IN RE: Death Penalty Disparity Claims (2013)

MEMORANDUM OF DECISION

In 2003, the Connecticut Supreme Court decided the appeal in the case of State v. Reynolds, 264 Conn. 1, in which the defendant attempted to raise the issue of whether the death penalty which was imposed upon him was produced by a legal system which violated the state constitutional ban on discriminatory punishment.   He contended that death sentences in Connecticut were influenced by arbitrary factors such as race.  Id., 226–28.   He argued that he was entitled to an evidentiary hearing, before sentencing, in order to establish his claim that the capital punishment process in Connecticut was “illegal, arbitrary, discriminatory, disproportionate, wanton, and freakish.”  Id.

Although it rejected the notion that such a hearing needed to precede sentencing, our Supreme Court announced that inmates sentenced to death had the right to such a hearing through a consolidated habeas corpus proceeding.   Id., 232–34.   Sedrick Cobb, Daniel Webb, Richard Reynolds, Robert Breton, and Todd Rizzo have availed themselves of this procedure.   On several dates from September to December 2012, this court heard the evidence relating to this consolidated habeas action.

On April 2, 2010, the petitioners filed their revised claims regarding the unconstitutionality of their death sentences based on 1. the “wanton, freakish, arbitrary and capricious manner” in which the death penalty was imposed;  and 2. that the “capital punishment procedures are impermissibly infected by racial, geographic, and/or gender bias and disparities.”   More specifically, the revised petition asserts that the death penalty system in Connecticut denies the petitioners their rights to due process and equal protection under the law and amounts to cruel and unusual punishment in contravention of Article I, §§ 7, 8, 9, and 20 of the Constitution of the State of Connecticut;  the Fifth, Sixth, Eighth, and Fourteenth Amendments to the federal constitution;  Articles 1, 2, 3, 5, and 7 of the Universal Declaration of Human Rights;  Articles 1, 2, 17, 18, 25, and 26 of the American Declaration on the Rights and Duties of Man;  Articles 6(1), 7, 10(1), 14(1), and 26 of the International Covenant on Civil and Political Rights;  and Articles 1, 2, 3, 4, 5, and 6 of the International Convention on the Elimination of All Forms of Racial Discrimination.   A post-trial amendment, allowed by the court, also sets forth a claimed violation of General Statutes § 53a–46b(b)(1).

I

As to the allegations that the Connecticut capital punishment scheme generates violations of various international treaties or compacts, such documents have no legal effect in this case.   Federal courts have overwhelmingly spurned the proposition that these international declarations, to some of which the United States is a signatory, impose any legal obligation on the domestic courts of this country.   In Sosa v. Alvarez–Machain, 542 U.S. 692 (2004), the United States Supreme Court held that the Universal Declaration of Human Rights and the International Covenant on Civil and Political Rights are merely statements of moral ideals with “little utility” in criminal cases arising under the laws of the United States.  Id., 734–35.   In Garza v. Lappin, 253 F.3d 918 (7th Cir.2001), a Court of Appeals held that the American Declaration of the Rights and Duties of Man “is an aspirational document which ․ did not create any enforceable obligations on the part of any of the O.A.S. member nations.”   Id., 925.   The D.C. Court of Appeals affirmed a federal district court decision which went further and ruled that domestic courts lacked subject matter jurisdiction to consider claims made under the International Convention on the Elimination of All Forms of Racial Discrimination, Johnson v. Quander, 370 F.Sup.2d 79, 101 (D.D.C.2005), affirmed 440 F.3d 489 (D.C.Cir.2006), cert. denied, 549 U.S. 945 (2006).

The trial courts of Connecticut have followed these federal precedents and held that these international statements concerning human rights posit no adjudicative obligations on state tribunals.   Judge Skolnick ruled that the Universal Declaration of Human Rights created no private right of enforcement in federal or state courts, Belic v. Amtrak, Superior Court, New Haven J.D., d.n. CV–01–275697 (March 11, 2002) [31 Conn. L. Rptr. 548].   More pointedly, Judge Schuman recently wrote that the “various international human rights documents ․ do not create enforceable rights in American courts” when resolving habeas corpus cases filed by two of the petitioners in the present habeas action, Breton v. Warden, Superior Court, Tolland J.D., d.n. CV–034261 (August 15, 2011);  and Reynolds v. Warden, Superior Court, Tolland J.D., d.n. CV–044417 (June 28, 2011).

In Moore v. Ganim, 233 Conn. 557 (1995), our Supreme Court refused “to read into the [state] constitution unenumerated affirmative governmental obligations” based on international statements of contemporary societal norms.  Id., 594–96.   That case involved an attack on the constitutionality of time limits on financial aid to indigent persons.  Id.

This court, therefore, holds that if the petitioners are to succeed in their claims of the illegal imposition of the death penalty, they must do so under interpretation of domestic constitutional and statutory provisions rather than relying on any international statements of human rights.   See, also, State v. Webb, 238 Conn. 389, 407 (1996), citing the case of Moore v. Ganim, supra.

II

The court now addresses the alleged violations of specific state and federal constitutional and statutory provisions.

A.

In Furman v. Georgia, the United States Supreme Court issued a one-paragraph, per curiam decision overturning “the imposition and carrying out of the death penalty” pertaining to three cases arising from the state courts of Georgia and Texas.   The basis for these reversals was a violation of the Eighth and Fourteenth Amendment prohibition of cruel and unusual punishment, 408 U.S. 238, 239–40 (1972).   This terse and, consequently, enigmatic decision required further explication by that high court.

Some state legislatures interpreted the constitutional flaw condemned in Furman to be a lack of uniformity in the imposition of the death penalty.   Those states responded to the Furman decision by enacting mandatory death sentences for crimes which entailed a specified set of circumstances.   See, Roberts v. Louisiana, 428 U.S. 325, 331 (1976).   In that appeal, the United States Supreme Court clarified that the evil which was to be eradicated under Furman was not the existence of discretion to be exercised by sentencing bodies, but it was the absence of well-defined guidelines which would focus the use of that discretion.  Id., 333–35.

Other legal authorities thought that Furman signaled the end of capital punishment, but the high court dispelled that notion in the appeal of Gregg v. Georgia, 428 U.S. 153 (1976), which was decided on the same day as Roberts v. Louisiana, supra.   Georgia adopted a detailed, statutory rubric designed to provide prosecutors, judges, and juries with specific instructions to follow when considering whether a sentence to death was warranted.

An inmate condemned to death under that new regimen contended that any legal procedure which invested the decision-makers, such as prosecutors, juries, judges, and other state officers, with discretion to impose a death sentence or not was intrinsically flawed and prone to generate arbitrary and capricious dispositions.  Id., 198–99.   The Supreme Court repudiated that reading of Furman and held that the “existence of these discretionary stages is not determinative.”  Id., 199.   The Court elucidated further that “Furman held only that, in order to minimize the risk that the death penalty would be imposed on a capriciously selected group of offenders, the decision to impose it had to be guided by standards so that the sentencing authority would focus on the particularized circumstances of the crime and the defendant.”  Id.

Because the new sentencing procedures directed the sentencer's “attention to the nature or circumstances of the crime committed or to the character or record of the defendant,” no longer could the sentencing entity “wantonly and freakishly impose the death sentence;  it is always circumscribed by the legislative guidelines.”  Id., 206.   Therefore, the Georgia death penalty provisions passed constitutional muster.  Id., 207.

After the Furman decision, the Connecticut legislature enacted General Statutes §§ 53a–46a through 53a–46d, which statutory scheme facially complied with the mandates of the federal constitution and did not breach the ban on cruel and unusual punishment required by due process, State v. Peeler, 271 Conn. 338, 447–48 (2004);  State v. Reynolds, 264 Conn. 1, 62–63 (2003);  State v. Webb, 238 Conn. 389, 403 (1996);  State v. Ross, 230 Conn. 183, 235–44 (1994).   Furthermore, the parallel provisions of the Connecticut constitution “incorporate the principles underlying a constitutionally permissible death penalty statute that the United States Supreme Court has articulated,” State v. Ross, supra, 252.   That is, the rights afforded capital murder defendants by our state constitution are analytically identical to those protections prescribed under the federal constitution, State v. Santiago, 305 Conn. 101, 252–53 (2012);  State v. Reynolds, supra, 236 (2003);  State v. Griffin, 251 Conn. 671, 691–709 (1999).

