OPINION OF THE COURT
Sotero Sanes appeals from a final judgment and commitment order of the district court. On December 29, 1993, an information was filed charging Sanes with burglary in the first degree, in violation of 14 V.I.C. § 442(1); rape in the first degree, in violation of 14 V.I.C. § 1701(3); robbery in the first degree, in violation of 14 V.I.C. § 1862(2); and possession of a dangerous weapon during the commission of a violent crime, in violation of 14 V.I.C. § 2251(a)(2)(B).
Sanes was accused of attacking Carmen Velez on January 13, 1993 and again on July 14,1993. Counts I, II, III and IV stem from the first attack and Counts V, VI and VII from the second. Sanes was convicted on all counts.
Sanes argues that the district court erred by denying his motion to suppress. He contends that the voice identification procedure was impermissibly suggestive. Further, he argues that the district court abused its discretion by limiting his expert witness' testimony about the validity of the voice identification procedures. Finally, Sanes contends that the district court erred by denying his pretrial motion to sever Counts I, II, III and IV from Counts V, VI and VII, arguing that the charges arose from separate incidents and their joinder was highly prejudicial to him. We will affirm.
I.
Sanes argues that the trial court violated his right to due process by denying his motion to suppress the voice identification. He contends that his voice sample was tainted by outside factors, *464technical and otherwise, which could have influenced the listener, making it more likely that his voice would be selected. He cites background voices, longer delivery time, different transition time between the statements made in Spanish and those made in English, his monotone and a number of interruptions, as examples of the factors which he believes, when taken cumulatively, compromised the reliability of the identification.
We find no evidence that the procedures used were impermissibly suggestive or that there was a substantial likelihood of misidentification. Ms. Velez heard the voice of her attacker on both January 13, 1993 and July 14, 1993. On each occasion, she conversed with the intruder for ten minutes. She listened to the voice array three times. She was ready to identify Sample No. 4 after the second time, but was encouraged by the police to listen to the tape one more time. Velez then positively identified the fourth voice as that of her attacker. Sample No. 4 was Sanes' voice.
The district court listened to the voice array and found that all of the voices contained unique factors. After listening to Sanes' expert witness on voice identification procedures, the district further found that, although Sample No. 4 was different horn the other samples, the differences did not present a sufficiently significant contrast as to cause misidentification. Before denying Sanes' motion to suppress, the district court applied the eyewitness identification test, enunciated in Neil v. Biggers, 409 U.S. 188, 93 S. Ct. 375 (1972) and Manson v. Brathwaite, 432 U.S. 98, 97 S. Ct. 2243 (1977). Sanes argues that this test is inapposite. We disagree.
In Neil, the Supreme Court developed a five-part test to test the reliability of an eyewitness identification. The Neil factors include the opportunity of the witness to view the criminal at the time of the crime; the witness' degree of attention; the accuracy of the witness' prior description of the criminal; the level of certainty demonstrated by the witness at the time of the confrontation; and the length of time between the crime and the confrontation. 93 S. Ct. at 382. Expanding on Neil, Manson held that admitting testimony even following a suggestive identification procedure does not violate due process if the identification has sufficient indicia of reliability. 97 S. Ct. at 2252-53. Although voice identification *465obviously differs from eyewitness identification (for example, what is at issue in the first part of the Neil and Manson test is opportunity to view rather than opportunity to hear), we conclude that the Neil and Manson eyewitness identification test, adapted to voice identification, provides a standardized source of guidance to district courts for assessing the reliability of voice identification as well.
