*449 Opinion of the Court
WISS, Judge:A special1 court-martial composed of officer and enlisted members convicted appellant, over his pleas, of wrongful use of cocaine and sentenced him to a bad-conduct discharge and reduction to the lowest enlisted grade. See Art. 112a, Uniform Code of Military Justice, 10 USC § 912a. The convening authority approved these results, and the Court of Military Review affirmed. 34 MJ 562 (1991).
On appellant’s petition, we granted review of the following issue:
WHETHER THE ARMY COURT ERRED IN HOLDING THAT THE MILITARY JUDGE DID NOT ABUSE HIS DISCRETION IN PERMITTING A GOVERNMENT POLYGRAPH EXAMINER TO TESTIFY IN REBUTTAL, OVER DEFENSE OBJECTION, THAT APPELLANT SUBMITTED TO A POLYGRAPH EXAMINATION AND THAT THE EXAMINATION INDICATED THAT APPELLANT WAS UNTRUTHFUL WHEN HE DENIED USING COCAINE.
We hold that the court below did err in finding that the military judge did not abuse his discretion in admitting the rebuttal testimony. The Government, as the party offering the evidence, did not carry its burden to establish the requisite foundation of its reliability. See United States v. Gipson, 24 MJ 246 (CMA 1987); see also Daubert v. Merrell Dow Pharmaceuticals, Inc., — U.S. -, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). Because we are unable to conclude that there is no fair risk that appellant was prejudiced by this error, the findings and sentence must be set aside.
I
Appellant is a senior noncommissioned officer with nearly 20 years of active military service. On July 25, 1989, while serving in Panama, he submitted a urine sample to military authorities during an unannounced urine inspection of members of his unit, see Mil.R.Evid. 313(b), Manual for Courts-Martial, United States, 1984. Subsequent analysis of the sample revealed presence of cocaine metabolites, so appellant was charged with wrongful use of cocaine.
Prior to any charges being preferred and after consulting with his defense counsel, appellant requested that criminal investigators administer a polygraph examination of him in an effort to clear his name. Defense counsel had advised appellant that, regardless of the outcome, the results of the examination likely could not be admitted in a court-martial but could be considered against him at proceedings under Article 15, UCMJ, 10 USC § 815.
During the “preinstrument phase of the examination,” appellant denied knowing use of cocaine and asserted the possibility of innocent ingestion. He explained that, a few days prior to the urinalysis, he had visited an off-post drinking establishment and had consumed both alcoholic beverages and cigarettes. He suggested “that it was at that particular establishment where he may have unknowingly consumed either through the liquor or the cigarettes the cocaine which caused the positive urinalysis result.”
Immediately after the examination, the polygrapher evaluated the results and advised appellant that he believed appellant had been “practicing deception while answering the relevant questions” during the examination. Under agreement between defense counsel and the polygrapher, however, there was no “post-instrument phase” of the examination — that is, no interview of appellant after he was removed from the polygraph. Subsequently on that same day, the instant charge was preferred against appellant.
At trial, the Government moved in limine to admit the results of the examination under Mil.R.Evid. 702. The Government indicated that it would use this evidence only if appellant testified and denied use of cocaine during the relevant time *450period. Special Agent Widup — a qualified Army criminal investigator and the polygrapher who had administered the examination to appellant — testified to lay a foundation to support the motion.
Widup stated during his testimony that he utilizes “the theory of psychological set” in his testing. This approach “holds that when a person is presented with a set of questions, that they will focus their concern on a question or a set of questions which pose the most harm to their well-being.” Specifically, when testing appellant, he had used the “modified general question technique,” which includes three types of questions: relevant questions, which deal with the particular incident under investigation; control questions, which deal with subject matter “similar to” the relevant issue but “separated” from it “by category and/or time”; and irrelevant questions, which have nothing to do with the relevant issue.
In theory, if a person is involved in the incident under investigation, he will be more threatened by the relevant questions; if he is not involved in the incident, he will be more troubled by the control questions. The irrelevant questions simply “reestablish a base line” on the chart during the examination. In other words, “an old lie will trouble an innocent man more than a crime he is charged with committing” if he is innocent of that crime.
Widup testified that “the relevant questions” asked appellant were as follows:
Did you knowingly possess any cocaine within 30 days of that urinalysis?
Did you knowingly use any cocaine within 30 days of that urinalysis?
Are you lying about receiving advance knowledge of that urinalysis?
Did you knowingly ingest any cocaine in any manner within 30 days of that urinalysis?
He concluded that “deception” was indicated in appellant’s negative answers to these questions.
Widup acknowledged, however, that “you cannot deduce from that particular conclusion that he may be lying about one issue and not lying about another.” Specifically, Widup was not able to “differentiate whether he’s lying about advance knowledge [of the inspection] or whether he’s lying about use.” Further, Widup recognized that “physiological reaction^]” like those measured in a polygraph can be caused by any number of things other than lying. Factors such as “extraneous noises,” “[v]isual stimuli,” “emotional complexes associated with a particular word or a question,” or even “random thoughts” could cause physical reactions that are measured as positive responses for deception on a polygraph.
