Opinion of the Court
SULLIVAN, Judge:Appellant was tried by a general court-martial composed of officer and enlisted members1 at Dyess Air Force Base, Texas, during August and September 1992. Despite his pleas of not guilty, he stands guilty before this Court of attempting to kidnap a 13-year-old female and taking indecent liberties with her, in violation of Articles 80 and 134, Uniform Code of Military Justice, 10 USC §§ 880 and 934, respectively. He was sentenced to a bad-conduct discharge, 2 years’ confinement, total forfeitures, and reduction to Airman Basic. On December 23,1992, the convening authority approved this sentence. The Court of Military Review2 affirmed on April 29,1994. 40MJ533.
On April 11, 1995, we granted review on the following Issue raised by appellate defense counsel and an issue modified by this Court:
I
WHETHER THE AIR FORCE COURT OF MILITARY REVIEW ERRED IN APPLYING AN INCORRECT STANDARD OF REVIEW AND IN FAILING TO APPLY THE CORRECT STANDARD OF CHAPMAN v. CALIFORNIA: WHETHER APPELLANT WAS PREJUDICED BEYOND A REASONABLE DOUBT BY AN ERROR OF CONSTITUTIONAL MAGNITUDE.
II (modified)
WHETHER THE AIR FORCE COURT OF MILITARY REVIEW WAS ERRONEOUS IN RULING THE MILITARY JUDGE ERRED IN EXCLUDING EXPERT EYEWITNESS TESTIMONY; AND IF THE EXCLUSION WAS ERROR, WAS THE COURT OF MILITARY REVIEW ERRONEOUS IN HOLDING IT HARMLESS.
On further consideration, this Court on May 17, 1995, vacated the order dated April 11, 1995. On June 14, 1995, this Court again granted review of the two Issues previously granted and specified an issue as follows:
WHETHER THIS COURT MAY CONSIDER, ON A PETITION FOR GRANT OF REVIEW, THE CORRECTNESS OF THE PORTION OF THE DECISION BY THE COURT OF MILITARY REVIEW FAVORABLE TO THE ACCUSED WHEN NO CERTIFICATE FOR REVIEW HAS BEEN FILED RAISING THAT ISSUE.
We hold that any error committed by the military judge in excluding the preferred defense evidence was, in light of the entire record of trial, harmless beyond a reasonable doubt. See United States v. Wilson, 20 MJ 31, 32 (CMA 1985) (harmless-beyond-a-reasonable-doubt standard applied in abundance of caution).3
The following facts are set forth by the Court of Military Review in its opinion:
On March 6, 1992, in Wichita Falls, Texas, a man in a small car accosted LMJ, who was walking home from school. After exposing himself and making indecent comments to her, he jumped from the car and tried to force her into it. Fortunately, she broke free and ran a short distance to her home where she reported the attack to her parents.
LMJ and her father, AJ, immediately went looking for the attacker in the family car. LMJ told AJ that the man drove a small blue car with tinted windows and the *29word “BUG” written in white across the windshield. She added that the car was like the one they used to own, which AJ recalled was a Volkswagen “bug.” They were unsuccessful in their quick search and returned home. Leaving LMJ to await the police, AJ resumed the search. About a block from his home, he saw a Volkswagen that fit the description with a license plate number that he recorded as “355 ONE.” He saw the male driver but lost the trail when the driver made a quick U-turn.
Back at home, LMJ and her mother, CSJ, stood in the front yard waiting for the police when LMJ saw a small blue Volkswagen with tinted windows and the work “BUG” written in white across the windshield drive by. LMJ told CSJ, “Oh my gosh, Mom, he’s coming.” CSJ saw the driver and wrote the license plate number down as “355 ONE.” The male driver waved and smiled as he drove'by.
LMJ described her attacker for the police as a short white male, possibly tanned, with bushy, collar-length dark hair. Appellant is Hispanic, wore a mustache at the time, and, like most male military members, did not have bushy hair. However, the police investigation focused on appellant because he drove a small blue Volkswagen “bug” with tinted windows, the word “BUG” written in white across the windshield, and license number “355 ONE.”
