McGhee v. State

Antonio McGhee v. State of Maryland, No. 64, September Term, 2021.
Opinion by Biran, J.

CRIMINAL LAW – INEFFECTIVE ASSISTANCE OF COUNSEL –
“CSI-EFFECT” VOIR DIRE QUESTION – During jury selection at Petitioner’s 2007
murder trial, Petitioner’s attorney did not object to a voir dire question that asked: “Does
any member of this panel believe that the State has got to present fingerprint evidence,
DNA, blood sample evidence, ballistic evidence, any scientific evidence in order to
convince you of the defendant’s guilt? In other words, do you think the State has a
requirement to do that in all cases?” The jury found Petitioner guilty. In a trilogy of cases
that the Court of Appeals decided in the years following Petitioner’s trial, the Court held
that so-called “CSI-effect” voir dire questions and similar jury instructions can improperly
intrude on the province of the jury. Charles v. State, 414 Md. 726 (2010); Atkins v. State,
421 Md. 434 (2011); Stabb v. State, 423 Md. 454 (2011). In 2014, Petitioner filed a
post-conviction claim alleging that his trial counsel provided ineffective assistance of
counsel by not objecting to the CSI-effect voir dire question at his trial. The Court of
Appeals held that, under the prevailing professional norms that existed in 2007, defense
counsel’s failure to object to a CSI-effect voir dire question did not render counsel’s
performance constitutionally deficient under Strickland v. Washington, 466 U.S. 668
(1984). The Court declined to address the retroactivity of Charles, Atkins, and Stabb in
determining whether counsel’s conduct was objectively unreasonable, because Strickland
requires a post-conviction court to assess an attorney’s performance based on the prevailing
professional norms at the time of the contested conduct. Cases that are decided after a
defendant’s trial do not shed light on the professional norms that existed at the time of the
defendant’s trial.
Circuit Court for Prince George’s County
Case No.: CT07-1096X
Argued: September 12, 2022

IN THE COURT OF APPEALS

OF MARYLAND

No. 64

September Term, 2021

ANTONIO MCGHEE

v.

STATE OF MARYLAND

Fader, C.J.
Watts
Hotten
Booth
Biran
Gould
Eaves,

JJ.

Opinion by Biran, J.

Filed: October 24, 2022

Pursuant to Maryland Uniform Electronic Legal
Materials Act
(§§ 10-1601 et seq. of the State Government Article) this document is authentic.

2022-10-24
11:21-04:00

Suzanne C. Johnson, Clerk
 This case arises from a post-conviction court’s grant of a new trial to Petitioner

Antonio McGhee based on ineffective assistance of counsel. In December 2007, a jury in

the Circuit Court for Prince George’s County convicted McGhee of the murder of Keith

Dreher. The basis of McGhee’s ineffective assistance of counsel claim is his counsel’s

failure to object to what courts and commentators have called a “CSI-effect” voir dire

question.

The “CSI effect” describes the theorized impact of television crime scene dramas

on jurors. The theory suggests that, based on the proliferation of programs such as CSI,

jurors in criminal cases now expect the prosecution to produce DNA evidence and/or other

forensic evidence to prove a defendant’s guilt, and that juries are prone to wrongfully acquit

criminal defendants where the prosecution does not produce such evidence. See, e.g.,

Robinson v. State, 436 Md. 560, 570 n.11 (2014) (citing Donald E. Shelton, Juror

Expectations for Scientific Evidence in Criminal Cases: Perceptions and Reality About the

‘CSI Effect’ Myth, 27 T.M. COOLEY L. REV. 1, 3 (2010)).

In 2010 and 2011 – more than two years after McGhee’s trial – this Court considered

three cases related to the CSI effect, and held in each that a CSI-effect message from the

bench constituted reversible error. See Charles v. State, 414 Md. 726 (2010); Atkins v.

State, 421 Md. 434 (2011); Stabb v. State, 423 Md. 454 (2011). One of the questions before

this Court is whether to apply these three cases retroactively in the context of an ineffective

assistance of counsel claim.

In Strickland v. Washington, the Supreme Court set out the controlling test for

evaluating an ineffective assistance of counsel claim under the Sixth Amendment. 466 U.S.
668 (1984). The Strickland test requires a petitioner claiming ineffective assistance of

counsel to make two showings. First, the petitioner must show that counsel’s performance

was deficient. Id. at 687. “This requires showing that counsel made errors so serious that

counsel was not functioning as the ‘counsel’ guaranteed the defendant by the Sixth

Amendment.” Id. Second, “the defendant must show that the deficient performance

prejudiced the defense.” Id.

Under Strickland, in considering the first “performance” prong of the test for

ineffective assistance, we are bound to evaluate defense counsel’s conduct according to

professional norms that existed at the time of the contested action (or inaction). Id. at 689.

This analysis precludes us from evaluating counsel’s conduct based on law that did not

exist at the time. Thus, we do not assess counsel’s performance at McGhee’s trial as if it

occurred after this Court decided Charles, Stabb, and Atkins. Under the professional norms

that existed at the time of McGhee’s trial, defense counsel’s failure to object to a CSI-effect

voir dire question did not render her performance constitutionally deficient.

I

Background

A. Maryland Jurisprudence Concerning the “CSI Effect”

1. CSI-Effect Jury Instructions

The term “CSI effect” emerged in 2002. Robinson, 436 Md. at 570. Due to the

popularity of forensic crime scene television series such as CSI: Crime Scene

2
Investigation,1 commentators speculated that such programs may heighten juror

expectations for forensic evidence. Id. at 570-71 (citing Jenny Wise, Providing the CSI

Treatment: Criminal Justice Practitioners and the CSI Effect, 21 CURRENT ISSUES CRIM.

JUST. 383, 383-84 (2010); Simon A. Cole & Rachel Dioso-Villa, Investigating the ‘CSI

Effect’ Effect: Media and Litigation Crisis in Criminal Law, 61 STAN. L. REV. 1335,

1338-39 (2009)). Studies that considered whether viewing CSI-type programs affected

jurors’ verdicts yielded inconclusive results. See Robinson, 436 Md. at 571-72.