B.

The United States Supreme Court has also established the constitutional boundaries for the death penalty as applied with respect to the equal protection clause of the Fourteenth Amendment as well as the incorporation of the ban on cruel and unusual punishment contained in the Eighth Amendment.   In the appeal of McCleskey v. Kemp, 481 U.S. 279 (1987), the high court confronted “the question of whether a complex statistical study that indicates a risk that racial considerations enter into capital sentencing determinations proves that petitioner McCleskey's capital sentence is unconstitutional under the Eighth or Fourteenth Amendments.”  Id., 282–83.   The Supreme Court rejected all aspects of this proposition.

McCleskey argued that a statistical study, embracing consideration of over 2,000 murder cases in Georgia, which suggested “that defendants charged with killing white persons received the death penalty in eleven per cent of the cases, but defendants charged with killing blacks received the death penalty in only one per cent of the cases.”  Id., 286.   That “study indicates that black defendants, such as McCleskey, who kill white victims have the greatest likelihood of receiving the death penalty.”  Id., 287.   McCleskey attacked every stage of the legal system including the actions of a state legislature in permitting a death sentence, prosecutorial decision-making, and jury deliberations which result in a sentence to death.  Id., 292.

The United States Supreme Court instructed that its “analysis begins with the basic principle that a defendant who alleges an equal protection violation has the burden of proving ‘the existence of purposeful discrimination,’ “ Id., 292 (emphasis added).  “A corollary to this principle is that a criminal defendant must prove that the purposeful discrimination ‘had a discriminatory effect’ on him.”  Id. “[T]o prevail under the Equal Protection Clause, [the defendant] must prove that the decision makers in his case acted with discriminatory purpose.”  Id. (emphasis in original).

McCleskey submitted that the multiple regression analysis which informed the statistical study in his case was sufficient to demonstrate that he was denied equal protection based on his race and that of the victim.   The Supreme Court disagreed.  Id., 293–94.   That Court acknowledged that it had “accepted statistics as proof of intent to discriminate in certain limited contexts,” such as in the summoning of venire persons and in employment discrimination cases.  Id. But the Supreme Court found that such statistically-based evidence was of little utility in an equal protection challenge to imposition of the death penalty.  Id., 294.  “[T]he nature of the capital sentencing decision, and the relationship of the statistics to that decision, are fundamentally different from the corresponding elements in the venire-selection or [employment] cases.”  Id. (emphasis added).  “Thus, the application of an inference drawn from the general statistics to a specific decision in a trial and sentencing simply is not comparable to the application of an inference drawn from general statistics to a specific venire-selection or [employment] case.”  Id., 294–95.

In that decision, the United States Supreme Court also observed that “policy considerations behind a prosecutor's traditionally ‘wide discretion’ suggest the impropriety of our requiring prosecutors to defend their decisions to seek death penalties.”  Id., 296 (emphasis added).   Far stronger proof of purposeful discrimination than that produced by a general statistical study is necessary to impugn a prosecutor's charging judgment, especially because the defendant was convicted of the murder “for which the United States Constitution and Georgia's laws permit imposition of the death penalty.”  Id., 297.   That is, the outcome itself, that a factfinder later imposed the death penalty, vindicates, to a significant extent, the earlier charging decision by the prosecutor to seek the death penalty for that individual defendant.  Id.

The United States Supreme Court further determined that “discretion is essential to the criminal justice process,” and “[t]he unique nature of the decision at issue [in death penalty cases] counsels against adopting such an inference [of purposeful discrimination] from the disparities indicated” by statistical studies.  Id., 297.   Absent “exceptionally clear proof,” such statistical analyses fail to prove that the decisionmaker “acted with discriminatory purpose.”  Id.

McCleskey also attempted to convince the high court that a state actor's inaction in the face of knowledge of disparate outcomes based on the race of the alleged victim and perpetrator violates the Equal Protection Clause of the Fourteenth Amendment.   The Supreme Court resoundingly rejected that argument.   Id., 298.  “ ‘Discriminatory purpose’ implies more than intent as volition or intent as awareness of consequences.   It implies that the decisonmaker ․ selected or reaffirmed a particular course of action at least in part ‘because of,’ not merely ‘in spite of’ its adverse effects upon an identifiable group.”  Id., 298 (emphasis added).   That means that the prosecutor and jury must have arrived at charging or sentencing decisions “to further a racially discriminatory purpose.”  Id.

McCleskey also averred that, even if Georgia's death penalty procedure was facially correct, the generic statistical study pointed to racial considerations influencing imposition of the death penalty, resulting in an “arbitrary and capricious” application of that statutory procedure in violation of the Eighth Amendment proscription against cruel and unusual punishment and the Fourteenth Amendment's grant of equal treatment under the law.  Id., 308.   The Supreme Court also found fault with this argument.  Id., 308–20.

The United States Supreme Court demarcated the line distinguishing the impermissibly freakish imposition of death sentences, as identified by the Furman decision, from the situation of statistically disparate outcomes based on race or other irrelevant characteristics which are inherent in any judicial system which confers discretion upon decision-makers.  “There is, of course, some risk of racial prejudice influencing a jury's decision.”  Id., 308.   “The question is at what point that risk becomes constitutionally unacceptable.”  Id., 308–09.

In grappling with this issue, the United States Supreme Court noted that it is an impartial jury selection system which itself provides the safeguards against the “unacceptable risk.”  Id., 310.   The very discretion which McCleskey assailed because of its capacity to produce racially disparate outcomes was regarded, instead, by the Supreme Court as the primary bulwark against invidious discrimination.  Id., 311–12.

Statistical evidence which merely shows “only a likelihood that a particular factor entered into” the capital sentencing process fails to establish a denial of equal protection.  Id., 308–09.   The person facing execution must prove that racial animus effectuated the sentence to death in that individual's case in particular.  Id. “McCleskey asks us to accept the likelihood allegedly shown by the [statistical study] as the constitutional measure of an unacceptable risk of racial prejudice influencing capital sentencing decisions.   This we decline to do.”  Id., 309.

Prosecutors are vested with the authority, and responsibility, to exercise leniency and decline to seek the death penalty or to press for its imposition in an individual case.   That authority is an essential part of every judicial system, and the extirpation of such discretion “would be totally alien to our notion of criminal justice.”  Id., 312.   Thus, the “power to discriminate” is insufficient to render a death penalty procedure constitutionally infirm.   Id. The constitutional rights to equal protection and against infliction of cruel and unusual punishment are breached only if racial disparity was the purpose and goal of the actions taken by the prosecutor when charging a defendant or the jury or judge in rendering a death sentence.  Id., 298.

Statistical investigations which disclose “apparent disparities in sentencing are an inevitable part of our criminal justice system.”  Id., 312.  “No perfect procedure” exists which will guarantee the eradication of racial disparity generally.  Id. Analyses of data which show a correlation of the imposition of the death penalty and racial considerations are insufficient to demonstrate invidious discrimination rather than the inevitable disparities which arise in any pluralistic society.  Id., 313.   That death penalty decisions may be influenced, in some manner, by race is an important, and perhaps compelling, factor when legislative bodies contemplate and debate if, and under what circumstances, the death penalty will be part of a jurisdiction's sentencing spectrum.  Id., 319.   However, “[i]t is not the responsibility—or indeed even the right—of [the United States Supreme Court] to determine the appropriate punishment for particular crimes.”  Id. “Unexplained” disparities alone cannot be assumed to be the product of purposeful discrimination.  Id., 313.

“[T]he ultimate duty of courts [is] to determine whether [such] laws are applied consistently with the Constitution.”  Id., 319.   Constitutional rights to equal protection and to be free from cruel and unusual punishment are abrogated only when an individual death sentence is applied with the intent to discriminate based on racial factors.  Id., 313.   The Supreme Court, therefore, affirmed the lower federal court's denial of McCleskey's challenge to his death sentence founded on a statistical study from which one can merely infer that race influences capital sentencing generally.  Id., 319–20.