Though the defense did not challenge the qualifications of Agent Widup, it did object to the motion to admit the polygraph result on several grounds. Principally, the essence of the objections was as follows: First, the evidence was not relevant because the examiner’s conclusion of deception could not be narrowed to questions relating to wrongful use, as opposed to simply being a broad conclusion of deception, see Mil.R.Evid. 401 and 402; second, for similar reasons, the evidence was not helpful to the trier of fact, see Mil.R.Evid. 702; third, again for similar reasons, the evidence would confuse and mislead the members and would pose risks of unfair prejudice that far outweighed its probative value, see Mil.R.Evid. 403; and fourth, reliability of the polygraph had not been established by scientific proof, so the evidence was not valid in any event.
Nonetheless, the military judge ruled that, if appellant testified and denied use of. cocaine, he would permit the Government to use Widup’s testimony about his polygraph conclusions in order “to assist the trier of fact to determine the credibility of the accused under those circumstances.” The military judge reasoned:
I am convinced of the soundness and reliability of the process and technique used in forming the polygrapher’s opinion in this case. I do not think that the evidence will overwhelm or confuse or mislead the jury, and I believe that if presented it will help them to determine *451the credibility of the — of the accused under the circumstances.
After the Government had concluded its case on the merits, appellant renewed his objection to this evidence, without success. Thereafter, appellant testified in his own defense; he swore that he had never knowingly used cocaine and had no idea how he had come up “hot” on the urinalysis. He also indicated that he had known of the impending test — with the arguable implication that under that circumstance, appellant would not have risked his career by knowingly ingesting cocaine. He acknowledged having taken a polygraph examination and acknowledged, too, that the examiner had informed him that the test had indicated deception.
At that point, the defense again sought to exclude Widup’s testimony. Citing Mil. R.Evid. 403, the defense argued that the anticipated evidence from the prosecution would be “overkill,” in light of appellant’s candid admissions regarding the polygraph examination. Not convinced, the military judge renewed his earlier ruling. Ultimately, Widup did testify, in prosecution rebuttal, that appellant had been “practicing deception while answering the relevant questions.” Additionally, he was permitted to testify that, in his opinion, appellant “was not being truthful” during the examination.
Prior to the members’ deliberating on the question of guilt, the military judge instructed, inter alia, as follows:
You have also heard the testimony of Special Agent Widup concerning polygraph examinations. He too has been accepted by me as an expert witness because his knowledge, skill, training, or education which may assist you in understanding the evidence or in determining a fact in issue. Again, you are not required to accept the testimony of any expert witness or give it more weight than the testimony of any other witness. You should, however, consider the qualifications of this expert witness. Further, you must consider the sole purpose of which the polygraph evidence was introduced and that is to assist you in evaluating the credibility of the court— of the accused’s in-court testimony. This evidence was introduced for whatever tendency it may have to rebut the accused’s testimony that he never knowingly ingested cocaine, which would, of course, include a denial that he knowingly ingested it during the period charged. Polygraph examinations have historically been viewed with great skepticism and you should examine this evidence very critically. You have heard the testimony of Mr. Widup as to how the test was conducted and the theory of how a polygraph operates. The weight to be given to this evidence, if any, is solely within your discretion. You should assess the theory of physiological responses upon which the polygraph analysis is based along with all the other evidence that Mr. Widup provided. Now, the polygraph has been commonly and incorrectly referred to as a lie detector. At best, and you don't have to consider this to be true, but at best it is capable of measuring physiological responses at a particular point in time. Further, in evaluating this evidence you should consider the fact that you have had the opportunity to observe the accused and his demeanor in court, while you were not present and did not have the opportunity to observe the polygraph examination.
II
In United States v. Gipson, 24 MJ 246 (1987), this Court concluded that results of polygraph examinations are not per se inadmissible as scientifically unreliable.2 Accord United States v. West, 27 MJ *452223, 225 (CMA 1988). The majority in Gipson observed, “The greater weight of authority indicates that it can be a helpful scientific tool.” Neither, however, is polygraph examination at the top level of reliability as scientific evidence — the level at which “the principles underlying the expertise are so judicially recognized that it is unnecessary to reestablish those principles in each and every case,” such as fingerprint or ballistic evidence. Rather, polygraph principles fall in the middle level, the “range of scientific and technical endeavor that can neither be accepted nor rejected out of hand.” 24 MJ at 249.
In any given case in which polygraph evidence is offered, the benchmarks against which admissibility must be determined are Mil.R.Evid. 401, 402, 403, and 702. “Taken together, the[se] rules seem to describe a comprehensive scheme for processing expert testimony.” 24 MJ at 251. Addressing Mil.R.Evid. 401 and 702, Judge Cox observed:
For any type of evidence to have logical relevance [under Mil.R.Evid 401], however — scientific evidence included— some degree of reliability is implicit. [United States v. Downing,] 753 F.2d [1224] at 1238 [(3d Cir.1985)]. In addition, the helpfulness standard of Fed. R.Evid. 702 [and Mil.R.Evid. 702] is said to “impl[y] a quantum of reliability beyond that required to meet a standard of bare logical relevance.” Id. [753 F.2d] at 1235.