Nevertheless, some courts began giving jury instructions to guard against a potential

CSI effect. These were sometimes referred to as “anti-CSI effect” instructions. See id. at

572. The advent of these jury instructions generated unique questions about the interaction

of pop culture, the role of the jury, and the State’s burden of proof, prompting consideration

in Maryland’s appellate courts.

In 2007, the Court of Special Appeals considered for the first time whether the

giving of a CSI-effect jury instruction constituted reversible error – specifically, whether

the instruction relieved the State of its burden of proof. Evans v. State, 174 Md. App. 549

1
CSI: Crime Scene Investigation aired on CBS from 2000 through 2015. Set in Las
Vegas, CSI was immensely popular and spawned several spinoff series, including
CSI: Miami, CSI: NY, and CSI: Cyber. See CSI: Crime Scene Investigation, Editors of
Encyclopaedia Britannica, BRITANNICA, available at https://perma.cc/WA85-DK3A; see
also Atkins v. State, 421 Md. 434, 457-58 (Harrell, J., concurring) (observing that the
“success of ‘forensic’ dramas … skyrocketed in 2000 with the debut of CSI: Crime Scene
Investigation, referred to as ‘the most popular television show in the world’ at one time”
and noting that, in a 2006 Nielsen rating, 30 million people watched CSI in one night; 70
million people watched one of the three CSI shows then in production; and 40 million
people watched two other forensic dramas, Without a Trace and Cold Case) (citations
omitted). In 2021, the Las Vegas iteration of the CSI franchise returned to television as
CSI: Vegas. See CSI: Vegas, CBS, available at https://perma.cc/B9F6-ENEJ.

3
(2007). The State charged Evans and another man with possession and distribution of

heroin after conducting an undercover “buy bust” operation. Id. at 552-53. At trial, Evans’s

attorney cross-examined the investigating detective concerning specific investigative

techniques that the detective had not used, including the failure to capture the drug

transaction through the use of video or audio recording equipment. Id. at 562. This line of

cross-examination prompted the trial court to instruct the jury as follows:

During this trial, you have heard testimony of witnesses and may hear
argument of counsel that the State did not utilize a specific investigative
technique or scientific test. You may consider these facts in deciding whether
the State has met its burden of proof. You should consider all of the evidence
or lack of evidence in deciding whether a defendant is guilty. However, I
instruct you that there is no legal requirement that the State utilize any
specific investigative technique or scientific test to prove its case. Your
responsibility as jurors is to determine whether the State has proven, based
on the evidence, the defendants’ guilt beyond a reasonable doubt.

Id. Evans’s trial counsel did not object to the giving of this instruction. Id. at 564-65.

In his closing argument, Evans’s counsel highlighted the detective’s failure to

record the alleged transaction. See id. at 562-63. The co-defendant’s attorney echoed this

point in his closing argument, also noting the lack of forensic evidence: “You have a

situation where there are absolutely no scientific tests that implicate my client in any way.

There’s no audio. There’s no video. There’s no fingerprints. There is nothing.” Id. at

563-64.

On appeal, Evans argued that the instruction concerning specific investigative

techniques and scientific tests improperly relieved the State of its burden to prove his guilt

beyond a reasonable doubt. Id. at 562. Because Evans had failed to preserve this issue

through objection in the trial court, the Court of Special Appeals held that it was precluded

4
from considering it. Id. at 566. In dicta, the intermediate appellate court noted that the jury

instruction was a correct statement of the law and that it did not relieve the State of its

burden of proof. Id. at 570. However, the court “stress[ed] that the salutary effect of the

instruction is found in the advisement that the absence of such evidence should be factored

into the juror’s determination of whether the State has shouldered its burden if, and only if,

the absence of such evidence, itself, creates reasonable doubt.” Id. at 571. The court

continued: “The risk is greatest that such an instruction will run afoul of the prohibition

against relieving the State of its burden where the instruction is predominant in the overall

instructions and its relation to the reasonable doubt standard unclear.” Id. Thus, the court

advised, “the preferable practice is for the … instruction to be promulgated in conjunction

with the explication of the State’s burden to prove the defendant guilty beyond a reasonable

doubt.” Id.

Between 2007 and 2011, no Maryland appellate court addressed the CSI effect in

the context of a jury instruction.2 In 2011, this Court considered such an instruction for the

first time in Atkins v. State, 421 Md. at 434. Atkins was charged with three counts of assault

after he was involved in an altercation during which he stabbed three people. See id. at

438-39. Three days after the incident, police executed a search warrant at Atkins’s home

and found a large, non-foldable “Rambo-type” knife in Atkins’s bedroom. Id. at 439, 439

n.3. Police did not perform any scientific or forensic tests on the knife, and there was no

witness testimony at trial linking that knife to the stabbings. Id. at 439. Nevertheless, the

2
As discussed below, in 2010 this Court considered a CSI-effect voir dire question
in Charles v. State, 414 Md. 726 (2010).

5
State told the jury in its opening statement that Atkins used the knife found in his bedroom

in the stabbings. Id. at 439, 439 n.4. During cross-examination of a detective who found

the knife in Atkins’s bedroom, defense counsel probed the lack of DNA testing conducted

on the knife. Id. at 440. In his defense case, Atkins testified that, as he was being punched

and kicked during the fight, he took a pocketknife out of his pocket, opened it, and swung

it at his attackers in self-defense. Id. at 438-39.

The trial court granted the State’s request for a CSI-effect jury instruction over

defense objection. The court instructed the jury:

During this trial, you have heard testimony of witnesses and may hear
argument of counsel that the State did not utilize a specific investigative
technique or scientific test. You may consider these facts in deciding whether
the State has met its burden of proof. You should consider all of the evidence
or lack of evidence in deciding whether the defendant is guilty. However, I
instruct you that there is no legal requirement that the State utilize any
specific investigative technique or scientific test to prove its case. Your
responsibility as jurors is to determine whether the State has proven based
upon the evidence the defendant’s guilt beyond a reasonable doubt.

Id. at 441-42 (emphasis omitted). The jury convicted Atkins of the assault charges, and the

Court of Special Appeals affirmed the convictions, discerning no error in the CSI-effect

jury instruction. See id. at 442.

This Court reversed, concluding that Atkins’s case was distinguishable from Evans.