The Second Circuit Court of Appeals cited the McCleskey principle in ruling in a case in which the plaintiffs alleged, inter alia, a denial of equal protection by the conduct of the Connecticut Department of Public Health, Knight v. Connecticut Department of Public Health, 275 F.3d 156 (2d Cir.2001).   The Court of Appeals observed that “[t]o show an equal protection violation, appellants must show they were selectively treated compared with other similarly situated employees, and that selective treatment ‘was based on impermissible considerations such as race, or religion,’ “ Id., 166.   “Appellants must also prove ‘that the decision-makers in their cases acted with discriminatory purpose,’ “ Id., citing McCleskey v. Kemp, supra, (emphasis added).

In another Second Circuit case arising out of Connecticut, the Court of Appeals stated, “It is well established that a claimant under the Fourteenth Amendment must establish intentional discrimination.   See McCleskey v. Kemp ․” Ricketts v. Hartford, 74 F.3d 1397, 1407 (2d Cir.1996), (emphasis added).   More recently, that same Court of Appeals reiterated that the McCleskey holding requires proof of “intentional discrimination,” Hayden v. Paterson, 594 F.3d 150, 165 (2d Cir.2010).   Clearly, then, the standard for evaluating the utility of statistical evidence in the context of claims of violations of the federal constitution must entail an assessment of whether that evidence points to purposeful discrimination rather than simply reflects an association between race, gender, etc. and outcomes.

C.

The court now explores the application of the McCleskey principles by the courts of Connecticut.   In Abdullah v. Commissioner, 123 Conn.App. 197 (2010) cert. denied, 298 Conn. 930 (2010), the habeas petitioner attacked Connecticut's charging and sentencing procedure by alleging that black defendants in criminal cases who elected trial received “disproportionately longer sentences” than white defendants who made the same election.  Id., 198.   The petitioner argued that this disproportionality offended the equal protection provisions of both the state and federal constitutions.  Id., 198–99.   The petitioner's evidence, purporting to validate his claims, came in the form of statistical analysis and data reports.  Id., 200.   The habeas court dismissed the petition for failure to state a claim for which habeas relief was available.  Id., 201.

Our Appellate Court relied heavily on McCleskey v. Kemp, supra, in affirming that dismissal.  Id., 202–05.   That Court declared that no viable constitutional challenge on equal protection grounds can rest on disparate outcome alone with respect to race.  Id. The habeas claim failed because the petitioner's statistical evidence, and his allegations based on disparate outcomes, failed to assert or prove that persons acting on behalf of the state “ha[d] displayed purpose or intent to create or to maintain” race-based inequities.  Id.

Citing McCleskey, our Appellate Court ruled that the petitioner failed to offer “specific evidence that would support an inference that racial bias played a role in his sentence.”  Id., 203 (emphasis added).   The Court held that McCleskey required an inmate to prove that his or her sentence was the result of purposeful racial discrimination.  Id., 203–04.   The Court rejected the petitioner's suggestion that the McCleskey holding was an outlier unworthy of precedential value.  Id., 205.

Further, the Appellate Court recognized that, “as a general matter, this state's constitutional equal protection jurisprudence follows that of the federal constitution.”  Id., 206.   In footnote five, the Court adopted the United States Supreme Court's “disavowal” of statistical analysis as sufficient proof of intentional discrimination in sentencing.  Id., 204.   In footnote seven, the Court reiterated its understanding that state constitutional equal protection challenges cannot be premised on “disparate impact” alone.  Id., 209.

Indeed, the unique procedure employed in the present case grew out of our Supreme Court's interpretation of the holding of McCleskey v. Kemp, supra.   See In Re Application for Petition for Writ of Habeas Corpus, 272 Conn. 676, 681 (2005);  State v. Breton, 264 Conn. 327, 399 (2003);  State v. Cobb, 251 Conn. 285, 504 (1991);  State v. Cobb, 234 Conn. 735, 748–49 and 761–62 (1995).   This court concludes that the state constitutional provisions bearing on equal protection, due process, and cruel and unusual punishment confer no greater nor lesser protection than their federal counterparts.   Consequently, the appropriate scrutiny of the petitioners' claims in this case is identical under either document and the standards set forth in McCleskey v. Kemp, supra, control.   See State v. Santiago, supra, 252–53;  State v. Webb, supra, 146–47.

D.

1.

In their post-trial brief, the petitioners acknowledge that our state appellate tribunals have explicitly adopted and applied the federal constitutional concepts concerning cruel and unusual punishment, equal protection, and due process when interpreting comparable state constitution provisions.   The petitioners politely, and perhaps sheepishly, suggest that this court treat these decisions as nugatory because they were wrongly decided.   They candidly recognize, however, that abandonment by an inferior court of prevailing precedent issued by a higher authority is usually ill-advised and legally improper.   This court perceives no imminent seachange in this area of the law or hints by the appellate courts of the state that such a judicial evolution is in the offing.   Therefore, this court declines the petitioners' invitation to disregard pertinent and binding precedents in the interpretation of our state constitution.

2.

The petitioners have amended their petition to assert a statutory approach to avoid the need to prove that purposeful discrimination occurred in each of the petitioners' cases individually.   They vigorously assert that the statutory review of a death sentence mandated by General Statutes § 53a–46b(b)(1) alters the contours of what comprises cruel and unusual punishment and what evidence will suffice to overturn that death sentence.   This court holds, however, that § 53a–46b(b)(1) also requires proof of intentional discrimination before a death sentence will be invalidated.

Subsection 53a–46b(b) states, in salient part:

[T]he Supreme Court shall affirm the sentence of death unless it determines that:  (1) the sentence was the product of passion, prejudice or any other arbitrary factor;  ․

Until 1995, this statute also included a proportionality review clause.   The removal of that review clause has left the text and meaning of the present version of § 53a–46b somewhat perplexing.   Subsection (a) of § 53a–46b requires death sentence review by our Supreme Court “pursuant to its rules,” (emphasis added).   That subsection expressly confers upon the Supreme Court the power and duty to examine and affirm or vacate each sentence of death “[i]n addition to its authority to correct errors at trial.”   By comparison, subsection (b) obligates the Supreme Court to affirm a death sentence “unless it determines” that the penalty was produced by passion, prejudice, or any other arbitrary factor or “(2) the evidence fails to support the finding of an aggravating factor ․” (emphasis added).

Insufficiency of proof of an aggravating factor would appear to be a legal error occurring at trial and for which the Supreme Court already has the power to adjudicate and rectify.   Is the phrase “in addition to” merely a superfluous remnant, unintentionally left over because of a poorly reconciled repeal process?   Did our legislature intend to allow the Supreme Court to substitute its own factual conclusions for those of the factfinder at trial with respect to the quality of the evidence pertinent to the existence of an aggravating factor or to opine that the sentence was the product of passion, prejudice, or other improper factor?   This is a usurpation which the Supreme Court vigilantly eschews in nearly all other situations.   What are the Supreme Court “rules” to which the subsection refers?   Are they the existing rules dealing with appeals generally or was the Supreme Court supposed to create a separate set of rules to implement § 53a–46b.

Subsection (c) of § 53a–46b provides:  “The sentence review shall be in addition to direct appeal and, if an appeal is taken, the review and appeal shall be consolidated for consideration.   The court shall then render its decision on the legal errors claimed and the validity of the sentence” (emphasis added).   How can a death sentence be legally correct and invalid simultaneously?   Can a death sentence acquire invalidity from some source besides legal error?   It is curious that the legislature employed the word “valid” rather than “proportional” or “appropriate.”   Our Supreme Court has sagely remarked that “the scope and meaning of § 53a–46b(b) have remained uncertain,” State v. Cobb, 234 Conn. 735, 763 (1995).