Accord Daubert v. Merrell Dow Pharmaceuticals, Inc., supra.
If a military judge does admit polygraph evidence consistent with these principles, the scope of the permissible use of that evidence is clear:
First and foremost, while polygraph evidence relates to the credibility of a certain statement, it does not relate to the declarant’s character. At best, the expert can opine whether the examinee was being truthful or deceptive in making a particular assertion at the time of the polygraph exam. It is then for the fact-finder to decide whether to draw an inference regarding the truthfulness of the examinee’s trial testimony.
24 MJ at 252-53 (footnote omitted).
While the military judge’s instructions, as a whole, do appear to limit the proper use of this evidence consistent with the opinion in Gipson, we conclude that the Government here did not succeed in laying the necessary foundation of reliability that is inherent in satisfying Mil.R.Evid. 401, 402, and 702. 24 MJ at 251.
First, the polygraph examination here did not permit the examiner’s conclusion of deception to focus and differentiate between questions relating to criminal conduct (that is, possession and knowing use of cocaine) and innocent conduct (that is, learning in advance of the impending urinalysis). It is true that appellant testified on both areas at trial', so one might argue that a showing of deception in either event would be relevant to factfinders in judging the credibility of appellant’s in-court testimony. The impact on the factfinders might differ substantially, however.
For instance, if appellant were deceptive about claiming advance knowledge of the test, factfinders might not place much weight on that deception; instead, they might even view it as a human-instinct response of an innocent man to grasp at any straw — even a fabrication — to convince his peers of his innocence. The impact from a conclusion of deception relating to criminal conduct at issue in the trial, however, might be dramatically more serious. If the polygraph examiner could have con-*453eluded that appellant was deceptive in responding to the questions relating to his knowing ingestion of cocaine, the factfinders might be predicted to view his similar in-court disclaimers with a good deal of skepticism. The failure of the Government here to differentiate between the possible areas of deception undermines the reliability of the evidence offered in this case.
Second, no “post-instrument” interview of appellant was conducted here as part of the examination. Widup, however, testified that such an interview “is normally required.” Regardless that the interview was omitted under agreement between defense counsel and the polygrapher, the fact remains that such a significant departure from the procedure that is “normally required” makes it incumbent on the preferring party to demonstrate that that omission does not undermine the examination’s reliability. The Government did not do so here. In fact, the indications at trial were to the contrary: that a post-instrument interview would have been helpful to differentiate the basis for appellant’s deception on the relevant questions — the point just discussed above.
Finally, it was revealed at trial that the polygraph report incorrectly reflects appellant’s response to a critical control question. When asked about this, Widup explained that the error was typographical. The fact remains, however, that an answer to an important control question was erroneously recorded — a fact that does nothing to prop up the reliability of this polygraph examination.
Ill
We are unable to conclude, as the Government would have us do, that there is no fair risk that appellant was not prejudiced by this error. Appellant forthrightly put his lengthy good service and his explicit denial of ever having used cocaine up against the prosecution’s case, which consisted of nothing more than the urinalysis results plus expert interpretation. That would seem, under normal circumstances, to be an interesting contest for the factfinders to resolve.
Admission of scientific evidence, accompanied by an opinion of an expert witness interpreting that evidence, indicating that appellant had been deceptive during a polygraph examination that related to the charges to be decided by the factfinders, however, removed any interesting aspect from that contest. The credibility of appellant’s denials — surely a critical part of his case — was devastated. Given the failure of the Government to establish the reliability of this weapon of devastation, we cannot fairly conclude that appellant was unharmed by it.
IV
The decision of the United States Army Court of Military Review is reversed. The findings and sentence are set aside. The record of trial is returned to the Judge Advocate General of the Army. A rehearing may be ordered.
Chief Judge SULLIVAN and Judges COX and GIERKE concur.. The decision below erroneously says this was a general court-martial. 34 MJ 562.
. Subsequent to appellant’s trial, the President approved Mil.R.Evid. 707, Manual for Courts-Martial, United States, 1984, effective July 6, 1991. See Art. 36(a), Uniform Code of Military Justice, 10 USC § 836(a). Subsection (a) of that rule provides:
Notwithstanding any other provision of law, the results of a polygraph examination, the opinion of a polygraph examiner, or any reference to an offer to take, failure to take, or taking of a polygraph examination, shall not be admitted into evidence.
*452In Judge Wiss’ view, to the extent that, consistent with Gipson, see opinion, infra, an accused is able to carry his foundation burden of demonstrating relevance, reliability, helpfulness to the factfinder, and relatively minor risk of confusion of the factfinder, due process and fundamental fairness might seem to compel admission of exculpatory polygraph evidence, notwithstanding this rule. See United States v. Gipson, 24 MJ 246, 252 (CMA 1987). These same concerns, of course, do not weigh as heavily in favor of the prosecution. But cf. United States v. Barger, 931 F.2d 359, 370 (6th Cir.1991). Ironically, then, it seems to him that Mil.R.Evid. 707(a) might be a rule of evidentiary exclusion that applies only to the Government.