We observed that, in Evans, the defense “distorted the law” through its extensive

cross-examination concerning the lack of video or audio evidence of the drug purchase, id.

at 451, and through its “robust and vehement closing argument” in which it improperly

suggested that the prosecution was required to produce such evidence. Id. at 450 (quoting

Evans, 174 Md. App. at 570). Whereas, in Evans, the defense’s cross-examination and

6
closing argument made a curative jury instruction proper, see id. at 450-51, in Atkins

defense counsel “merely pointed out” during cross-examination the lack of specific

forensic tests, which made a curative instruction unnecessary. Id. In addition, the

instruction in Evans was improper because it invaded the province of the jury, suggesting

the inferences they should draw from the State’s failure to test the recovered knife for DNA

evidence. Id. at 453-54. However, the Court stated that its “conclusion that the instruction

as given was invalid [was] based on the particular facts in this case,” and explained that it

was “not hold[ing] that an investigative techniques instruction would never be proper.” Id.

at 454. Consistent with Evans, this Court observed that “the key to producing a valid jury

instruction is ensuring that the State is properly held to its burden, and any instruction

regarding what the State must produce in proving its case must be properly related to the

reasonable doubt standard.” Id.

A few months later, in Stabb v. State, this Court again considered a CSI-effect jury

instruction case, and again held that the giving of the instruction constituted reversible

error. 423 Md. at 471-72. Stabb was charged with a third-degree sex offense involving a

minor. See id. at 457. After the alleged assault was reported to law enforcement, the

authorities did not refer the victim for a Sexual Assault Forensics Exam (“SAFE”). Id. at

458-59. At trial, Stabb’s defense counsel “argued properly and without undue emphasis the

lack of corroborating physical evidence of the crime, and questioned [two detectives] as to

the likelihood of the existence of such evidence and why a SAFE was not performed, but

did not ‘harp’ impermissibly on the lack of physical evidence in its case-in-chief or during

7
closing arguments.” Id. at 471. Nevertheless, the trial court issued a preemptive CSI-effect

jury instruction. Id. at 471-72.

This Court held that the use of a preemptive CSI-effect instruction was improper,

again distinguishing Evans. See id. at 471. We noted that, after defense counsel’s

cross-examination of the detectives, “[t]he State responded, during recross-examination of

witnesses and in closing arguments, to defense counsel’s implication regarding the lack of

a SAFE, i.e., why a SAFE was not administered and the unlikelihood that a SAFE, had it

been administered, would have yielded testable DNA or fingerprints.” Id. at 472. We

observed that “[r]ebuttal by the State was the proper approach.” Id. However, “[w]hen the

trial judge injected the pertinent instruction into the jury’s calculus, it had more force and

effect than if merely presented by counsel, and could have influenced impermissibly the

drawing by the jury of inferences regarding the absence of physical evidence.” Id. (internal

quotation marks and citation omitted). We concluded:

In giving the “anti-CSI effect” instruction to the jury, the trial court directed
effectively the jurors not to consider the absence of a SAFE or corroborating
physical evidence. The trial court invaded impermissibly the province of the
jury deliberations with the given “anti-CSI effect” instruction under the
circumstances. The “anti-CSI effect” jury instruction given, in the
circumstances of this case, was improper because it relieved the State of its
burden to prove Stabb was guilty beyond a reasonable doubt, invaded the
province of the jury, and, thus, violated Stabb’s constitutional right to a fair
trial.

Id. We warned prosecutors that the use of a CSI-effect jury instruction is “fraught with the

potential for reversible error,” id. at 473, and advised that it should be “confined to

situations where it responds to correction of a preexisting overreaching by the defense, i.e.,

a curative instruction.” Id.

8
 Most recently, this Court decided Taylor v. State, 473 Md. 205 (2021). In Taylor,

we held that the trial court erred by giving a CSI-effect jury instruction sua sponte in

anticipation of a potential defense objection. Id. at 222, 231. Taylor involved a

complicating factor: Although Taylor’s trial had occurred in 2008, no timely appeal was

filed on his behalf. Id. at 224. A post-conviction court granted Taylor relief in the form of

a belated appeal more than eight years later. Thus, the case came to this Court as a direct

appeal, rather than on collateral review. See id. at 224-25. One of the questions presented

to us in Taylor was which law to apply: the law as it existed in 2008 at the time of Taylor’s

trial (i.e., only the dicta from Evans) or the law as it existed at the time we were considering

his belated appeal (i.e., Atkins and Stabb). We did not reach this question, however, because

we held that either way the jury instruction in Taylor’s trial was improper. Id. at 233-34.

2. CSI-Effect Voir Dire Questions

An alternative iteration of the CSI effect emerged in Maryland courtrooms

concerning voir dire questions. In 2010 in Charles v. State, the State convicted two

defendants of second-degree murder and use of a handgun in the commission of a felony

or crime of violence. 414 Md. 726 (2010). During voir dire, over defense counsel’s

objection, the trial judge asked the jurors: “[I]f you are currently of the opinion or belief

that you cannot convict a defendant without ‘scientific evidence,’ regardless of the other

evidence in the case and regardless of the instructions that I will give you as to the law,

please rise.” Id. at 730 (emphasis deleted).

This Court held that the trial court erred by using non-neutral language that

suggested the only option was to convict the defendant, without mention of the option to

9
acquit. Id. at 738; see also McFadden v. State, 197 Md. App. 238, 250-51 (2011) (holding

non-neutral CSI-effect voir dire question deprived defendant of right to fair and impartial

jury), abrogated in part on other grounds by State v. Stringfellow, 425 Md. 461 (2012).

Charles did not focus on the appropriateness of the CSI-effect message when posed to the

jury pool during voir dire. Rather, it focused on the appropriateness of non-neutral language

in comments from the bench. See Charles, 414 Md. at 733.

In 2018, the Court of Special Appeals considered a CSI-effect voir dire question in

the context of an ineffective assistance of counsel claim. State v. Armstead, 235 Md. App.

392 (2018). After reflecting on Evans, Charles, Atkins, and Stabb, the intermediate

appellate court concluded that trial counsel’s failure to object to a CSI-effect voir dire

question did not amount to ineffective assistance of counsel. The Armstead Court reasoned

that, at the time of Armstead’s trial in 2009, CSI-effect voir dire was permissible, and

arguably even “favored,” under Maryland common law (presumably referring to Evans).