LMJ could not pick her attacker out of a photo array which included appellant’s picture. However, she picked both appellant and another male out of a subsequent physical lineup even though appellant had shaved his mustache after the photo array. LMJ’s parents also failed to pick appellant out of the photo array, but they recognized appellant at the lineup. However, the lineup deviated from standard investigative procedure because the investigators did not use the same individuals from the photo array — only appellant and one other person LMJ picked from the lineup were also pictured in the photo array.
At trial LMJ could not identify appellant as her attacker. Her parents, on the other hand, were certain appellant was the driver of the Volkswagen they saw in the neighborhood.
Appellant admitted he drove in LMJ’s neighborhood on March 6th but denied committing the offenses or trying to elude AJ. Appellant also could not recall driving by LMJ’s house.
40 MJ at 53<b-35.
• Sergeant Garcia’s defense at trial was one of mistaken identity. To support this defense, he offered the testimony of Dr. Curtis Wills, a board-certified forensic psychologist and consultant to several law enforcement agencies, who was an expert in eyewitness identification. In a session under Article 39(a), UCMJ, 10 USC § 839(a), conducted without members, the defense proffered Dr. Wills’ testimony. Appellant in his final brief notes the intended scope of direct examination, which he proffered at the Article 39(a) session, as follows:
The appellant proffered testimony from Dr. Wills regarding the stages of memory, R. 465, the effect of stress on these stages, R. 466, the accuracy of recollection of startling events, R. 468, the effect of questioning by authority figures, R. 471, the phenomenon of insertion into the memory by post-event questioning, R. 469, the reliability of lineup procedures, R 473, and the correlation between the accuracy and confidence of a memory, R. 475.
Final Brief at 8.
At trial Dr. Wills testified that stress has a negative impact on the accuracy of eyewitness identification. Moreover, the military judge allowed him to testify that suggestive questioning can impact eyewitness identification and that questioning by authority figures such as parents can affect one’s responses. Relying on Mil.R.Evid. 403 and 702, Manual for Courts-Martial, United States, 1984, the judge did not allow the expert to testify about the “inverse” relationship between confidence and accuracy in identification and the theory of memory transference and transposition. See Appendix. Later, defense counsel made a motion for the judge to recuse herself because of her *30rulings on defense evidence. That motion was denied.
The Court of Military Review found the military judge erred in excluding defense evidence in this case. It then determined that appellant suffered no prejudice from these erroneous rulings. Judge Pearson said:
PREJUDICE
Even though we determine the military judge abused her discretion in excluding relevant portions of the expert’s testimony and restricting the rest, we do not grant relief unless we find prejudice. Article 59(a), UCMJ, 10 USC § 859(a) (1988); Combs, 35 MJ at 827; United States v. Sawyer, 32 MJ 917 (AFCMR 1991); see United States v. Weeks, 20 MJ 22 (CMA 1985). We do not find it here.
In assessing prejudice from excluded expert testimony, we believe the two most important questions to answer are: (1) Was appellant able to adequately present the issue to the members in some other suitable form? and (2) Was the Government’s case strong and conclusive when weighed against the defense theory of the case? See Weeks, 20 MJ at 25.
We start our review at defense counsel’s voir dire. Defense counsel expertly questioned the members extensively about identification evidence, tailoring questions to the facts of the ease. For example, the members agreed “that if somebody is stressed or frightened, their memory and their perceptions may not be very strong.” While. voir dire questions and answers aren’t evidence in the case, the questions and answers show the members were amenable to the defense theory of the ease. In presenting that theory, Dr. Wills did testify in general terms about most of the areas in issue except the suggestiveness of the lineup which counsel fully explored through cross-examination of the investigator who conducted it. In closing argument, defense counsel honed the issue of the reliability of the eyewitness’ identifications for the members, including how LMJ and her parents had improved on their identifications over time. Consequently, we conclude appellant was able to adequately present the issue in spite of the military judge’s ruling.