Applying Neil, the district court found that Ms. Velez was alert and had a sufficient opportunity to view or hear her attacker on both occasions and that she presented an accurate description of the defendant and his clothing (which she later identified.) Moreover, she identified the items taken from appellant as looking like items the perpetrator had worn, and there is no evidence indicating that these items, when presented to Ms. Velez, were suggestively linked to either the appellant or to Sample No. 4. Further, Ms. Velez stated that she was certain that Sample No. 4 was the voice of her attacker. Finally, although there was a six-month period between the incidents, only fifteen days elapsed between the last attack and her positive voice identification. We also note that Ms. Velez not only listened to her attacker for a considerable period of time during the two incidents but testified that she engagaed him in conversation in the hope that she could identify his voice, and hence the opportunity and degree of attention criteria of Neil/Manson are met. We also listened to the tape during oral argument, and conclude that the evidence supports the district court's findings and conclusions. We conclude that the district court properly applied the Neil/Manson factors and properly denied Sanes' motion to suppress the voice identification.
II.
Sanes next argues that the district court erred by prohibiting expert testimony comparing voice identification and eyewitness identification. The district court permitted the testimony of Sanes' expert witness, Dr. Harry Hollien, a professor of linguistics, on the question whether the voice exemplar containing the defendant's voice was suggestive. Dr. Hollien was also permitted to comment on the distinguishing characteristics of Sanes' voice sample and *466their impact, if any, on Ms. Velez's positive identification. However, the district court would not permit Dr. Hollien to explain to the jury why voice identification was not as accurate as eyewitness identification under ideal circumstances.
Under Federal Rule of Evidence 702, "[i]f scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training or education, may testify thereto in the form of opinion or otherwise." Whether to allow scientific or technical expert testimony, however, is within the discretion of the district court and is reviewed only for abuse. United States v. Downing, 753 F.2d 1224, 1232 (3d Cir. 1985).
Under Daubert v. Merrell Dow Pharmaceuticals, Inc.,_U.S--113 S. Ct. 2786 (1993), when faced with the proffer of expert scientific testimony under Rule 702, the district court must preliminarily assess whether the testimony underlying the reasoning or methodology is scientifically valid and can properly be applied to the facts at issue. The district court properly held a Daubert hearing in which the nature of Dr. Hollien's testimony was fully considered. We conclude that the district court appropriately limited Dr. Hollien's expert testimony to the pertinent issue of whether the distinguishing factors on Sample No. 4 unduly influenced Ms. Velez's selection. We also conclude that the district court did not abuse its discretion in excluding from the jury's consideration the testimony that would have compared eyewitness and voice identification; such testimony was of limited probative value, and that value would have been outweighed by waste of time and confusion. See Downing, 753 F.2d at 1232.
III.
Finally, Sanes argues that the district court abused its discretion by denying his pretrial motion to sever Counts I through IV from Counts V, VI and VII. He contends that joinder of the counts involving two separate incidents substantially impaired his right to a fair trial. We disagree. Ms. Velez's testimony and a common modus operandi provided sufficient reason to deny Sanes' motion to sever the counts and to conduct a single trial.
*467Federal Rule of Criminal Procedure 14 provides:
[I]f it appears that a defendant or the government is prejudiced by a joinder of offenses or of defendants in an indictment or information or by such joinder for trial together, the court may order an election of separate trials of counts, grant a severance of defendants or provide whatever other relief justice requires. In ruling on a motion by a defendant for severance the court may order the attorney for the government to deliver to the court for inspection in camera any statements or confessions made by the defendant which the government intends to introduce in evidence at the trial.
Sanes must demonstrate that joinder would result in a "manifestly unfair trial," beyond a mere showing that he would have had a better chance of acquittal with separate trials. United States v. Eufrasio, 935 F.2d 553, 568 (3d Cir.), cert. denied, 502 U.S. 925, 112 S. Ct. 340 (1991). Here, Ms. Velez identified Sanes as the person who attacked her both times. Moreover, during the second attack Sanes inculpated himself in the first offense by referring to Ms. Velez' failure to keep quiet following the first attack. The modus operandi of the crimes is similar and it was evident that Ms. Velez would have to testify about both incidents, which would have required the government to offer essentially the same proof in both trials. We conclude that the district court did not abuse its discretion by denying Sanes' motion to sever the counts that stem from the first incident from those of the second.
IV.
In sum, we have examined all contentions raised by the appellant and conclude that the district court's decisions were proper in all respects. We will affirm.