Id. at 406. That being the case, it could not be said that Armstead’s trial counsel fell short

of the professional norms that prevailed at the time. See id. at 417 (“It is unreasonable to

require trial counsel to see the change that remained beyond the horizon.”).

Equipped with an understanding of the development of CSI-effect case law in

Maryland, we turn now to the CSI-effect voir dire question that the trial court asked in the

present case.

B. McGhee’s Trial and the Contested Voir Dire Question

McGhee’s murder trial went forward in December 2007. During voir dire, the trial

court asked the potential jurors the following question related to the CSI effect:

10
 Does any member of this panel believe that the State has got to present
fingerprint evidence, DNA, blood sample evidence, ballistic evidence, any
scientific evidence in order to convince you of the defendant’s guilt? In other
words, do you think the State has a requirement to do that in all cases?

McGhee’s attorney did not object to this question.

After selection of the jurors, the State presented evidence that, on the evening of

March 17, 2007, Keith Dreher was killed by a gunshot to the head outside a Papa Johns

restaurant in Oxon Hill, Maryland.

Detective Paul Dougherty testified that after the shooting, the Assistant Manager of

the restaurant, Jerrone Joyner,3 gave a statement to him, in which Jerrone said that he saw

the shooter and the victim in the restaurant for approximately 20 minutes prior to the

shooting. He also told Detective Dougherty that he would be able to identify the shooter if

he were shown a picture of him. One month later, Detectives Andre Brooks and Michael

Delaney met with Jerrone and presented him a photo array. When Jerrone saw McGhee’s

photograph he asked, “[w]as that the guy?” Detective Delaney responded, “I couldn’t tell

you that. I wasn’t there.” The detectives testified that Jerrone then identified McGhee and

said he was “a hundred percent” certain that McGhee was the shooter.

Jerrone’s testimony at trial differed from that of Detectives Dougherty, Brooks, and

Delaney. Jerrone recalled giving a statement to Detective Dougherty the night of the

shooting, in which he described the shooter. He recalled telling Detective Dougherty that

it was possible he could identify the shooter, but he did not recall telling police that the

3
Because two of the witnesses share the last name Joyner, we will refer to them by
their first names. We mean no disrespect.

11
shooter was in the restaurant before the shooting. Additionally, while Jerrone remembered

identifying McGhee to police, he testified that he only did so because he “knew the other

guys” and did not know McGhee. He also disputed that he made the identification with

certainty. While acknowledging that he saw the barrel of the gun, he said that he did not

get a good look at the shooter and did not see the shooting.

Two other witnesses who were at the Papa Johns restaurant at the time of the

shooting – Shamell Joyner and Demetrius Young – knew McGhee personally. Shamell,

who was working at the Papa Johns on the night of the murder, testified that he saw

McGhee in the restaurant shortly before the shooting. He recalled hearing an argument

going on outside before the gunshot, but he did not know who was arguing or what the

argument was about.

Young testified that he was at the Papa Johns restaurant the night of the shooting

waiting for Shamell to get off work, when he saw McGhee enter the restaurant. Young saw

McGhee ask Dreher for a cigarette before Dreher went outside. He heard a gunshot two

minutes later.

On March 21, 2007, Shamell accompanied Detective Delaney to Martin Luther

King Elementary School in Washington, D.C., where he identified McGhee. After

members of the Prince George’s County Police Department warrant squad identified

themselves to McGhee as police, McGhee ran from them. Detective Wayne Martin saw

McGhee running into a wood line with a gun protruding from McGhee’s clothing. Martin

testified that he lost sight of McGhee for approximately 15 to 30 seconds during the chase.

12
A short time later, Martin apprehended McGhee. McGhee did not have a gun on his person

at that time.

An evidence technician for the Prince George’s County Police Department testified

that he recovered a loaded, “reasonably clean” sawed-off shotgun from underneath leaves

along a fence at the rear of the school on March 21, 2007. Although the gun was submitted

for DNA analysis, there was not sufficient genetic material on it from which to develop a

genetic profile. Similarly, latent fingerprints lifted from the weapon did not contain enough

characteristics for an examination to be conducted. However, Detective Martin testified

that the recovered firearm appeared to be the gun he saw in McGhee’s possession while

McGhee was running.

Susan Lee, a Prince George’s County Police Department firearms examiner,

testified that plastic wadding and shell fragments recovered from Dreher’s body were

consistent with material contained within the cartridges that were loaded in the shotgun at

the time it was recovered on March 21; however, she could not conclude that the recovered

shotgun was the murder weapon.

Finally, the State presented evidence of incriminating statements McGhee made to

police officers on April 5, 2007. Detective Delaney testified that, in the course of McGhee’s

swabbing for DNA on that date, he asked McGhee if he had an attorney. According to

Delaney, McGhee said his family would be wasting their money on an attorney because

the police had the “pump.” Delaney understood this to mean a sawed-off shotgun.

Additionally, Delaney testified that McGhee asked about cooperating with the

investigation and that Delaney said he could not make any promises.

13
 McGhee presented an alibi witness, Dayontae Duncan, who testified that McGhee

was with him at a party on the evening of the shooting and that they spent the night at

Duncan’s home. McGhee also testified that he was with Duncan at the time of the shooting

and was not at the Papa Johns restaurant that night. He testified that he ran from the police

on March 21, 2007, because police had “been coming around our neighborhood harassing

young dudes because of the colors we wear,” and he suspected police thought he was

affiliated with a gang. He admitted telling police at the time of the DNA swabbing that he

thought his family was hiring an attorney, but he denied saying the attorney was a waste of

money because the police had “the pump,” and he denied making any other statements to

police during that interaction.

The jury found McGhee guilty of first-degree murder, and he was sentenced to life

in prison on January 18, 2008. The Court of Special Appeals affirmed McGhee’s

conviction on direct appeal, McGhee v. State, Sept. Term, 2007, No. 2827, and this Court

denied McGhee’s petition for certiorari. McGhee v. State, 410 Md. 561 (2009).

C. Post-Conviction Proceeding

On June 10, 2014, McGhee filed a pro se petition for post-conviction relief. Defense

counsel later supplemented this petition with additional issues. McGhee contended that he

received ineffective assistance of counsel based on:

1. trial counsel’s failure to object to the trial court’s “CSI-effect” voir
dire question;
2. appellate counsel’s failure to raise the preserved issue that the trial
court erred in admitting testimony about inconclusive forensic
evidence; and
3. trial counsel’s failure to object to incomplete or missing jury
instructions.