As for the quality of the Government’s case, we find the evidence establishing appellant’s guilt overwhelming. See Moore, 786 F.2d at 1313; Smith, 736 F.2d at 1107-1108. In this regard, we will not recount appellant’s testimony other than to say he was his own worst enemy on the witness stand. Appellant also made several incriminating admissions to a Wichita Falls detective during a telephone conversation and to an Office of Special Investigations special agent during an interview. He told the detective, “nobody drives his Volkswagen but him,” and appeared to know facts about the incident only the perpetrator would know. He admitted to the agent that he “might have stopped and talked to a girl” in LMJ’s neighborhood on the date of the offenses, and he “stopped for extended periods of time near [her] school.”
40 MJ at 538-39.
Appellant’s basic contention on this appeal is that the Court of Military Review applied the wrong test for determining whether error by the military judge in excluding certain defense evidence was prejudicial. He asserts that the appellate court below found an error of constitutional magnitude in this case, i.e., a violation of his due process right to present a defense under the Fifth Amendment. He then argues that the court below erred in applying the decision of this Court in United States v. Weeks, 20 MJ at 25, to determine prejudice arising from such a constitutional error. Finally, he contends that even if the four-factor test of Weeks for finding prejudice is appropriate in constitutional error eases, the Court of Military Review did not properly apply all four factors of that test in his case.
We initially note that the Court of Military Review did not hold that constitutional error occurred in this case. See United States v. Lane, 474 U.S. 438, 446 nn. 8 & 9, 106 S.Ct. 725, 730 nn. 8 & 9, 88 L.Ed.2d 814 *31(1986). Instead, it held that the military judge violated several rules of evidence by restricting Doctor Wills’ expert testimony (ie., Mil.R.Evid. 403 and 702-04). 40 MJ at 537-38. Then, it assessed the impact of the judge’s errors in light of other evidence in this case and found no prejudice. Id. at 538-39. This is standard nonconstitutional evidentiary analysis. Mil.R.Evid. 103(a); United States v. Wimberly, 60 F.3d 281, 286 (7th Cir.1995), cert. denied, — U.S. —, 116 S.Ct. 744, 133 L.Ed.2d 693 (1996); United States v. Hanson, 994 F.2d 403, 407 (7th Cir.1993); United States v. Hogan, 886 F.2d 1497, 1512 (7th Cir.1989); United States v. Peak, 856 F.2d 825, 834-35 (7th Cir.), cert. denied, 488 U.S. 969, 109 S.Ct. 499, 102 L.Ed.2d 535 (1988). Accordingly, we reject appellant’s initial assumption that the Court of Military Review found constitutional error in this case.
We further hold that appellant was not denied his constitutional right to present his defense by the military judge’s evidentiary rulings limiting Dr. Wills’ testimony. Appellant does not challenge the legitimacy of Mil.R.Evid. 403 or 702, or the discretion they confer upon a military judge to admit or exclude evidence. See Rock v. Arkansas, 483 U.S. 44, 107 S.Ct. 2704, 97 L.Ed.2d 37 (1987); Chambers v. Mississippi 410 U.S. 284, 93 S.Ct. 1038, 35 L.Ed.2d 297 (1973); Washington v. Texas, 388 U.S. 14, 87 S.Ct. 1920, 18 L.Ed.2d 1019 (1967). Instead, he simply notes that the Court of Military Review found that the military judge abused her discretion in applying those rules to the excluded portions of Dr. Wills’ testimony. See United States v. Blum, 62 F.3d 63, 68 (2d Cir.1995); United States v. Stevens, 935 F.2d 1380, 1400-01 (3d Cir.1991); see also United States v. Abel 469 U.S. 45, 55, 105 S.Ct. 465, 470, 83 L.Ed.2d 450 (1984). It is well established, however, that not every error in the exclusion of defense evidence constitutes constitutional error. See United States v. Cruz-Jiminez, 977 F.2d 95, 100 (3d Cir.1992); Rosario v. Kuhlman, 839 F.2d 918, 924 (2d Cir.1988); Flanders v. Meachum, 824 F.Supp. 290, 299 n. 8 (D.Conn.1993), rev’d on other grounds, 13 F.3d 600 (2d Cir.1994). To establish constitutional error, exclusion of material or vital defense evidence must usually be shown.4 See Taylor v. Illinois, 484 U.S. 400, 409, 108 S.Ct. 646, 653, 98 L.Ed.2d 798 (1988); United States v. Valenzuela-Bernal 458 U.S. 858, 868-69, 102 S.Ct. 3440, 3446-48, 73 L.Ed.2d 1193 (1982); United States v. Robaina, 39 F.3d 858, 862 (8th Cir.1994).