Fortunately, some aspects of § 53a–46b(b)(1) have been clarified by our Supreme Court.   When subsection (3) of § 53a–46b(b) was still in effect, that Court held that subsection (1) fails to create “a nonwaivable right to a review of any and all claims implicating the arbitrariness of a death sentence ․” In Re Application for Petition for Writ of Habeas Corpus, 272 Conn. 676, 685 (2005).   Waiver was available to Michael Ross even though other death row inmates wanted to file habeas claims governing the Ross case in order to address “systemic racial bias.”  Id., 689–90.   Because no claim was asserted by Ross that race-based disparity produced his death sentence, § 53a–46b review of that type of claim was unnecessary and waivable.  Id.

The notion that sentence review under § 53a–46b(b) is confined to an individualized consideration of whether racial bias or other improper factors played a role in producing a particular death sentence was also evinced in the decision of State v. Cobb, 234 Conn. 735, 761 (1995).  “[H]e could have proceeded pursuant to § 53a–46b(b)(1) claiming that the statistical analysis he seeks to establish demonstrates that his death sentence was the product of ‘prejudice or any other arbitrary factor’ “ Id. (emphasis added).

Our Supreme Court explicitly rejected the idea that § 53a–46b(b) permits a universal attack on death sentences, collectively, based on a statistical analysis of data proffered under this statutory provision, State v. Breton, 264 Conn. 444 (2003).  “We note that statutory proportionality review pursuant to § 53a–46b does not contemplate review of this type of claim.”  Id. Generic allegations that imposition of the death penalty in Connecticut is inherently flawed “is not incorporated in § 53a–46b, but is simply another term for analysis under the eighth amendment's prohibition against cruel and unusual punishment or under our state constitutional counterpart.”  Id. (emphasis added).

Also, § 53a–46b(b)(1) and former subsection (3) differ in another important respect.   Repealed subsection (3) necessitated a comparison of “trial court records of similar cases,” In Re Application for Petition for Writ of Habeas Corpus, 272 Conn. 676, 682 (2005).   In contrast, subsection (1) is confined to evaluation by [the Supreme Court] of the trial court record of “the case on appeal.”  Id. (emphasis added).

This court concludes that § 53a–46b(b) directs that an individualized inquiry for a particular death sentence be conducted by the Supreme Court rather than reliance solely on a global review of the death penalty in general.   It is, of course, true that statistical evidence of systemic racial or other improper bias can be probative evidence toward the evaluation required by 53a–46b(b)(1), and the McCleskey standard for that matter.   However, such generic evidence, by itself, is insufficient to overturn a death sentence.   Instead, each petitioner must demonstrate the specific way in which such systemic disparity contributed to an inference of purposeful discrimination in that petitioner's prosecution and sentencing.

This consolidated habeas corpus action can efficiently accommodate the statistical evidence regarding global disparities with respect to certain criteria, but the court discerns no intent by our Supreme Court to modify the judicial scrutiny set forth by the McCleskey case, supra, which at its core demands proof of personal impact produced by intentional discrimination.   As recounted above, § 53a–46b(b)(1) requires the same kind of evidence as needed for a resolution of constitutional violations, State v. Cobb, 234 Conn. 735, 762 (1995);  State v. Breton, 264 Conn. 327, 444 (2003).

E.

1.

With these legal principles in mind, the court now discusses the statistical evidence submitted by the petitioners.

In 1998, the Office of the Chief Public Defender (OCPD) sponsored the collection of raw data pertaining to death penalty cases.   Around that time, Professor Neil Weiner, now deceased, was conducting similar studies for the state of New Jersey and the city of Philadelphia.   The OCPD retained Professor Weiner for the purpose of assembling an appropriate and accurate list of cases from which pertinent death penalty data might be extracted.

Professor Weiner's team collected all cases in Connecticut in which a defendant was chargeable with capital murder from 1973 on.   The source material for potential inclusion of a case was appellate reports;  Judicial District clerical records;  Department of Corrections sentencing lists;  medical examiner reports;  public defender office reports;  and FBI homicide victim information records.   The preliminary assemblage of cases was screened to exclude those cases where no death penalty was legally feasible, e.g. because of the lack of any aggravating factor.   Also, acquittals were removed from consideration.

Over a four-year period, 2,500 cases arising from 1973 to 1998 were filtered in this manner, and 105 cases were identified as satisfying the criterion of potential imposition of the death penalty.   For each selected case the team created a single data file which contained police reports, court transcripts, presentence investigation reports, pleadings, etc.   Then, a Data Collection Instrument (DCI) was prepared for the file from that material.   The DCI had essentially a questionnaire format which called for numerically valued responses.   In 2001, the chosen cases were independently reviewed by two lawyers.   Whenever the lawyers concurred on the result, that result was coded into a computer-readable document.   By 2003, Dr. Weiner authored a report.   In that report, Dr. Weiner indicated that the 105 cases supplied too little data from which to draw any meaningful statistical inferences concerning the effects of race or other impermissible factors in death penalty decision making.

Following Dr. Weiner's departure, the OCPD hired Professor John J. Donahue, III, to oversee and complete the statistical study.   Mr. Donahue is a law professor at Stanford Law School, where he teaches courses in law and statistics, torts, and the legal and empirical aspects of the death penalty.   He has also taught at Yale Law School and other institutions.   He graduated, with honors, from Harvard Law School and acquired both masters and doctorate degrees in economics at Yale University.   The court will henceforth refer to him as Dr. Donahue.

He is highly trained, experienced, and skilled in the field of econometrics as well as the law.   Econometrics applies mathematical and, particularly, statistical methods to economic or other social phenomena.   Econometric approaches are essential if one wishes to test economic or social theories against real-world data.   Econometric techniques are often used to evaluate claims of workplace discrimination or to assess whether civil rights legislation programs actually produce the desired economic results.   The power of econometrics is the ability to devise methods to discern general patterns or trends from masses of empirical data.   It can provide the existential foundation upon which policy makers can attempt to construct effective solutions to complex economic and social problems.

Subsequently, a similar process was used to screen an additional 1,500 homicide cases which were adjudicated between 1998 and 2007.   The team culled from those cases a second batch of 100 death eligible candidates for study.   Combining the 105 original selections with these new cases produced a total of 205 DCIs. From this set of DCIs, a second, “scrubbed” set of DCIs was prepared.   A “scrubbed” DCI is one from which all racial information was deleted.

Of the 205 cases studied, only twenty-eight reached a penalty phase which considered imposition of a death sentence.   Of those twenty-eight cases in which a penalty phase was litigated, twelve murderers were sentenced to death.   Three of the death sentences were reversed on direct appeal, and Doctor Donahue chose to eliminate these three cases from his analysis.   The court will discuss that choice later in this decision.

Doctor Donahue then performed his statistical evaluation of the collected data using so-called regression analysis.   Doctor Donahue did not employ the least squares regression method, but preferred to use the logit technique.   The goal of any regression analysis is to tease out from the data whether certain independent variables are statistically associated with the production of the dependent variable under scrutiny.   In this case, the dependent variable is, of course, receiving a sentence of death.

While accurate and sufficient data are being amassed and organized, the analyst familiarizes herself or himself with the system or processes which result in the dependent variable.   In this case, that process is the Connecticut criminal justice system applicable to capital felony cases.   Once the analyst thoroughly comprehends the process or processes involved, the analyst proposes relevant and seemingly important factors which may potentially explain the outcomes observed.   These theoretical, explanatory factors become the independent variables tested using some regression-like method in order to substantiate, negate, and/or in some instances, quantify the contribution, if any, made by a particular independent variable with respect to the dependent variable or outcome of interest.

The nature of the data collected may impede or prevent usable regression analysis for a given proposed factor or for the dependent variable itself.   For example, there must exist a sufficient number of outcomes or the analysis will be unreliable.   One cannot flip a coin twice, observe two heads, and declare with assurance that the coin is a fair or biased one.   Both Doctor Donahue and the respondent's statistical analyst, Doctor Stephan Michaelson, agreed that the 200 or so cases whose outcomes were utilized are a large enough collection to conduct a regression study, even though these experts quarrel over the inclusion or exclusion of a few individual cases.