14
 The post-conviction court held a hearing, at which McGhee’s trial counsel testified.

She explained that “[t]he defense in this case was that Mr. McGhee was not the shooter in

the incident, that he was not present at the place and time in which the decedent was killed.”

Trial counsel did not recall the CSI-effect voir dire question.

On June 11, 2020, the post-conviction court granted McGhee’s petition for

post-conviction relief and ordered a new trial. The court stated that “[a]lthough at the time

of [McGhee’s] trial, the CSI question was allowed, the Court believes the question must

still be analyzed based on its prejudicial effect on the jury.” The post-conviction court

found that, in this case, the lack of scientific evidence was material to the question of

McGhee’s guilt. For this reason, the court held that the failure to object to the CSI-effect

question constituted ineffective assistance of counsel. The post-conviction court also

determined that McGhee’s trial counsel provided ineffective assistance by failing to object

to certain incomplete or missing jury instructions. The court held that McGhee’s appellate

counsel was not ineffective.

The Court of Special Appeals reversed the post-conviction court’s decision with

respect to McGhee’s trial counsel in an unreported opinion. State v. McGhee, Sept. Term

2020, No. 638 (Md. Ct. Spec. App. Nov. 30, 2021). The intermediate appellate court

reasoned:

As in Armstead, appellee presented no evidence establishing that the
prevailing professional norm at the time of his trial was to object to “CSI
effect” messages to the venire or jury. Consequently, we conclude that trial
counsel’s failure to object to the trial court’s “CSI effect” voir dire question
in this case was not deficient performance.

15
Slip op. at 12. The Court of Special Appeals also held that the post-conviction court erred

in alternatively granting relief to McGhee based on trial counsel’s failure to object to

missing or incomplete jury instructions. Id. at 13-26.

McGhee petitioned this Court for a writ of certiorari, which this Court granted on

March 8, 2022. McGhee v. State, 478 Md. 187 (2022). McGhee presents the following

questions for our review, which we have rephrased slightly:4

1. Do Charles v. State, 414 Md. 726 (2010), Atkins v. State, 421 Md. 434
(2011), and Stabb v. State, 423 Md. 454 (2011), which concern the
propriety of CSI-effect voir dire questions and jury instructions, apply to
cases that became final before those decisions were issued?

2. Did McGhee’s trial counsel render ineffective assistance of counsel
when, at McGhee’s 2007 trial, she failed to object to a CSI-effect voir
dire question?

II

Standard of Review

“This Court reviews ‘a post-conviction court’s findings regarding ineffective

assistance of counsel [a]s a mixed question of law and fact. The factual findings of the

post-conviction court are reviewed for clear error. The legal conclusions, however, are

reviewed de novo.’” Wallace v. State, 475 Md. 639, 653 (2021) (quoting State v. Syed, 463

Md. 60, 73 (2019)).

4
McGhee did not seek further review of the portion of the Court of Special Appeals’
opinion which held that trial counsel’s failure to object with respect to several jury
instructions did not constitute ineffective assistance of counsel. Therefore, we consider
only whether counsel’s failure to object to the CSI-effect voir dire question violated
McGhee’s right to effective assistance of counsel.

16
 III

Discussion

The Sixth Amendment to the United States Constitution and Article 21 of the

Maryland Declaration of Rights guarantee criminal defendants the right to “effective

assistance of counsel.” Strickland v. Washington, 466 U.S. at 686 (internal quotation marks

and citation omitted); Taylor v. State, 428 Md. 386, 399 (2012). To meet their burden on

an ineffective assistance of counsel claim, a petitioner must make two showings. First, the

petitioner “must show that counsel’s performance was deficient. This requires showing that

counsel made errors so serious that counsel was not functioning as the ‘counsel’ guaranteed

the defendant by the Sixth Amendment. Second, the defendant must show that the deficient

performance prejudiced the defense.” Taylor, 428 Md. at 399 (quoting Strickland, 466 U.S.

at 687).

Under the first, “performance” prong, a defendant must show that “counsel’s alleged

acts or omissions, … ‘viewed as of the time of counsel’s conduct,’ fell ‘outside the wide

range of professionally competent assistance.’” Taylor, 428 Md. at 399 (quoting Strickland,

466 U.S. at 690). Under the second, “prejudice” prong, a defendant must show “that there

is a reasonable probability that, but for counsel’s unprofessional errors, the result of the

proceeding would have been different.” Id. at 399-400 (quoting Strickland, 466 U.S. at

694).

McGhee argues that his trial counsel’s failure to object to a CSI-effect voir dire

question violated both the performance prong and prejudice prong of the Strickland test.

McGhee contends that Charles, Atkins, and Stabb apply retroactively to cases on collateral

17
review. According to McGhee, with the benefit of the Atkins/Charles/Stabb trilogy of

CSI-effect cases, it is clear both that his counsel’s failure to object to the CSI-effect voir

dire question was deficient performance and that this error prejudiced him because an

appellate court would have reversed his conviction under Charles, Atkins, and Stabb.

McGhee alternatively argues that, even if the Court considers only the CSI-effect case law

that existed at the time of his trial – i.e., Evans – his counsel should have extrapolated from

Evans that it was necessary to object to the CSI-effect voir dire question that the trial court

put to the potential jurors.

The State argues that McGhee’s retroactivity argument is incompatible with an

analysis of the performance prong under Strickland, which requires a reviewing court to

consider counsel’s acts or omissions at the time of the contested conduct, not with the

benefit of hindsight. According to the State, that analysis leads to the conclusion that

McGhee’s trial counsel acted consistently with the professional norms that existed at the

time of McGhee’s trial in late 2007.

We agree with the State.

A. The Issue of Retroactivity

The retroactivity doctrine in criminal law considers whether a change in law should

be applied to a conviction that occurred before the change. See Wiggins v. State, 275 Md.

689, 698-99 (1975) (describing the “genesis of the modern retroactivity doctrine”). When

confronted with a question of whether a new case applies retroactively, this Court’s

precedent considers whether a new decision sets out a “new principle of law.” State v.