For several reasons, we conclude that the defense was not denied material evidence in this case.
First, we note that Dr. Wills was allowed to testify in support of the defense theory of misidentification. He provided testimony to the members about the effect of stress on memory and its suggestibility as induced by questioning of authority figures such as parents.
Second, appellant’s cross-examination of a police officer revealed admittedly suggestive circumstances surrounding the photograph spread and the subsequent in-person lineup, both involving appellant. It sufficiently allowed him to present a tainted-identification argument to the members. Dr. Wills’ testimony on the theory of transference and transposition was more theoretical and general in nature, so it added little to the police officer’s more practical concessions.
Third, the confidence-accuracy relationship testimony of Doctor Wills did not have great value in this case. See United States v. Holmes, 44 F.3d 1150, 1157 (2d Cir.1995). The victim could not identify appellant as her assailant from the photo spread, at the subsequent line-up, or at trial. 40 MJ at 535. Moreover, her identification of his car as the assailant’s car and his race as Spanish could hardly be described as confident. She admitted that her identification of her assailant’s car as a Volkswagon came from her father. Moreover, her testimony that she *32had .previously said her assailant was “like a Mexican” was contradicted by the arresting police officer. Cf. United States v. Stevens, 935 F.2d at 1400 (witness identification at trial was exceedingly confident). In these circumstances, we hold that there was no reasonable likelihood that the excluded testimony of Doctor Wills would change the result in this case. See generally Kyles v. Whitley, — U.S. —, —, 115 S.Ct. 1555, 1566, 131 L.Ed.2d 490 (1995) (Material means “Treasonable probability of a different result.”); see also People v. Sanders, 11 Cal.4th 475, 46 Cal.Rptr.2d 751, 774, 905 P.2d 420, 443 (1995).
We further hold that the appellate court below did not err in affirming appellant’s case on the basis of the four-pronged test for prejudice articulated in United States v. Weeks, supra. That decision does not expressly state whether its trial-impact factors are appropriate for determining prejudice in cases involving errors of constitutional magnitude or just nonconstitutional errors. Cf. United States v. Vandelinder, 20 MJ 41, 47 (CMA 1985); United States v. Wilson, 20 MJ 31, 32 (CMA 1985) (harmless-beyond-a-reasonable-doubt standard applied in abundance of caution). In any event, evidentiary error is normally tested to see if it resulted in actual prejudice, i.e., substantial and injurious effect or influence in determining the jury’s verdict.5 Likewise, a determination of reversible constitutional error with respect to an accused’s right to present a defense requires at some point consideration of the entire record of trial. See Delaware v. Van Arsdall, 475 U.S. 673, 684, 106 S.Ct. 1431, 1438, 89 L.Ed.2d 674 (1986). Cf. Kyles v. Whitley, supra.6 Accordingly, the factors in Weeks, as considered and supplemented by the Court of Military Review, were an appropriate and necessary basis upon which to decide appellant’s ease. See United States v. Gagan, 43 MJ 200, 203 (1995); cf. United States v. Levy-Cordero, 67 F.3d 1002, 1013-15 (1st Cir.1995).
Finally, we hold that the Court of Military Review fully and fairly applied the Weeks test in appellant’s case. In United States v. Weeks, supra, then-judge Cox clearly articulated the factors to be considered in determining prejudice or harm from an erroneous ruling by the judge excluding defense evidence. He said:.
We now adopt the following four-pronged analysis to test for prejudice:
First: Is the Government’s case against the accused strong and conclusive? United States v. Lewis, 482 F.2d 632, 644 (D.C.Cir.1973).
Second: Is the defense’s theory of the case feeble or implausible? United States v. Lewis, supra at 646.
Third: What is the materiality of the proffered testimony? ...
Fourth: What is the quality of the proffered defense evidence and is there any substitute for, it in the record of trial?
