[¶ 1] Tim Sorenson appealed from an amended judgment awarding primary residential responsibility of his son to Jana Slater, the child’s mother. We reverse and remand for further proceedings, concluding the district court did not exceed the scope of a prior remand but erred in admitting into evidence and relying upon the results of polygraph tests to conclude Sorenson had committed domestic violence.
I
[¶ 2] Sorenson and Slater, who were never married, had a child together in May 2008. The child initially lived with Slater, but Sorenson petitioned to establish paternity and sought custody of the child.
[¶ 3] In November 2008, the child suffered a broken clavicle. The parties disputed whether the injury occurred during a weekend visitation with Sorenson or after the child had been returned to Slater. The emergency room doctor who treated the child filed a report with social services, resulting in investigations into the incident by social services and law enforcement. Sorenson and Slater both submitted to polygraph examinations as part of the criminal investigation. No criminal charges were filed, but social services recommended that Sorenson and Slater attend parenting classes. See Sorenson v. Slater, 2010 ND 146, ¶ 3, 786 N.W.2d 739.
[¶ 4] In September 2009, following a two-day trial, the district court issued its order awarding primary custody to Soren-son, with reasonable visitation to Slater. The court’s order included findings of fact on each of the best interests factors under N.D.C.C. § 14-09-06.2(1). The court determined that the best interests of the child favored custody with Sorenson, finding that Sorenson’s home environment was more stable and organized than Slater’s. The results of the polygraph examinations were not introduced into evidence at the trial. On the issue of domestic violence, the court’s finding, in total, was: “This is *185not an issue.” Sorenson, 2010 ND 146, ¶ 5, 786 N.W.2d 739.
[¶ 5] Slater appealed to this Court, which reversed the judgment and remanded. Id. at ¶ 14. A majority of this Court concluded that the district court’s findings of fact on two of the best interests factors were clearly erroneous, that the findings on two factors did not adequately explain the court’s rationale, that the court had failed to acknowledge testimony it had received on some of the factors, and that the findings were sparse and conclusory. Id. at ¶¶ 9-10. The majority also concluded that the district court had applied the wrong version of the statute setting out the best interests factors. Id. at ¶ 11. The Court therefore reversed the judgment awarding Sorenson custody of the child and remanded to the district court for further findings. Id. at ¶ 14.
[¶ 6] On remand, the district court held a supplemental evidentiary hearing. Slater introduced extensive additional evidence regarding the broken clavicle incident, including testimony by medical experts and the law enforcement officer who had investigated the incident. Over Sor-enson’s objection, the court allowed the investigating officer, who had not conducted the polygraph tests, to testify regarding the results of the polygraph tests. The officer testified Sorenson showed signs of deception on his polygraph exam and Slater did not show signs of deception on her exam. The court did not, however, allow the actual written test results into evidence. The district court issued new findings of fact, expressly relying upon the polygraph results to find that Sorenson had committed domestic violence. The court concluded it was in the best interests of the child for Slater to have primary residential responsibility, with Sorenson to have parenting time on alternating weekends, one evening during the week, and alternating weeks during the summer. An amended judgment was entered, and Sor-enson appealed.
II
[¶ 7] Sorenson contends the district court exceeded the scope of the remand when it held a supplemental eviden-tiary hearing and allowed Slater to present additional evidence on issues other than the two findings of fact which this Court had expressly found to be clearly erroneous. Sorenson argues the court was strictly limited upon remand to merely making new findings of fact on best interests factors (b) and (c) under N.D.C.C. § 14-09-06.2(1), and could not consider additional evidence and make new findings of fact on the other best interests factors, including domestic violence. See N.D.C.C. § 14-09-06.2(l)(j).
[¶ 8] Sorenson has misconstrued the holding in the prior appeal. This Court did not merely determine that two findings of fact were clearly erroneous and remand for the limited purpose of new findings on those isolated factors. A review of the prior opinion in its entirety demonstrates that the holding was much broader. This Court first concluded that the district court’s findings of fact on best interests factors (b) and (c), assessing the parents’ ability to provide food, clothing, and shelter and the parents’ ability to meet the child’s developmental needs, were “not in-accord with the record” and were clearly erroneous. Sorenson, 2010 ND 146, ¶ 8, 786 N.W.2d 739. The Court then noted that the district court’s findings on factors (c) and (d), assessing the child’s development, the stability of the home environment, and the impact of extended family, did not “properly explain the court’s rationale.” Id. at ¶ 9. The Court also noted that the district court’s rationale was “especially unclear, given it failed to acknowl*186edge the testimony it received concerning some of the best interests factors.” Id. The Court then concluded that the district court’s findings in general were “sparse and conclusory and [did] not allow this Court to properly perform its appellate function.” Id. at ¶ 10. Finally, the Court concluded that the district court had applied the wrong version of the best interests factors under N.D.C.C. § 14-09-06.2(1), noting that the best interests factors had been substantively changed and the district court should have applied the prior version of the factors. Sorenson, at ¶ 11. Thus, this Court’s direction that the case be remanded to the district court “for further findings consistent with this opinion,” id. at ¶ 14, was broader than a mere directive to make additional findings on two isolated best interests factors. The district court was directed generally to “make more specific findings, clearly explaining its rationale,” id. at ¶ 10, and to apply the substantively correct version of the best interests factors, id. at ¶ 11.
