People v. Bryant, 491 Mich. 575

Defendant was convicted of first-degree criminal sexual conduct, armed robbery, and marijuana possession.
He challenged the jury venire for racial composition, noting only one African American among 45 summoned and few appeared; 132 appeared, 45 were drawn for defendant’s venire.
The Kent County jury-selection system used a state database of eligible residents; race was not recorded and race did not influence selection.
A remand revealed a computer programming error reduced the usable pool from 453,414 to 118,169 names, biasing juror questionnaires toward zip codes with fewer African Americans.
Expert analyses showed underrepresentation of African Americans in jury summons (4.17%) versus community population (8.25%), with various statistical tests applied over three months.
The Michigan Court of Appeals found a Sixth Amendment fair-cross-section violation; the Michigan Supreme Court reversed, holding the second prong was not violated and reinstated the convictions.

Issue	Plaintiff's Argument	Defendant's Argument	Held
Whether under Duren’s second prong the representation of African-Americans was not fair and reasonable.	Bryant contends underrepresentation in venires over time violated Duren.	Bryant argues the underrepresentation is not fair and reasonable when measured across venires.	No, the Court reverses the Court of Appeals; under the most reliable data, representation was fair and reasonable.
Whether the Court of Appeals erred by using only defendant’s venire data to judge the second prong.	Bryant asserts broader venire data over time should be considered.	City argues the data set was sufficient to assess the second prong.	No, the Court held Duren requires analysis over venires over time with reliable data.
Whether the tests used (absolute disparity, comparative disparity, standard deviation, disparity of risk) support a finding of underrepresentation.	Bryant argues results show underrepresentation.	The tests, especially disparity methods, are unreliable with small minority populations.	In sum, the tests do not establish unfair and unreasonable representation under the second prong.
Whether there was systematic exclusion under the third prong.	Bryant claims the zip-code bias from the programming error shows systematic exclusion.	The error was nonbenign but not reflective of the selection process’ intent.	Court finds, under third prong, that exclusion was systematic and therefore supports relief.

Duren v. Missouri, 439 U.S. 357 (1979) (established three-prong test for fair-cross-section challenges)
Taylor v. Louisiana, 419 U.S. 522 (1975) (requires juries drawn from a fair cross section, not perfect proportionality)
People v. Smith, 463 Mich. 199 (2000) (case adopting case-by-case, all-tests approach to Duren prongs; rejects exclusive reliance on any single test)
Hubbard (After Remand), 217 Mich. App. 459 (1996) (limits of second-prong evaluation and concerns about nonbenign causes influencing results)
Berghuis v. Smith, 559 U.S. 314 (2010) (u.S. Supreme Court held no single method is mandated to measure underrepresentation)