A second consideration is whether a prospective independent variable occurs often enough in the data set to perform a valid regression analysis with respect to that proposed factor.   In other words, the collection of information must contain a large enough number of cases where the putative explanatory variable existed in order to evaluate, statistically, if that factor does, indeed, affect the outcome under consideration.   Doctor Donahue acknowledged that no one can legitimately infer, using regression analysis, whether the gender of the perpetrator is an independent variable associated with imposition of the death penalty or not because too few Connecticut cases from 1973 to 2007 involved female capital felony murderers.

No other type of evidence was adduced pertinent to the issue of gender disparity.   Consequently, the petitioners cannot prevail on their allegation that the gender of the perpetrator unconstitutionally impacts the death penalty scheme in Connecticut or violates § 53a–46b(b)(1).

A third matter for consideration regarding regression analysis is that the number of independent variables used in such analysis must be limited in order to produce useful information.   Under the least squares regression analysis, one seeks a linear equation for which the explanatory variables appear with mathematically appropriate coefficients so as to maximize the analytical power, and therefore predictive utility, of the equation to match a plot of the dependent variable, i.e. outcomes.   In the case of logit, the logistical probability curve of the dependent variable indicated by the actual data is compared with the mathematical model produced by the proposed independent variable.   In other words, under either linear regression or logit, the analyst attempts to ascertain those coefficients for the independent variables which generate an equation which best “fits” the empirical data.

Too many proposed explanatory variables, or factors, dilute the effect of each explanatory variable so as to render the regression analysis uninformative.   Therefore, great care and diligent research must precede the choice of each independent variable because of the need to limit the number of variables included in the analysis.   The object of regression analysis is to extract genuine signals from the clatter of background noise created by the wealth of information gathered.   Excess explanatory variables retain, in this sense, too much noise.   Out of the hundreds of potential factors, the statistical experts in this case selected, at various times, from ten to thirty-five explanatory variables in the various iterations of their respective regression analyses.   It should be pointed out that very useful and important deductions can be drawn through regression analysis even though less than all conceivable explanatory variables are utilized.

“Perfect predictors” must also be avoided.   A perfect predictor, as the name implies, is an explanatory variable that correlates highly with the dependent variable.   Use of such variables may skew the coefficient calculations and drown out more subtle contributory factors.   Perfect predictors are, themselves, typically not very informative.   This is because perfect predictors add little or nothing to what the investigator already knew about the dependent variable.   As a blatant example, if one were attempting to learn which factors explain longevity using a regression method, the analyst would not include an independent variable such as having a smallpox vaccination scar.   While that criterion would correlate quite nicely with longevity, that correlation is utterly bereft of new information and the inclusion of that as an independent variable might alter the calculation of coefficients of other, more interesting, explanatory variables.

Along the same lines, the analyst ought to identify and eliminate two or more proposed explanatory variables which really measure the same circumstance or condition.   The regression analyst would likely avoid using both acreage and frontage as independent variables when attempting to determine the factors which contribute to the resale value of a residential property because typically, although not always, frontage and acreage are both ways of measuring lot and building size, and these variables fluctuate sympathetically.   Inclusion of independent variables which really classify, categorize, or measure the same characteristics can also skew the regression equation coefficients and produce inaccurate results.

Another problem facing the analyst is the use of explanatory variables which require subjective evaluation.   In the present case, one such explanatory variable is egregiousness of the crime or crimes.   Students from Yale Law School and the University of Connecticut School of Law perused the selected cases to gauge egregiousness.   These evaluators were, as a group, young and well-educated.   Doctor Donahue justified this procedure by noting that studies have shown that assessments of severity of crimes and torts by young, educated persons match those of our nation's population as a whole as far as the numerical mean, or average, value of egregiousness of a given scenario.   Although the mean value is roughly the same, the “spread” surrounding that mean is different in that it is more dispersed for the general public and narrower for the younger and more educated group.   In other words, the egregiousness scores of the larger group will vary more and include more extreme opinions for the same hypothetical, while the more homogeneous evaluators' scores will cluster closer to the mean value.   The average score, however, will be about the same for both groups of evaluators.

2.

In footnote four of State v. Cobb, 234 Conn. 735, 738 (1995), our Supreme Court enumerated “[f]our general conventions of statistical analysis [that] may affect the significance of the [petitioners'] preliminary data.” as derived from the McCleskey case, supra.

1. Any statistical analysis “must reasonably account for racially neutral variables which could have produced the effect observed”;

2. “[S]tatistical evidence must show the likelihood of discriminatory treatment by the decision-makers who made the judgments in question”;

3. “[T]he underlying data must be shown to be accurate”;  and

4. “[T]he results should be statistically significant.”

Our Supreme Court recognized that these McCleskey conventions are guidelines which allow for methods of proof besides statistical analysis to establish purposeful discrimination.  Id. However, in the present case the petitioners have relied exclusively on statistical inferences to prove their claims, and, therefore, these conventions provide a useful backdrop by which to weigh that evidence.  Id.

3.

The court now discusses Doctor Donahue's conclusions as to the statistically significant results of his regression analysis.   The phrase “statistical significance” has no fixed scientific or legal meaning.   It simply denotes that the set of actual outcomes departs markedly from the expected outcomes if the outcomes were purely the result of random occurrences.   The statistician will usually declare, ahead of time, what numerical level of probability will be considered a marked departure.   For a variety of reasons, an array of outcomes which would occur only five percent or less among the many possible distributions produced by random chance alone is often used as the boundary line to demarcate statistical significance.   That .05 probability is neither talismanic nor mathematically compelled, but it is analytically convenient, objective, and utile.   If no other probability standard is announced, most statisticians assume that .05 probability is to be employed.   Doctor Donahue used this figure in drawing his conclusions as to statistically significant findings.

Another threshold comment is warranted.   Because “race” is an artificial and elusive description, the evaluators were instructed to assign a particular racial status according to whether the evaluator perceived that the person to be classified possessed stereotypically recognized physical characteristics.

In Doctor Donahue's view, his regression analysis demonstrated that, in general:

1.  A black perpetrator who slew a white victim in Connecticut had a statistically significant higher probability of being charged with capital murder than a white perpetrator who killed a white victim in Connecticut.

2. A black perpetrator who murdered a white victim in Connecticut had a statistically significant higher probability of receiving a death sentence than a white murderer whose victim was also white.

3. Any defendant convicted of capital murder in the Waterbury Judicial District had a statistically significant higher probability of receiving a death sentence.

4. The egregiousness of the crime had little impact on the outcome with respect to charging or sentencing decisions in Connecticut.

Doctor Donahue inferred from his findings:

1.  No benign factors explain these disparities;  and

2.  The death penalty system in Connecticut operated in an arbitrary and unfair manner.

Doctor Donahue also made other statistically noteworthy observations:

1.  No statistically significant evidence supported an inference that the Waterbury State's Attorneys Office charged defendants with capital murder any differently than prosecutors from other judicial districts in Connecticut, i.e. the disparity concerning Waterbury appeared only as to sentencing decisions by the factfinder.   This discrepancy appears unrelated to the perceived race of the perpetrator or victim.

2. Black perpetrators who killed minority victims received the death sentence slightly less often than white perpetrators who murdered white victims.

3. The disparity levels calculated by Doctor Donahue for Connecticut regarding black perpetrators and white victims are the same as calculated for every other state for which a similar regression analysis has been conducted and the same as the level for a composite of jurisdictions nationally.   That is, the death penalty outcomes in Connecticut with regard to the circumstance of a minority perpetrator and a majority victim exhibit the same degree of disparity as elsewhere in the country;  no better, no worse.

4. Having a previous criminal record and being sentenced to prison for other crimes appeared to be a negative indicator, i.e. analyzed as a group, it was less likely in Connecticut, during the pertinent time period, for a previously incarcerated perpetrator to be sentenced to death than a perpetrator without such a history.   See Exhibit 769.

Doctor Donahue never conducted a study or offered an opinion regarding whether the death penalty procedure in Connecticut influenced the filing of capital murder charges or the imposition of the death penalty in any individual case.   No statistical evidence was adduced by anyone focused on whether a particular death sentence was inappropriately sought or imposed as a result of racial or geographic discrimination.   In other words, Doctor Donahue's statistical analysis was confined to an examination of the potential factors associated with the death penalty generically.