[¶ 9] This Court has repeatedly held that, when we remand for redetermi-nation of an issue without specifying the procedure to be followed, the district court may decide the issue based on the evidence already before it or may take additional evidence. E.g., Livinggood v. Balsdon, 2006 ND 215, ¶ 5, 722 N.W.2d 716; Frisk v. Frisk, 2006 ND 165, ¶ 17, 719 N.W.2d 332; Kautzman v. Kautzman, 2000 ND 116, ¶ 7, 611 N.W.2d 883. The decision whether to take additional evidence is within the district court’s discretion, and its determination will be reversed on appeal only for an abuse of discretion. Livinggood, at ¶ 5; Frisk, at ¶¶ 17-18, Kautzman, at ¶ 7. This Court did not specify the procedure to be followed upon remand, and it was therefore within the district court’s discretion to decide whether to allow introduction of additional evidence.
[¶ 10] Sorenson also contends he was unfairly surprised when Slater sought to introduce additional testimony on the broken clavicle incident at the supplemental evidentiary hearing. This Court has consistently held, however, that a continuance is the proper remedy when a party claims that introduction of evidence constitutes unfair surprise. E.g., State v. Procive, 2009 ND 151, ¶ 23, 771 N.W.2d 259; Paulson v. Paulson, 2005 ND 72, ¶ 16, 694 N.W.2d 681. Sorenson did not seek a continuance when Slater attempted to introduce evidence regarding the broken clavicle incident.
[¶ 11] We conclude that the district court did not abuse its discretion when it held a supplemental evidentiary hearing and did not exceed the scope of the remand when it allowed introduction of additional evidence relating to the broken clavicle incident.
Ill
[¶ 12] Sorenson challenges the district court’s finding that he had committed domestic violence in the broken clavicle incident, and particularly alleges the district court erred in admitting into evidence, and relying upon, the results of polygraph examinations.
[¶ 13] Lieutenant Kvande of the Williams County Sheriffs Office conducted the criminal investigation of the broken clavicle incident. During the course of the investigation he asked Sorenson and Slater to submit to polygraph examinations, and they consented. The polygraph examinations were conducted by Special Agent Pfenning, not by Kvande. The results of the polygraph examinations were communicated to Kvande, and copies of the written polygraph reports were included in Kvande’s investigative report.
*187[¶ 14] At the supplemental evidentiary hearing on remand, Slater sought to introduce the results of the polygraph examinations through testimony of Kvande. Slater’s counsel asked Kvande, “what did the polygraph results indicate?” Sorenson objected, and although no foundation had been laid for admission of the polygraph results, the objection was overruled and Kvande testified: “The results indicated that Mr. Sorenson showed signs of deception and that Jana Slater did not show any signs of deception.” Slater’s counsel then offered into evidence Kvande’s investigative report, which included the written report of the polygraph results. Sorenson again objected, and the district court sustained the objection and the written report was not admitted.
[¶ 15] Although this Court has indicated that the results of a polygraph examination are generally not admissible in a criminal trial unless the parties have stipulated to their admissibility, the Court has not fashioned a definitive rule for admissibility of polygraph results in other contexts. See, e.g., State v. Weatherspoon, 1998 ND 148, ¶ 6, 583 N.W.2d 391; Healy v. Healy, 397 N.W.2d 71, 74 n. 1 (N.D.1986). This Court has noted, however, that polygraph evidence has generally been considered to be unreliable, quoting the conclusion reached by the United States Supreme Court:
The contentions of respondent and the dissent notwithstanding, there is simply no consensus that polygraph evidence is reliable. To this day, the scientific community remains extremely polarized about the reliability of polygraph techniques .... The approach taken by ... Rule 707 — excluding polygraph evidence in all military trials — is a rational and proportional means of advancing the legitimate interest in barring unreliable evidence.
Weatherspoon, at ¶ 7 (quoting United States v. Scheffer, 523 U.S. 303, 309, 312, 118 S.Ct. 1261, 140 L.Ed.2d 413 (1998)). In Scheffer, at 312, 118 S.Ct. 1261, the United States Supreme Court concluded:
Although the degree of reliability of polygraph evidence may depend upon a variety of identifiable factors, there is simply no way to know in a particular case whether a polygraph examiner’s conclusion is accurate, because certain doubts and uncertainties plague even the best polygraph exams.