Daughtry, 419 Md. 35, 78 (2011) (quoting Houghton v. Cnty. Comm’rs of Kent Cnty., 307

18
Md. 216, 220 (1986)); Denisyuk v. State, 422 Md. 462 (2011), abrogated on other grounds

by Miller v. State, 435 Md. 174 (2013). When the decision announces a new principle of

law, it does not necessarily apply retroactively. See Daughtry, 419 Md. at 78. In contrast,

“where a decision has applied settled precedent to new and different factual situations, the

decision always applies retroactively.” Potts v. State, 300 Md. 567, 577 (1984) (citation

omitted).

The Supreme Court has not explicitly addressed whether the principle of

retroactivity may be applied within an ineffective assistance of counsel claim, such that a

new rule of law would render counsel’s conduct defective even if the conduct was

reasonable and appropriate at the time. McGhee asks this Court to dive into the substance

of this analysis and determine whether Charles, Atkins, and Stabb applied settled precedent

to new facts or announced a new principle of law. Thus, McGhee implicitly asks this Court

to conclude that such an application of retroactivity principles is compatible with a

Strickland analysis. For the reasons set out below, we conclude it is not.

Strickland’s performance prong requires this Court to assess trial counsel’s

performance based on prevailing professional norms at the time of the contested conduct.

Strickland, 466 U.S. at 689-90. At the outset, this appears incompatible with the notion of

retroactivity.

McGhee disclaims that he is asking this Court to judge his trial counsel’s conduct

based on future case law. Instead, McGhee suggests the practical effect of applying these

cases retroactively in an ineffective assistance of counsel claim is only that a

post-conviction court can assume the direct appeal would have been successful. Applied

19
here, McGhee argues, the Court of Special Appeals on direct review of McGhee’s

conviction would have held that the failure to object to a CSI-effect voir dire question was

reversible error, had it had the benefit of Charles, Atkins, and Stabb. It necessarily follows

(according to McGhee) that his counsel provided constitutionally deficient performance.

McGhee’s argument thus verges into the second Strickland prong, the prejudice

prong. McGhee argues this is appropriate when evaluating trial counsel’s failure to

preserve an issue for appeal. In such a situation, the “performance and prejudice prongs

naturally overlap because the questions of whether counsel’s performance was adequate

and whether it prejudiced the petitioner both will turn on the viability of the omitted

claims.” Gross v. State, 371 Md. 334, 350 (2002).

However, it does not follow from Gross that we consider “the viability of the

omitted claims” based on case law that did not exist at the time of the attorney’s challenged

performance. To the contrary, doing so is inconsistent with Strickland. McGhee has

identified no post-Strickland precedent of the Supreme Court or this Court that would

countenance such an application of the Strickland test.

The Supreme Court’s decision in Lockhart v. Fretwell, 506 U.S. 364 (1993), informs

our analysis. In Fretwell, a defendant received a death sentence based on an aggravating

circumstance (murder committed for pecuniary gain), id. at 366, which was improper under

then-existing law because the aggravating circumstance was also an element of the

underlying crime. Collins v. Lockhart, 754 F.2d 258, 265 (8th Cir. 1985) (holding such

“double counting” was impermissible), overruled in Perry v. Lockhart, 871 F.2d 1384 (8th

Cir. 1989). By the time the defendant brought his habeas corpus petition alleging

20
ineffective assistance of counsel, Collins had been overruled by Perry. Fretwell, 506 U.S.

at 368. Thus, the Supreme Court considered which law the federal district court should

have applied to an ineffective assistance of counsel claim on collateral review: the law as

it existed at the time of Fretwell’s trial (Collins) or the law as it existed when Fretwell’s

post-conviction petition was filed (Perry). Notably, the parties had agreed that trial

counsel’s conduct was deficient under Strickland’s performance prong, so the Court’s

inquiry was focused solely on the prejudice prong. Id. at 369 n.1.

The Supreme Court held that defense counsel’s failure to object to the pecuniary

gain aggravating factor did not prejudice the defendant, even if the objection would have

been meritorious at the time of the trial and direct appeal. The Court distinguished between

the performance and prejudice prongs of the Strickland test for purposes of its analysis:

Respondent argues that the use of hindsight is inappropriate in
determining “prejudice” under Strickland, and that this element should be
determined under the laws existing at the time of trial. For support, he relies
upon language used in Strickland in discussing the first part of the necessary
showing—deficient performance. We held that in order to determine whether
counsel performed below the level expected from a reasonably competent
attorney, it is necessary to “judge ... counsel’s challenged conduct on the
facts of the particular case, viewed as of the time of counsel’s conduct.”
Strickland, 466 U.S. at 690, 104 S.Ct. at 2066.

Ineffective-assistance-of-counsel claims will be raised only in those
cases where a defendant has been found guilty of the offense charged, and
from the perspective of hindsight there is a natural tendency to speculate as
to whether a different trial strategy might have been more successful. We
adopted the rule of contemporary assessment of counsel’s conduct because a
more rigid requirement “could dampen the ardor and impair the
independence of defense counsel, discourage the acceptance of assigned
cases, and undermine the trust between attorney and client.” Ibid. But the
“prejudice” component of the Strickland test does not implicate these
concerns. It focuses on the question whether counsel’s deficient performance
renders the result of the trial unreliable or the proceeding fundamentally

21
 unfair. Unreliability or unfairness does not result if the ineffectiveness of
counsel does not deprive the defendant of any substantive or procedural right
to which the law entitles him. As we have noted, it was the premise of our
grant in this case that Perry was correctly decided, i.e., that respondent was
not entitled to an objection based on “double counting.” Respondent
therefore suffered no prejudice from his counsel’s deficient performance.

Id. at 371-72 (citation omitted). The Fretwell Court’s reiteration of the importance of

contemporaneous assessment of counsel’s conduct under the performance prong is

significant. See also Williams v. Taylor, 529 U.S. 362, 391 (2000) (observing that Fretwell

did not “modif[y] or in some way supplant[] the rule set down in Strickland”).