Doctor Donahue completed the statistical study as it was requested by the OCPD. However, that global statistical overview fails to generate the evidence needed to establish any constitutional or statutory violation as explicated by the McCleskey decision, viz, that a particular petitioner's death sentence was the product of purposeful discrimination.   Put another way, Doctor Donahue's testimony was a relevant starting point but a far cry from the proof necessary to establish the invalidity of the death sentence imposed upon the individual petitioners.

The salient question is not whether race may influence death penalty outcomes in Connecticut.   The critical issue is the inverse question of whether the death sentence of any of the petitioners in this case was the product of intentional discrimination.

The petitioners recognize the shortcomings of their statistical evidence.   In their posttrial brief, they argue that the United States Supreme Court and the Connecticut Supreme Court have wrongly continued to follow the standards of proof set forth in McCleskey with regard to allegations of violations of the federal constitutional rights discussed earlier;  that the Connecticut Supreme Court has wrongly declared that the federal constitutional analysis applies to similar state constitutional claims;  and that § 53a–46b(b) creates a distinct route by which to challenge death sentences without the need to prove purposeful discrimination.   The petitioners also acknowledge that this habeas court is legally bound to apply the legal principles set forth in the appellate precedents.

The court points out that, of the nine death penalty cases considered by Doctor Donahue's statistical study, only three fit within the global findings inferred by Doctor Donahue based on considerations of the races of the perpetrator and victim, namely the cases of Webb, Cobb, and Reynolds.   The other six capital murders for which the death penalty was imposed and upheld on appeal did not involve a black defendant who murdered a white victim, including those for the petitioners Breton and Rizzo who are classified as white in their DCIs.

Neither can any of the petitioners rely on any geographic “disparity” to support their allegations because Doctor Donahue found no statistical evidence that Waterbury prosecutors charged defendants any differently than prosecutors from other districts.   As to any supposed Waterbury sentencing anomaly, this court can safely assume that the same set of twelve jurors did not sit on more than one death penalty case.   No evidence of that astounding coincidence was introduced, nor that even a single juror sat on more than one capital murder case arising in Waterbury.

The petitioners cannot point to any feature or condition of the Waterbury Judicial District which would indicate that Waterbury venire persons are more receptive to the death penalty than elsewhere.   Convention two, as described in footnote four of State v. Cobb, supra, obligates a petitioner whose death sentence was imposed in the Waterbury Judicial District and who claims geographical prejudice to offer statistical evidence which shows “the likelihood of discriminatory treatment by the decision-makers who made the judgments in question.”  Id. No evidence, statistical or otherwise, was presented that sentencers for Waterbury cases were treating the defendants in Waterbury cases more harshly because the case arose in the Waterbury area.

Assuming, arguendo, that Doctor Donahue's analysis was reliably derived from the data collected, his conclusions leave the question of unconstitutional discrimination unproven as to any of the petitioners in this habeas case or any other individual on death row in Connecticut.   Perhaps, the following example will crystallize this deficiency as discerned by the court.

Statistical analysis of many outcomes of the flip of a coin may accurately demonstrate that the coin is a fair one in the sense that it is equally likely to show a head or tail.   That analysis can also reliably predict that, given many future trials, a tail will show in about half the flips.   But that statistical analysis cannot predict which particular future flip will be a tail or not.   Neither can the analyst explain why flip number 638 displayed a tail rather than a head in the past.   Analogously, using Doctor Donahue's statistical study, as conducted in accordance with the commission of the study by the OCPD, no one can say that a particular defendant's death sentence was or was not unfairly produced by racial animus.   The most one can say from that study is that some capital murder cases were influenced by the racial dynamics of a minority perpetrator and a majority victim, but one cannot say, based on the results of the study alone, that a particular inmate received the death penalty due to intentional racial discrimination or that any petitioner did.

The petitioners attempt to treat all capital murder cases as interchangeably vulnerable in that they wish the court to view every case as being equally impacted by improper racial considerations rather than trying to establish which individual minority petitioner did not deserve to be charged with a capital crime or sentenced to death.   Under McCleskey, untoward racial factors cannot be spread among all death row prisoners as insurance spreads economic risk among all insureds.

As noted above, convention two of the McCleskey analysis indicates that the statistical study in question “must show the likelihood of discriminatory treatment by the decision-makers who made the judgments in question.”  State v. Cobb, 234 Conn. 735, 738 n.4 (emphasis added).   These decision-makers would be the prosecutors, judges, and/or jurors in the petitioners ' cases.   In order to prevail in this habeas case each petitioner bears the burden of proving, through the use of statistical analysis as well as other evidence, that his death sentence was the product of racial animus on the part of one or more of the decision-makers.   If the racial disparity, uncovered through statistical studies is purely the result of demographic, economic, sociological, psychological, or other nonpurposive forces which are manifest only through examining all capital cases, collectively, the petitioners cannot prevail in their claims of the unconstitutionally of their convictions and sentences, McCleskey v. Kemp, supra, 292.   Although such culture-based disparity may be important and even determinative for policy-makers, such as a state legislature, it is immaterial to the question before this court.

The court concludes that none of the petitioners has met his burden based on the evidence admitted at the habeas hearing and the law as elaborated above.

III

Ordinarily, that conclusion would relieve the habeas court from addressing other issues in the case.   However, this proceeding is an extraordinary one.   The petitioners seek to have the Connecticut Supreme Court diverge from the constitutional principles espoused in McCleskey and also contend that § 53a–46b(b)(1) affords a separate avenue for relief which dispenses with the need to demonstrate that intentional discrimination based on race has polluted the judicial process for any of the petitioners.   If our Supreme Court should decide to shed the jurisprudence enunciated in previous caselaw, then the failure of this court to discuss in greater detail its assessment of Doctor Donahue's opinions might necessitate relitigation of this case.   Prudence and judicial efficiency dictate that this court make findings and rulings beyond those strictly needed to resolve this case under the McCleskey decision.

A.

In order to evaluate Doctor Donahue's statistical study further, the court will peruse the remaining conventions delineated in footnote four of State v. Cobb, 234 Conn. 735, 738, viz. conventions one, three, and four.   To reiterate, convention one inquires into whether the study reasonably accounts for racially neutral variables which could have produced the observed effect;  convention three deals with the accuracy of the underlying data;  and convention four asks whether the results are statistically significant.

Doctor Donahue's statistical study is accurate to the extent that the underlying data collected from the more than 200 cases included in his analysis correctly summarizes the factual information pertaining to these cases in general.   Also, the regression analyses performed by him were, for the most part, computationally correct, i.e. the calculations derived were without mathematical errors.   However, the court also concludes that some of the steps he took were flawed and some of the inferences he drew from the outcomes were questionable and lacked probative value.

As mentioned earlier, the respondent presented the testimony of Doctor Stephan Michaelson to critique the analysis conducted by Doctor Donahue.   Doctor Michaelson is also a statistical analyst who possesses impressive credentials.   He obtained a Ph.D. in Economics from Stanford University and did postgraduate work at Harvard University.   He has taught economics and related subjects and has a wealth of experience in the statistical investigation of the relationship of education, race, and gender with respect to employment and employment discrimination.   He has handled statistical analyses related to litigation, including cases in which the jury summoning process was challenged.   He has also been consulted and testified in cases involving scrutiny of the death penalty in the past.   The respondent retained Doctor Michaelson in 2007 to review Doctor Donahue's work.

Again, the experts agreed that the 200 plus cases which formed the data set were sufficient in number to execute a useful regression analysis.   However, in only nine cases was a death sentence imposed and upheld on appeal.   This court lacks confidence that valid inferences regarding causation, as opposed to mere association or correlation, can be gleaned from this paucity of death sentence cases.   With such a small number of death penalty outcomes, the addition or omission of one or two cases which contain or lack a pertinent circumstance can greatly disturb the regression results.

