LOWE, J., delivered the opinion of the Court.

The appeal of Van Gregory Johnson from his conviction by a jury in the Circuit Court for Anne Arundel County of assault with intent to rape and of committing unnatural and perverted sex practices, raises eight grounds for reversal. Because we reverse the conviction on one of those grounds it is not necessary to respond to the other seven. We are also spared the necessity of reciting the revolting facts of the heinous crime involved because the error compelling reversal is an evidentiary one relating to the voluntariness of Johnson's confession.

For our purposes it need only be noted that an inculpatory statement was given to the police by appellant during a rather lengthy interrogation. A polygraph or lie detector machine was used twice during the questioning. Officer Beck conducted the questioning while using the polygraph. The interrogating officer would look at the polygraph results after each question and advise appellant that he had failed the test and that the officer didn't believe him. The officer also questioned appellant as to what had caused the physiological responses indicating that he had made a false statement. After this procedure had been followed, the interrogator left appellant alone in the room with the polygraph for approximately twenty minutes. Then he returned with a second set of questions and repeated the same process.

Soon afterwards, while the machine was not functioning and appellant and the officer "were completely relaxed" and "talking one-on-one", appellant gave the officer a statement which was inculpatory, though falling short of a confession. When asked at the suppression hearing the reason for doing so, appellant answered:

". . . because I got tired of sitting there and it seemed like everybody was against me. You know, nobody would believe what I said but yet I was supposed to believe what they were saying."

It is apparent that the use of the deception testing device was intended to produce a psychological effect upon the accused in order to obtain the relevant facts if known by appellant. Appellant's expressed reason for confessing does not gainsay the test's effectiveness. It is clear, however, that the use of such a procedure for that purpose would not as a matter of law require the exclusion of a confession so obtained, United States v. McDevitt, 328 F.2d 282 (6th Cir.); Commonwealth v. Hipple, 3 A.2d 353 (Pa.); but see People v. Lettrich, 108 N.E.2d 488 (Ill.), involving a confession after several such tests. Indeed, when these facts became known to the trial judge out of the jury's presence during a suppression hearing, he decided that such circumstances, even in addition to the other factors offered, did not overcome the preponderant proof by the State that the confession was freely and voluntarily given. Mulligan v. State, 18 Md. App. 588, 602. The judge having made that preliminary finding, the question of voluntariness should have then been submitted to the jury on the same evidence, although the jury is to use a higher standard for determining voluntariness — it must find beyond a reasonable doubt that the confession was freely and voluntarily given. Gill v. State, 265 Md. 350, 358.

Appellant contends, and the State did not dispute, that they had "agreed between the parties that the results of the purported polygraph examination should not be disclosed to the jury . . . ." However, the judge did not permit the jury to be told even the fact that a polygraph examination was given. The officer testified to the questions he asked appellant and the answers elicited, but did not reveal the part played by the deception detecting device. At no time was the fact that a polygraph examination was used in the questioning of appellant imparted to the jury. The exclusion of this information was over the vehement objection of appellant. We are, therefore, faced with the narrow question of whether the trial judge properly excluded from the jury's consideration of the voluntariness of appellant's confession the fact that a polygraph was used in the obtention of that confession. We hold that such exclusion was error.

A trial judge may not select which circumstances relative to voluntariness may go to the jury and which may not. Once its preliminary determination of voluntariness is made, see Jackson v. Denno, 378 U.S. 368, all of the relevant evidence bearing on voluntariness must then be submitted to the jury, Day v. State, 196 Md. 384, 399:

"The practice in this State, approved in many cases, is that the court first hears evidence without the jury to determine whether a confession is voluntary and should be admitted. If it decides to admit it, the same evidence is then given to the jury, as it has the final determination, irrespective of the court's preliminary decision, whether or not the confession is voluntary, and whether it should be believed. In so doing, the jury is entitled to have before it all of the evidence which affects the voluntary character of the document, and which the court passed upon in admitting it." (Emphasis added).

The State argues that the trial judge has the right to keep from the jury the fact that the polygraph was utilized in the interrogation because it was "only a motivating factor as opposed to a force factor affecting voluntariness." That, however, is precisely the question the jury must decide. It is within its province to decide that the psychological impact of a polygraph is a mere motivating factor, but it may also determine that it had so coercive an effect as to render a confession thus obtained involuntary. The decision is one for the fact finder alone, especially in light of the Maryland jury's broad role as judge not only of fact, but of law as well. Md. Const., Art. XV, § 5.