McGhee relies on dicta in Allen v. State, 204 Md. App. 701 (2012), to support the

applicability of a retroactivity analysis in this context. The Allen Court considered whether

the Atkins and Stabb holdings applied to cases that were pending on direct appeal at the

time Atkins and Stabb were decided. Id. at 706. The intermediate appellate court answered

this question affirmatively. Id. at 721-22. McGhee focuses on dicta in Allen which

suggested – based on Denisyuk – that Atkins and Stabb also would apply retroactively to

cases on collateral review:

In Denisyuk, a post conviction appeal, the Court of Appeals concluded
that the Supreme Court’s decision in Padilla v. Kentucky, ––– U.S. –––, 130
S.Ct. 1473, 176 L.Ed.2d 284 (2010), that defense counsel is constitutionally
ineffective by failing to advise a criminal defendant that deportation is a
likely consequence of a guilty pleas, applied settled principles to new factual
situations, and thus, it applied to convictions that were final. 422 Md. at
481-82, 30 A.3d 914…. The Denisyuk analysis and result appears to be
applicable to Atkins and Stabb.

Id. at 723.

We disagree with this dictum in Allen to the extent it might be thought to apply to

an ineffective assistance of counsel claim. The potential application of Charles, Atkins, and

22
Stabb retroactively is inconsistent with a proper analysis under Strickland’s performance

prong. Adopting McGhee’s argument, in practice, would require attorneys not only to

comply with prevailing professional norms of the time, but also to predict future

developments in case law. Neither the Sixth Amendment of the United States Constitution

nor Article 21 of the Maryland Declaration of Rights imposes such a requirement on

counsel. See, e.g., Strickland, 466 U.S. at 690; State v. Calhoun, 306 Md. 692, 735 (1986)

(“There was no duty on counsel to foresee that we might hold as we held in [future]

case[s].”); State v. Davis, 249 Md. App. 217, 230 (2021) (post-conviction court erred in

applying this Court’s 2014 holding in Pearson v. State, 437 Md. 350 (2014), retroactively

when assessing trial counsel’s performance in 2007); Toledo v. United States, 581 F.3d

678, 681 (8th Cir. 2009) (“Counsel is not accountable for unknown future changes in the

law.”). We do not read Denisyuk as holding otherwise.5 Accordingly, we decline McGhee’s

request to apply Charles, Atkins, and Stabb in assessing the performance prong of

McGhee’s ineffective assistance of counsel claim.6

5
Notably, this Court abrogated Denisyuk in Miller v. State, 435 Md. 174 (2013),
following the Supreme Court’s holding in Chaidez v. United States, 568 U.S. 342 (2013),
that Padilla v. Kentucky does not apply retroactively to cases on collateral review. In any
event, we agree with the analysis of Denisyuk that the State provided in its brief: “The issue
in Denisyuk was not that defense counsel should have predicted Padilla’s holding about
what constitutes ineffective assistance – it was that defense counsel should have known
that Denisyuk’s plea would subject him to deportation under statutory law that was
perfectly clear at the time of the plea…. [Denisyuk] does not involve whether counsel is
ineffective for failing to predict future legal developments.”
6
Because we conclude that a retroactivity analysis is inapplicable in this context,
we do not decide whether Charles, Atkins, and Stabb articulated a new principle of law or
applied settled precedent to new facts.

23
 B. Application of the Strickland Test to Trial Counsel’s Failure to Object to the
CSI-Effect Voir Dire Question at McGhee’s Trial

The right to effective assistance of counsel aims to ensure a fair trial. As such, the

“benchmark for judging any claim of ineffectiveness must be whether counsel’s conduct

so undermined the proper functioning of the adversarial process that the trial cannot be

relied on as having produced a just result.” Strickland, 466 U.S. at 686.

The petitioner bears the burden to overcome the presumption of effective assistance

of counsel by satisfying both prongs of the Strickland test. Failure to prove either of the

two Strickland prongs is fatal to the claim. State v. Syed, 463 Md. 60, 75 (2019) (citing

Strickland, 466 U.S. at 687). As explained below, we hold that McGhee failed to meet his

burden to prove that his trial counsel’s failure to object to a CSI-effect voir dire question

in 2007 fell below an objective standard of reasonableness. Because McGhee has not

satisfied the first prong of the Strickland test, we decline to address the question of

prejudice.

When scrutinizing counsel’s conduct, “a court must indulge a strong presumption

that counsel’s conduct falls within the wide range of reasonable professional assistance.”

Strickland, 466 U.S. at 689. A “fair assessment of attorney performance requires” a court

to “reconstruct the circumstances of counsel’s challenged conduct, and to evaluate the

conduct from counsel’s perspective at the time,” considering all the surrounding

circumstances. Newton v. State, 455 Md. 341, 355 (2017) (quoting Strickland, 466 U.S. at

689). In conducting this analysis, the court “must be highly deferential” and take care to

avoid “the distorting effects of hindsight.” Strickland, 466 U.S. at 689.

24
 As McGhee bears the burden to overcome the presumption of reasonableness, we

now consider the arguments and evidence he presented: (1) that any reasonable attorney

should have extrapolated from Evans the need to object to a CSI-effect voir dire question;

(2) that McGhee’s trial counsel acted contrary to her express trial strategy; and (3) that

other objective indicia indicate counsel’s conduct was objectively unreasonable. We

conclude that, under the totality of the circumstances, McGhee has failed to meet his

burden to overcome the presumption of effective assistance of counsel.

1. Case Law in 2007

First, McGhee directs our attention to the only reported opinion at the time of

McGhee’s trial in 2007, Evans v. State.7 As discussed above, in Evans the Court of Special

Appeals in dicta discerned no error in the trial court’s instruction to the jury that “there is

no legal requirement that the State utilize any specific investigative technique or scientific

test to prove its case.” Evans, 174 Md. App. at 562, 570. This opinion, notably, did not use

the phrase “CSI effect.” Yet McGhee contends that any reasonable attorney should have

extrapolated from Evans that they were required to object to a CSI-effect voir dire question.

We disagree.