The uncertainty concerning inferences derived from the small number of death sentences was compounded by Doctor Donahue's decision to discard three cases in which the death penalty was sought and obtained but later overturned on appeal.   If one is examining the decisions to pursue the death penalty or to impose that penalty, in order to ascertain if race or some other impermissible factor influenced those decisions, then logically all cases in which the death penalty was sought must be included even though some of those sentences were vacated on appeal because of some legal deficiency.   The activities surrounding the pursuit of the death penalty by prosecutors and the judgment that a death sentence ought to be imposed were the critical conduct which needs examination.   Postsentencing appellate review is irrelevant when attempting to evaluate the behavior of the decision-makers which did, in fact, produce a death sentence.

The court deduces from the documentary evidence that the three cases excluded by Doctor Donahue were State v. Johnson, 253 Conn. 1 (2003);  State v. Colon, 272 Conn. 106 (2004);  and State v. Courchesne, 296 Conn. 622 (2010).   The DCIs for these cases, Exhibits 550, 624, and 625, respectively, disclose that none of those murderers were classified as black.   Johnson and Courchesne are labeled as white, and Colon is labeled as Hispanic.   As a result, Doctor Donahue erroneously eliminated from his regression analysis three out of the mere twelve case studies in which the death penalty was sought and imposed from 1973 to 2007, that is, twenty-five per cent of all such cases.   Also, because only three of the remaining nine cases involved a black perpetrator and white victim, the omission of the Johnson, Colon, and Courchesne cases creates the impression that three out of nine death sentences involved this racial polarity when actually three out of twelve is the correct ratio.   Doctor Donahue's exclusion casts significant doubt on the reliability of his findings and, more importantly, on his conclusions about black perpetrator/white victim disparity in Connecticut.

If the original number of death sentences were large, as it would be in some jurisdictions, the improper deletion of these cases might be insignificant.   With Connecticut's small number of death sentences to begin with, namely twelve from 1973 to 2007, the impact of omitting the three cases is magnified.   Suppose there had only been eight death sentences of which only three involved a black perpetrator.   Assume further that three of the other five cases were improperly discarded.   This would leave three black defendants sentenced to death and only two nonblack defendants.   No one would seriously argue that this type of erroneous elimination of cases was inconsequential, for it would transform a nonblack majority into a minority of death row inmates.   It seems to the court that refusing to consider the cases of Johnson, Colon, and Courchesne in the regression analysis leaves the results of that analysis suspect.

Also, the very small number of cases in which the death penalty occurred may account for the strange, even bizarre, findings which flowed from Doctor Donahue's regression analysis.   As noted earlier, the data produced the result that, although the Waterbury prosecutors charged capital crimes in conformity with the other judicial district prosecutors, the number of death sentences imposed was higher to a statistically significant degree.   In other words, the charging practice in Waterbury with respect to capital crimes was typical for the entire state;  yet the rate of death sentencing was, according to the study, dramatically higher for the Waterbury district.

Because the sentencing bodies in the Waterbury cases, either juries or panels of judges, were composed of different individuals for each case, it is difficult to imagine that those death sentences were “caused” by some impermissible conduct by the decision-makers.   Doctor Donahue arrived at his opinion that the source of this sentencing anomaly must derive from some systemic impropriety because he was unable to conceive of a benign cause for that outcome.   However, Doctor Donahue ignores the possibility that this strange result was merely a random distribution stemming from the small number of death sentences, viz, nine out more than 200 capital cases.

He also fails to consider the very plausible possibility that the prosecutors for the Waterbury district possessed greater experience and skill and/or greater willingness to devote resources with respect to capital cases.   Perhaps, that prosecutorial staff was more adept at choosing which cases to pursue to the fullest extent possible;  at selecting amenable jurors to hear these types of cases;  at effectively presenting evidence in such cases;  and/or in making persuasive arguments regarding aggravation and mitigation.   With so few death sentences actually meted out, one case added or subtracted may perturb the statistical results enough to give the impression of great discrepancy.

The sparsity of death sentences undermines the court's confidence in Doctor Donahue's opinion that the Waterbury anomaly, if it exists at all, is evidence of constitutional or statutory violations.   One might as well infer that there is something peculiar in the physical environment surrounding the Waterbury courthouses which spawns death sentences even though the charging decisions are typical of other districts.

The exceedingly small number of death sentences may also explain the weird finding that capital defendants who have served prison sentences for previous crimes are slightly less likely to receive a later death sentence in Connecticut from 1973 to 2007.   Can this outcome mean anything except that the dearth of death sentence cases, when compared to the number of capital murder cases where no death sentence was imposed, produces regression analysis results susceptible to noticeable, but inconsequential, fluctuations?   One cannot reasonably infer that the lack of a prior prison record, itself, is an aggravator rather than a mitigating circumstance.   Such an inference would be absurd.

That does not mean, however, that the lack of a previous record may not express, through regression analysis, an association with the imposition of a death sentence.   The court here is distinguishing correlation from causation.   It is quite possible that the statutory definitions of capital murder and/or aggravators or mitigators minimize or negate the influence of a defendant's criminal history.   For instance, in the case of murder for hire, the contract killer may be more likely to have lived a life of crime than the person who hired the killer, who wishes to eliminate a troublesome spouse, business associate, or rival while leaving the dirty work to others.

Also, gang warfare may generate a relatively high number of murders committed by persons with prior records when compared to thrill murders.   However, the latter type of killing is more likely to be performed in an especially heinous, cruel, or depraved manner, and the killer may select a more vulnerable and sympathetic victim than the victim in a battle for drug turf.   Also, such thrill murders may be more likely to entail sexual assault and/or kidnapping.

Had the data set of death sentences in Connecticut been large enough, perhaps any such nuances would be apparent.   The point is, that with only nine death sentences imposed, the useful information from regression analysis remains hidden, and neutral explanations cannot be eliminated with any level of assurance.

Doctor Michaelson's main criticism of Doctor Donahue's approach, and one that the court finds compelling, is that Doctor Donahue never attempts to parse out the root causes of the outcomes displayed by his regression analysis.   When one observes unusual statistical results, it behooves the analyst to investigate further to discern what events could trigger such unexpected or rare relationships.

In that endeavor, one must apply common sense and one's understanding of the ways of the world.   Of course, there is a danger that the analyst's subjective beliefs or subconscious biases will dissuade the investigator from following the path toward which the statistical calculations point and lead to erroneous conclusions.   One must be open-minded and receptive to unanticipated correlations.   But one cannot simply presume and declare that all disparities must be rooted in unlawful and unconstitutional activities on the part of governmental actors.   Even using identical numerical data, a reasoning person may arrive at divergent conclusions as to causation of the data sets, as the following example will illustrate.

Assume that for the period 1973 to 2007, substantial rain fell on only nine days in a remote desert area.   Suppose, also, that these nine days of heavy precipitation were distributed thusly:  three days in February;  two days in January;  two days in March;  and one day in October and June.

Also, assume that these days of rain fell three times on a Wednesday, two times on a Tuesday;  two times on a Thursday;  and one time on a Sunday and a Friday.

Note carefully that these two distribution patterns possess the identical numerical and statistical structure.   With the cluster of rainy days in January, February, and March, one could reasonably hypothesize that some seasonal condition or conditions play a role in developing the few days of rain which occurred.   But would any rational person draw the same conclusion that there exists some causal nexus surrounding Wednesday which increases the likelihood of precipitation?   Of course not.   One would attribute that midweek cluster to mere coincidence.   We apply our knowledge and experience of the way the world seems to work and deduce two different causal explanations for the exact same numerical data sets.   We confidently do so because it makes perfect sense that different seasons can, on average, produce different weather, while the days of the week seem interchangeable with respect to the weather.   If the desert were densely populated, one might justifiably suspect that some human activity peculiar to midweek existed.   This would demand further investigation.

