisel surveys, being the latter.
Acknowledging the difference in emotional state between the Free jury and the survey respondents, Free's experts testified that empirical research demonstrates that intensity and emotions eclipse the cognitive abilities necessary for comprehension of such complex instructions. See, e.g., Transcript at 441-42 ("One of the difficulties is that when people are emotionally overwrought -- and I guess what you were trying to say is that they are heavily emotionally involved -- it sometimes cuts against comprehension. . . . high emotional level interferes with comprehension.") (testimony of Prof. Diamond); see also Loftus, Eyewitness Testimony (1979) (emotions and intensity impair cognitive functions) (cited below by Free's experts). Further, Professor Diamond testified that the conditions afforded the survey respondents provided a superior environment for comprehension than did the conditions in Free, particularly because the survey respondents not only had the instructions read to them, but also had a written copy for reference while answering the survey questions. Transcript at 1146-47. In light the uncontradicted evidence, we concur in Magistrate Judge Weisberg's conclusion that such factors as "motivation" and "emotion" do not enhance juror comprehension, and that the survey respondents did not participate in an actual life or death decision does not undermine the results of the Zeisel surveys as applied to the Free jury.
Respecting the effect of deliberation and closing arguments upon juror comprehension, Magistrate Judge Weisberg found that "the evidence is not strong enough to rule out the possibility that comprehension among members of the Free jury on some key issues may have been substantially aided by hearing the evidence, the arguments of the attorneys and by the process of deliberation." Report and Recommendation at 24. After a careful review of the testimony of Professors Zeisel, Transcript at 37-41, 91, Hans, Transcript at 156-166, 221-22, and Diamond, Transcript at 306-07, we agree that the effect of these factors can only be assessed in conjunction with an analysis of the issues on which a likelihood of misunderstanding is claimed. Accordingly, we will revisit these factors infra Section III of this opinion.
Respondents maintain that a comparison of the 1992 survey results with the confidence intervals predicted on the basis of the 1990 survey demonstrates a mark of inconsistency or "volatility." We disagree.
More than one confidence interval was calculated on the basis of the 1990 survey. Professor Zeisel calculated 95% confidence intervals based on percentages of incorrect answers, while Professor Rossi calculated 99% confidence intervals based on percentages of correct answers. Each eliminated those questions for which no correct answer existed (questions 1, 13, 14, 15 and 16), as well as question 6, which was not included in the 1990 results because of a typographical error. Comparing the 1992 results to the Zeisel intervals, "the 1992 results fell within the predicted confidence intervals for 3 questions, fell slightly outside (1% to 2%) for 2 questions, outside by 4% to 6% for 3 questions and outside by 9% to 12% for 2 questions." Report and Recommendation at 17-18. "Using Rossi's confidence intervals, the 1992 results fell within predicted confidence intervals for 3 questions, slightly outside for 1 question, outside by 5% to 8% for 5 questions, and outside by 12% for 1 question." Id. at 18.
As an empirical matter, we agree with Magistrate Judge Weisberg and Professor Rossi that the levels of deviation in the 1992 survey from the results predicted on the basis of the 1990 survey (i.e., the confidence intervals) are "modest." Report and Recommendation at 18; Transcript at 763, lines 13-14 (testimony of Prof. Rossi). According to Professor Rossi, that the two surveys were not identical accounts for this modest deviation. Transcript at 770-73. Indeed, the predictive quality of the confidence interval rests on the assumption of identical surveys being given under identical conditions. In any event, the fact that both surveys, with few exceptions, depict very high levels of misunderstanding buttresses the Magistrate Judge's conclusion that the surveys are valid and reliable.
III. The Impact of the Zeisel Survey Results on Free
In determining the impact of the Zeisel data on Grounds 5, 10 and 14 of Free's petition for habeas relief, this court is guided by the holding in Boyde v. California, 494 U.S. 370, 380, 110 S. Ct. 1190, 1198, 108 L. Ed. 2d 316 (1990), that when presented with a claim that a capital sentencing instruction is ambiguous and therefore subject to an erroneous interpretation, the relevant inquiry is
whether there is a reasonable likelihood that the jury has applied the challenged instruction in a way that prevents the consideration of constitutionally relevant evidence. Although a defendant need not establish that the jury was more likely than not to have been impermissibly inhibited by the instruction, a capital sentencing proceeding is not inconsistent with the Eighth Amendment if there is only a possibility of such inhibition.
With this standard in mind, we turn to consider the conclusions to be drawn from the Zeisel survey results respecting each of the measured aspects of juror comprehension.
A. Existence of a Mitigating Factor
Designed to test whether jurors believe they must agree unanimously on the existence of a mitigating factor before that factor can be considered by an individual juror, questions 4 and 5 of the Zeisel surveys provide as follows:
4. A juror decides that the fact that Mr. Wood was only 25 years of age when he committed the murder is a mitigating factor sufficient to preclude the death penalty. However the other eleven jurors disagree and insist that his age is not a mitigating factor. The one juror believes that she cannot consider a mitigating factor unless the entire jury agrees upon it and votes for the death penalty. She votes for the death penalty.
5. A juror decides that the fact that Mr. Woods was good to his family is a mitigating factor sufficient to preclude the death penalty. However, the other eleven jurors disagree. The other jurors insist that no juror should consider the defendant's good relations with his family as a mitigating factor unless they all agree it is a mitigating factor. The one juror accepts this approach and votes for the death penalty.