Appellee further argues that the exclusion of any evidence concerning the use of a polygraph in an interrogation resulting in a confession is justified because the admission into evidence of the results of a polygraph examination is not proper, citing Rawlings v. State, 7 Md. App. 611. We are not here concerned with the results of a polygraph examination, but rather with the circumstance that it was used as a psychological tool in the interrogation process.

While the issue to our knowledge has never been decided in Maryland, other jurisdictions have held that the taking of a lie detector test may be considered by the jury in determining whether a confession was freely and voluntarily given. Leeks v. State, 245 P.2d 764 (Okla.). A jury should also consider if the method of examination was such as to constitute, in itself, coercion sufficient to find a resulting confession involuntary, Johnson v. State, 166 So.2d 798 (Fla. App.). The very fact that the accused was preparing to take a lie detector test was found to be a proper factor for consideration by the jury when deciding the voluntariness of a confession given before such test was administered. People v. McHenry, 22 Cal.Rptr. 621. See also annotation Evidence — Deception Tests, 23 A.L.R.2d 1306.

The reason for excluding the results of a polygraph examination is the questionable reliability of such evidence. Similarly, the admission into evidence of whether an accused agreed or refused to take such a test may give rise to jury speculation as to his reasons for submitting or refusing to submit to the test. In both cases, a determination of guilt or innocence may be affected by an accused's state of mind after the crime, rather than upon evidence produced related to the crime itself. But evidence of the use of a polygraph as a device to obtain a statement is substantially less prejudicial than either the impact of the questionable results of the device or the effect of the defendant's refusal to take the test. The importance of permitting the jury to weigh the coercive effect of every motivating circumstance surrounding the eliciting of a confession, far outweighs the importance of avoiding the possible prejudice from a reference to its use. Furthermore, as indicated by Lusby v. State, 217 Md. 191, any prejudice may be diminished by instructing the jury as to the limited purpose of the admission of the evidence relating to the device.

The possibility of speculation by jurors as to the results of the polygraph test, although diminished when merely used for purposes of voluntariness, causes the issue to be one charged with evidential cordite. This case presents a not atypical example of such a possibility. The officer stated that, after the accused responded to each question and he had viewed the results on the polygraph machine, he espoused his disbelief of appellant's answers. This could give rise to speculation by the jury that the officer was reacting to the actual test result. On the other hand, it is entirely possible that the officer might have been deceiving the accused as to the real result in order to obtain his reaction and possible confession. Either inference is speculative and gives rise to an obvious danger. Therefore, the jury should be admonished not to speculate upon the machine's results as they relate to the guilt or innocence of the accused, but to draw any logical inferences from the circumstances solely as to the question of whether the procedure used induced an involuntary confession.

We do not hold that the use of psychological devices to interrogate will, as a matter of law, make the resultant confession inadmissible. We do hold that if such devices are utilized, the State takes the chance that a given jury will consider the device or procedure so coercive as to declare the confession to have been involuntary. It would be unsound to exclude from the jury's consideration on the issue of voluntariness the psychological effect of the use of polygraph devices when we fail to exclude the effect of any other potentially coercive condition, person or device present during interrogation. We have frequently quoted Malloy v. Hogan, 378 U.S. 1, 7, to the effect that:

"The basic standard governing the admissibility of an extrajudicial statement is whether, considering the totality of the circumstances, the statement was voluntary. . . . To be voluntary, a statement cannot be `extracted by any sort of threats or violence, nor obtained by any direct or implied promises, however slight, nor by the exertion of any improper influence.'" (Emphasis added). Keller v. State, 2 Md. App. 623, 626-627.

What is an "improper influence" is a question of both law and fact, the evidence of which is twice presented; first to the judge alone, and later with the jury in attendance. The jury must have the opportunity to consider all of the evidence pertaining to the voluntariness of a confession before deciding the question of guilt or innocence. Day, supra. That is an inviolable jury function which may not be eliminated. Gill v. State, 265 Md. at 358-359.

Judgments reversed.

Case remanded for new trial.