It is important, then, to try to account for why the odd distributions occurred rather than simply ascribe improper motive or capriciousness to the death penalty scheme in Connecticut.   Again, this endeavor is especially important when one considers how few Connecticut cases were examined in which a death penalty was actually imposed.   Indeed, convention one from the McCleskey paradigm states that any statistical analysis regarding the death penalty “must reasonably account for racially neutral variables which could have produced the effect observed,” State v. Cobb, n.4, 234 Conn. 735, 738 (1995).   The court finds that Doctor Donahue's opinions are deficient with regard to the task set forth in convention one.   The court concludes that while this regression finding, that the black perpetrator/white victim racial tandem is associated, in some way, with the imposition of the death penalty is accurate, his opinion that such influence must be the result of capricious or other impermissible factors is without evidentiary support and is purely conjectural.

B.

Even if the McCleskey holding that only purposeful racial discrimination will vitiate an otherwise constitutional death sentence is ignored, and the approach urged by the petitioners concerning a collective rather than individualized statistical attack on Connecticut's capital punishment system were applied, Doctor Donahue's opinions still fall short.   In order for the petitioners to establish a violation of a constitutional or statutory provision, the petitioners must prove, using that global statistical analysis, that the disparities shown by the analysis was the product of Connecticut's capital punishment procedure and not merely a reflection of ambient social or psychological forces.   In other words, the petitioners must demonstrate that the arbitrary, freakish, or disparate outcomes resulted from some flaws in our legal framework or the positive acts of governmental agents or officials.   If extrinsic, nongovernmental circumstances give rise to the undesirable disparities, the petitioners cannot prevail in this action.   Such external pressures and factors may be highly relevant, and even determinative, for legislative bodies when weighing whether to enact or to repeal a death penalty, but such factors are immaterial to the question of whether that scheme, on its face or as applied, violates existing statutory or constitutional law.

As discussed earlier, the McCleskey decisions, adopted by our appellate courts, expressly declared that “the nature of the capital sentencing decision, and the relationship of the statistics to that decision, are fundamentally different” from discrimination analysis in the area of venire selection, employment, and educational opportunity, McCleskey v. Kemp, supra, 294–97.   But even racial disparity challenges in those area of the law require that the disproportionate outcome derives from official policies or conduct.

In the case of college admissions, “a state has an affirmative duty to remove policies, tied to the past, by which it continues to discriminate,” Hopwood v. Texas, (emphasis added), 78 F.3d 932, 955 (5th cir.1996), cert. denied, 518 U.S. 1033 (1996).  “[A]ny official action that treats a person differently on account of his race or ethnic origin is inherently suspect,” (emphasis added), Fisher v. University of Texas (June 24, 2013), p.8. However, racial imbalance which is not induced by governmental conduct or policies is not, in itself, unconstitutional.   Nor does that circumstance allow courts to countenance attempts to rectify that imbalance by resorting to “benign” race-based action.  Id., p. 6. “A university is not permitted to define diversity as some specified percentage of a particular group merely because of its race or ethnic origin.”  Id., p. 9. Racial balancing is “patently unconstitutional.”  Id.

Within the employment discrimination context, the United States Supreme Court expounded a similar principle in a case arising in Connecticut, Ricci v. DeStefano, 557 U.S. 557 (2009).   In that case, the city of New Haven administered a promotional examination for firefighters.   That exam produced an outcome such that no black candidates achieved eligibility for promotion and only one nonwhite candidate did qualify.   New Haven feared that this statistical disparity would subject the city to liability, and the city discarded the results.   The white candidates who had passed the exam brought suit.  Id., 562.   The high court concluded that rejection of the results, which results clearly had disparate impact on minorities, was itself discriminatory against whites.  Id., 563.

That Court held that if the disparate test results for minority groups was the product of some defect in the examination process, then the city's concern would be legitimate.  Id., 578.  “Based on the degree of adverse input reflected in the results, [the city was] compelled to take a hard look at the examination to determine whether certifying the results would have had an impermissible disparate impact.”  Id., 587 (emphasis added).  “The problem for [the city] is that a prima facie case of disparate-impact liability—essentially, a threshold showing of a significant statistical disparity ․ and nothing more —is far from a strong basis in evidence that the City would have been liable under Title VII had it certified the results.”  Id. (emphasis added).  “[T]he City could be liable for disparate-impact discrimination only if the examination were not job related and consistent with business necessity ․” Id. (emphasis added).  “[T]here is no strong basis in evidence to establish that the test was deficient ․” Id.

Thus, even for employment and educational opportunity cases of alleged discrimination, a mere showing of statistically significant, disparate results is insufficient evidence, as a matter of law, to establish liability for unlawful discrimination.  Id. In the present case, then, the petitioners need to show that the statistical disparity, about which they based their putative claims, was caused by some unnecessary or extraneous feature of the criminal justice system in Connecticut regarding the death penalty.

Some of Doctor Donahue's conclusions actually point away from Connecticut's legal system being the source of the death penalty disparity with respect to the minority perpetrator/majority victim variable.   Recall that Doctor Donahue noted that the disparate statistical outcome for black perpetrator/white victim cases in Connecticut mirrored the statistical outcome concerning this variable both nationally and in every other state or district of a state for which such a statistical study was done.

Also, one must recall that Doctor Donahue's study disclosed no statistical disparity with regard to minority defendants in capital cases generally.   His results found statistically significant deviation from the expected value only in the situation of a black perpetrator who was convicted of murdering a white victim.   This difference reasonably suggests that the source of the disparity in the mixed race case was not the race of the perpetrator alone.   Indeed, Doctor Donahue testified that statistical studies from other jurisdictions have shown that this divergence of outcomes probably reflects the tendency of members of the majority to be more empathetic to majority victims, who resemble themselves, and less sympathetic to minority perpetrators, with whom they identify least.

The court concurs that this potential explanation is the most likely reason why the minority perpetrator/majority victim murder cases have a greater probability of producing a death sentence.   But this explanation stems from a psychological phenomenon rather than some flaw created by the legal system and its procedures.   That phenomenon would be inherent in any jurisdiction which has a population containing a majority subgroup, no matter what death penalty protocol existed.   Again this type of subconscious influence, which manifests in the minority perpetrator/majority victim disparity, is of paramount concern to policy makers but has no bearing on the constitutionality of Connecticut's death penalty system because that disparity does not derive from that system under the holding of Ricci v. DeStefano, supra, 587.   Therefore, even under the petitioners' argument that the court ought to ignore the McCleskey standard for evaluating statistical evidence pertaining to the death penalty and apply the scrutiny used in employment discrimination cases, the petitioners' claim fails with respect to the minority perpetrator/majority victim disparity.

Constitutional and fair legal procedures may produce unequal results.   Inversely, disparate outcomes standing alone, do not necessarily mean that the legal processes which preceded that outcome were unfair or unlawful.

Summary of Conclusions

1. The appellate courts of Connecticut have construed the pertinent state constitutional provisions as analytically identical with their federal counterparts as set forth in the United States Supreme Court case of McCleskey v. Kemp, supra.

2. The holding of McCleskey required each petitioner to prove that a decision-maker in his own case engaged in purposeful discrimination based on the race of the perpetrator and/or victim in order to establish any constitutional violation.

3. Section 53a–46b(b)(1) also requires an individualized consideration of whether race or some other improper factor played a role in producing the death sentence in that petitioner's own case.

4. The petitioners only introduced global statistical evidence which pertained to no case in particular but simply found disparities relevant to capital cases in general.

5. Consequently, the petitioners have failed to satisfy the purposeful discrimination criterion under McCleskey v. Kemp, supra.

6. Even if the court had utilized the employment/educational opportunity discrimination analysis urged by the petitioners instead of the McCleskey standard for death penalty cases, the petitioners failed to prove that the collective, racial disparity about which they complain arose from some deficiency in the Connecticut capital punishment system rather than plausible neutral causes, such as subconscious associational bias.

7. Doctor Donahue's regression analysis was accurately executed, however his opinions based on that analysis are of doubtful value because of the inordinately small number of death sentences actually imposed from 1973 to 2007 and his erroneous exclusion of three cases, or twenty-five percent, of that already small number of cases.

8. As a result of that paucity of information, the Waterbury sentencing anomaly was most probably the result of innocuous coincidence.

The petitioners' claims for habeas corpus relief are, therefore, denied.

Sferrazza, S.J.

Sferrazza, Samuel J., S.J.