20 MJ at 25 (footnote omitted). The opinion of the court below contains a finding on each of these factors adverse to appellant, although perhaps, out of order and in conjunctive fashion. Moreover, even assuming constitutional error occurred in this case, we are *33convinced that the findings of the court below clearly demonstrate such error was harmless beyond a reasonable doubt. See United States v. Wilson, supra.
The decision of the United States Air Force Court of Military Review is affirmed.
Chief Judge COX, Judges CRAWFORD and GIERKE, and Senior Judge DARDEN concur.APPENDIX
The military judge did not admit all of this evidence from Doctor Wills. While the members were deliberating on sentence, she stated her reasoning in limiting his testimony, as follows:
I wanted to put on the record the reasons and some of the findings that I made factually from my rulings regarding his testimony.
First I’d note, of course, that Military Rule of Evidence 702 provides that if scientific, technical, or specialized knowledge will assist the trier of fact to understand the evidence or determine a fact in issue, then you can allow expert testimony. Of course, there’s several questions. First, is the testimony helpful to the members. And then always you have the issue of is the testimony relevant. And I looked at both of those. I determined that the expert testimony regarding the effects of stress and the general principles of how perceptions are acquired, stored, and retrieved were relevant regarding eyewitness identification in this case. However, I felt the relevance was very limited in that I would restrict the testimony beyond those areas and that I would restrict it in those areas in a very limited manner to avoid some factors in balancing using Military Rule of Evidence 403. And I’m going to go over those factors.
So initially I did determine that on a limited basis it was relevant, balancing it with the conditions that are talked about in Rule 403, the dangers of unfair prejudice, confusion of the issues, misleading of the members, or by considerations of undue delay, waste of time, or needless presentation of cumulative evidence. Clearly determining the credibility or believability of each witness is within the sole province of the court members. The ability of each witness to observe and recall what they observed is amply presented by examination or cross-examination of the witnesses. This is true regarding all of the witnesses that we’ve heard in this case and in particular regarding my exclusion of Dr. Wills’ testimony regarding the photo and physical lineups and his testimony regarding suggestibility there. I find that counsel for both the Government and defense had the opportunity to thoroughly examine and cross-examine the testimony of the police officers who presented the photo lineup to the various witnesses and how she picked those and how she presented that and, as well, the testimony of the OSI agent who conducted the personal lineups in this case. And that testimony included how he selected the members to be included, what steps he took to avoid possible suggestibility, and, on cross-examination, the counsel went into what the Air Force regulation provides regarding such and the questions regarding its suggestibility and why one shouldn’t necessarily do a photo lineup as well as in-person lineup. So all those areas were covered by the cross-examination and examination of the special agent.
Now although Dr. Wills read and he testified in the 39(a) session, I would not allow him to testify in the hearing in front of the members — read the OSI report and police reports, those reports don’t fully detail the dynamics of what occurred during those presentations. The members had the photo lineup and they had it, of course, during their findings deliberations as well, which is Prosecution Exhibit 6. And they had the photos from the physical lineup. Those are Prosecution Exhibits 7 and 8. They had those as evidence to observe for themselves any suggestibility, as well as they were able to observe the witnesses who talked about what occurred during those lineups. As factfinders, that is clearly their duty to determine any suggestibility regarding those. Dr. Wills’ testimony *34in this area would therefore not only be cumulative, but confusing and misleading, therefore not admissible, since it is their duty as factfinders to look at the credibility of the witnesses and determine their credibility. And they could themselves observe the suggestibility of any photos or the lineups. Therefore that was in their province as factfinders and Dr. Wills’ testimony would therefore invade their province.