In order to satisfy McGhee’s standard of objectively reasonable representation, an

attorney in December 2007 would have to have been familiar with the budding controversy

surrounding CSI-effect jury instructions. They would have to have been aware of the one

Court of Special Appeals opinion from May 2007 that touched on the CSI effect, but

7
The Court of Special Appeals issued its opinion in Evans on May 3, 2007.
McGhee’s trial occurred in December 2007.

25
without referring to it as the “CSI effect.” Assuming they were familiar with the Evans

opinion, the attorney would have to have looked past the fact that the Evans Court upheld

the CSI-effect jury instruction in that case. They also would have to have looked past the

fact that the intermediate appellate court’s discussion of this issue was dicta. From there,

they would have to have extrapolated that dicta related to jury instructions also applied to

voir dire questions. We do not believe it is reasonable to expect an attorney in December

2007 to have made all of these connections.

However, McGhee notes that defense counsel in Charles objected to a similar voir

dire question in a trial that occurred at approximately the same time as McGhee’s trial.

McGhee contends that his trial counsel therefore should have known to object as well.

While Charles’s attorney apparently had the knowledge and foresight in December 2007

to object to a CSI-effect voir dire question, we are not prepared to say that dicta in Evans

set a “prevailing professional norm,” such that the failure to object rendered McGhee’s

counsel’s performance constitutionally deficient in late 2007. As the Court of Special

Appeals noted in Armstead, merely showing that some attorneys were objecting and others

were not does not establish a prevailing professional norm. See Armstead, 235 Md. App.

at 422-23.

The logical conclusion of McGhee’s argument is that any attorney who failed to

object to a CSI-effect message between the time Evans was decided in 2007 and 2010,

when this Court decided Atkins, provided constitutionally deficient representation.

McGhee attempts to avoid this “floodgate” effect by pointing this Court to McGhee’s

26
post-conviction hearing, in which his trial counsel testified about her trial strategy. We

consider this argument next.

2. Trial Counsel’s Strategy

During the post-conviction hearing, McGhee’s trial counsel testified that the defense

at trial was that “Mr. McGhee was not the shooter in the incident, that he was not present

at the place and time in which the decedent was killed.” McGhee contends it was clear that

a voir dire question that insinuated the State need not present forensic evidence to obtain a

conviction undermined his defense counsel’s express trial strategy. Thus, McGhee argues,

even if it was not generally a professional norm at the time to object to a CSI-effect voir

dire question, his attorney should have recognized that the CSI-effect voir dire question the

trial court put to McGhee’s potential jurors might prejudice McGhee.

However, trial counsel’s theory of the case that McGhee was not the perpetrator

does not change matters. Trial counsel’s testimony at the post-conviction hearing does not

overcome the presumption that her conduct fell within “the wide range of professionally

competent assistance.” Strickland, 466 U.S. at 689. As discussed above, Evans would have

led a reasonably competent attorney to believe that an objection to the voir dire question at

issue would not have been meritorious. See Armstead, 235 Md. App. at 406 (explaining

that, at the time of Armstead’s trial in 2009, CSI-effect voir dire was permissible, and

arguably even “favored,” under Maryland law).8

8
At oral argument, McGhee contended that counsel’s objection to a jury instruction
concerning specific investigative techniques demonstrates that she knew or should have
known CSI-effect messages were problematic in voir dire. However, counsel objected to
the specific investigative techniques jury instruction not based on the contention that it

27
 In response, McGhee argues that, separate and apart from Evans, other objective

indicia lead to the conclusion that a reasonably competent attorney in 2007 should have

known to object to a CSI-effect voir dire question. We now consider that contention.

3. Other Objective Indicia

Although there were no Maryland cases in 2007 that would have suggested that

CSI-effect voir dire questions were inappropriate, McGhee may attempt to meet his burden

to show that his counsel’s performance violated professional norms by directing our

attention to other sources, such as news articles and scholarly articles, to the extent such

sources shed light on the general consciousness of defense counsel in 2007. In Charles,

this Court conducted a review of the state of literature and case law related to the CSI effect

in and around 2007. 414 Md. at 731-32. Of the articles the Charles Court located, only one

law review article and two news articles were published before McGhee’s trial. See Cynthia

Di Pasquale, Beyond the Smoking Gun, THE DAILY REC. (Sept. 7, 2006), available at

https://perma.cc/Y2G8-LX52 (noting a potential danger to prosecutors, while observing

that “[d]efense attorneys don’t seem to know what all the fuss is about”); Andrew P.

Thomas, The CSI Effect: Fact or Fiction, 115 YALE L.J. POCKET PART 70 (2006)

(advocating that “judges acknowledge the CSI effect and take steps during voir dire to

prevent biased jurors from improperly influencing the jury”); Jeffrey Toobin, The CSI

would relieve the State of its burden of proof, but rather because the trial court had ruled
that the State would be permitted to introduce expert testimony concerning DNA evidence,
and therefore the instruction was not warranted under the circumstances.

28
Effect: The Truth About Forensic Science, The New Yorker (May 7, 2007), available at

https://perma.cc/2NCV-U2EJ (criticizing forensic inductive sciences in general).

These articles do little to advance McGhee’s argument. First, they show that

Maryland defense attorneys in 2006 did not view the CSI effect as a problem, as they did

not “seem to know what all the fuss [was] about.” Second, if the CSI effect was generally

known to any extent, it was understood to be a phenomenon that harms prosecutors, not

defendants.

The Court of Special Appeals concluded in Armstead that it was not a prevailing

professional norm at the time of Armstead’s trial in March 2009 to object to CSI-effect

voir dire questions. We find Armstead’s reasoning on this point to be persuasive. It follows

that it was not a professional norm to object to a CSI-effect voir dire question when

McGhee went to trial in December 2007, 15 months earlier.

In sum, we conclude that McGhee’s counsel performed within the accepted

professional norms of 2007, which did not require objecting to the CSI-effect voir dire

question in this case. Thus, McGhee’s claim of ineffective assistance of counsel fails on

the performance prong, and we need not address McGhee’s arguments concerning the

prejudice prong.

IV

Conclusion

Because McGhee’s trial occurred in December 2007, we do not assess his counsel’s

performance under Charles, Atkins, and Stabb, which were all decided in the years that

followed. McGhee failed to demonstrate that, under the professional norms that existed at

29
the time of his trial, his attorney provided constitutionally deficient representation by

failing to object to a CSI-effect voir dire question. Thus, the post-conviction court erred in

determining that McGhee received ineffective assistance of counsel.

JUDGMENT OF THE COURT OF
SPECIAL APPEALS AFFIRMED; COSTS
TO BE PAID BY PETITIONER.

30