Dr. Wills’ expert testimony that there is an inverse correlation between the certainty of an eyewitness identification and its accuracy would also create a substantial danger of undue prejudice and confusion and have limited probative value. In examining — when you examine the credibility of witness instruction regarding eyewitness identification, it clearly instructs the members in part — and I’m going to read part of that — whether the witness had an adequate opportunity to observe the offender at the time of the offense will be affected by such matters as how long or short a time was available, how far or close the witness was, how good were lighting conditions, whether the witness had an occasion to see or know the person in the past. Such a thesis by Dr. Wills of this inverse correlation would only add confusion and be misleading to this instruction. The instruction clearly tells the members what facts or evidence they should consider in determining the credibility or believability of eyewitness identification and his testimony would not take into consideration these instructions and matters. I would also note — and I have examined the cases that the Government cited so I won’t recite them. I did examine U.S. v. Hicks [,7 MJ 561], which is the only military justice case I could find and I believe that is all counsel could find as well. And in there it noted — and that was a decision and a case that excluded all of the expert testimony. And it talks about Rules 702 and 403 and the balancing that the judge did in that case. It talked about the testimony of the expert would invade the province of the jury and would create a substantial danger of undue prejudice and confusion because of its aura of special reliability and trustworthiness and it would have limited probative value, directed as it is to the expert’s thesis rather than the actual eyewitness accounts. And in this case I would concur and my reasoning would be the same regarding that testimony as to his thesis of this inverse correlation. The limited evidence that I am allowing or did allow, since I’m adding and expanding and explaining my findings, but at the time I was going to allow Dr. Wills to testify regarding eyewitness identification to sufficiently educate the members in the possible pitfalls of eyewitness identification and the effects that such stress has on such identification. And by limiting it, they get that general background information, they’re educated, and then they get to evaluate the evidence — the eyewitness evidence or testimony, the testimony of the officers, as well, who were present during any lineups, photo or otherwise. And they can examine all of those witnesses to determine their credibility, suggestibility, and the believability in this case. Additionally, I would note that the Government was not notified of this expert witness until the first day of trial, and he was not available for them to interview until the next morning — early on Thursday morning. And I also considered that and that in terms of proceeding with the case that we’d started with members and the difficulty it might be for the Government to obtain their own expert. I considered all of that. However, that was not a major consideration, although I believe I’ve clearly outlined the factual basis for it. I did consider that. And, again, I noted that I had reviewed the cases. And that’s the basis upon which I made my decision.
(Emphasis added.)
. See Part I, United States v. McElroy, 40 MJ 368, 370-71 (CMA 1994).
. See 41 MJ 213, 229 n. * (1994).
. In view of our resolution of this case, it is not necessary to answer the specified issue.
. Our case law indicates that exclusion of certain types of evidence, i.e., defense character evidence (United States v. Gagan, 43 MJ 200, 203 (1995); United States v. Brown, 41 MJ 1, 4 (CMA 1994)) or victim-bias evidence (United States v. Bins, 43 MJ 79, 86 (1995)) can usually be considered constitutional error.
. An evidentiary error has not been considered harmless under this standard if it excluded the only or primary evidence in support of or in opposition to a claim or defense. See United States v. Cardascia, 951 F.2d 474, 489 (2d Cir. 1991); United States v. Peak, 856 F.2d 825, 834 (7th Cir.), cert. denied, 488 U.S. 969, 109 S.Ct. 499, 102 L.Ed.2d 535 (1988); United States v. Kohan, 806 F.2d 18, 22 (2d Cir. 1986). For the reasons stated by the Court of Military Review, 40 MJ 533, 538-39 (1994), we reject appellant’s argument that error in the military judge’s evidentiary rulings so prejudiced him in this case.
. Appellant suggests that a violation of his constitutional right to present a defense should be determined without regard to “outcome determinative” prejudice. See Delaware v. Van Arsdall, 475 U.S. 673, 680, 106 S.Ct. 1431, 1435-36, 89 L.Ed.2d 674 (1986); Davis v. Alaska, 415 U.S. 308, 318, 94 S.Ct. 1105, 1111, 39 L.Ed.2d 347 (1974). Even if this approach is correct, this constitutional error would still be tested for harmless error employing Weeks-like factors before a conviction is reversed. United States v. Gagan, 43 MJ 200 (1995); see also Delaware v. Van Arsdall, supra at 684, 106 S.Ct. at 1438; United States v. Owens, 789 F.2d 750, 761-62 (9th Cir.1986), rev'd on other grounds, 484 U.S. 554, 108 S.Ct. 838, 98 L.Ed.2d 951 (1988); United States v. Cerro, 775 F.2d 908, 915-16 (7th Cir.1985).