Based upon the lack of scientific acceptance of the reliability and accuracy of polygraph testing, courts have generally held that the results of a polygraph examination are not ordinarily admissible to establish the truth of statements made during the examination unless the parties have stipulated to admissibility prior to administration of the test. 31A C.J.S. Evidence § 317 (2008); see also United States v. Montgomery, 635 F.3d 1074, 1094 (8th Cir.2011) (“[although there is no per se ban on the use of polygraph evidence in this circuit, ‘[o]ur cases make clear that polygraph evidence is disfavored’ ”) (quoting United States v. Gill, 513 F.3d 836, 846 (8th Cir.2008)).
[¶ 16] This Court has also clarified that a party seeking to introduce polygraph evidence must, as a preliminary matter, present scientific evidence establishing the reliability of polygraph examination results. See Weatherspoon, 1998 ND 148, ¶ 6, 583 N.W.2d 391; see also State v. Newnam, 409 N.W.2d 79, 87 (N.D.1987); State v. Swanson, 225 N.W.2d 283, 285 (N.D.1974). In addition, certain types of evidence, including evidence of the results of scientific tests and testimony of an expert witness, require the proponent to establish a sufficient foundational basis for the evidence, and the determination whether an adequate foundation has been *188laid lies within the sound discretion of the district court. See, e.g., Piatz v. Austin Mut. Ins. Co., 2002 ND 115, ¶ 17, 646 N.W.2d 681; Cook v. Eggers, 1999 ND 97, ¶ 13, 593 N.W.2d 781.
[¶ 17] In this case, the district court allowed Kvande to testify about the polygraph results over Sorenson’s objection, and expressly relied upon those results to find that Sorenson was being untruthful and had committed domestic violence, when the record demonstrates there were blatant and obvious foundational and hearsay problems precluding admission of the evidence. The record is utterly devoid of any foundational evidence demonstrating the scientific reliability or accuracy of polygraph testing in general. There is absolutely no evidence about the training, experience, or competency of the officer who actually conducted the test. Nor is there any evidence that the test was conducted in accordance with established methods or protocols, or that precautions had been taken to ensure the polygraph machine was functioning properly. In short, no attempt was made to demonstrate the customary foundational prerequisites for introduction of testimony about scientific test results. Kvande was quite literally allowed to testify that another officer, whose credentials were unknown, had conducted some unspecified form of polygraph test on Sorenson and Slater, and that Kvande had been advised the tests indicated Sorenson had shown signs of deception and Slater had not shown signs of deception.
[¶ 18] If this unquestionable lack of foundation were not enough by itself, the record further shows obvious hearsay problems with Kvande’s testimony. Kvande testified that he did not administer the polygraph examinations and did not know what steps were used in conducting the tests, but only that another officer had conducted the tests and the results had been communicated to him. The district court nevertheless allowed Kvande to testify that the test results indicated Sorenson had been deceptive but Slater had not. The court then relied upon the indication of deception — the “truth of the matter asserted” by the polygraph results — to make findings regarding Sorenson’s lack of credibility. See N.D.R.Ev. 801(c). The testimony was obvious hearsay which was inadmissible under N.D.R.Ev. 802. See Healy, 397 N.W.2d at 74 (testimony about polygraph test results by someone other than the person who administered the test “was clearly inadmissible because it was hearsay”).
[¶ 19] We are aware that this was a bench trial, and in such cases the trial court should admit all evidence which is not clearly inadmissible but must then consider only competent evidence when making findings of fact:
We have said in a non-jury case the court should “admit all evidence which is not clearly inadmissible because a judge, when deliberating the ultimate decision, is capable of distinguishing between admissible and inadmissible evidence.” McKechnie v. Berg, 2003 ND 136, ¶ 7, 667 N.W.2d 628. In a bench trial, we presume the court only considered competent evidence, and it is not reversible error to admit incompetent evidence unless the evidence induced an improper finding.
In re B.B., 2007 ND 115, ¶ 10, 735 N.W.2d 855; accord Healy, 397 N.W.2d at 74. This Court in Healy therefore concluded that the erroneous admission of polygraph results in a non-jury case did not require reversal where the trial court’s findings did not mention the polygraph results and nothing in the record indicated that the incompetent evidence had influenced the court’s decision. Healy, at 74-75.
[¶ 20] In contrast, the district court in this case mentioned the polygraph results *189several times in its findings of fact, and expressly stated its reliance on the polygraph results to find that Sorenson was being untruthful about the broken clavicle incident:
Detective Kvande interviewed both parties as part of his investigation and requested polygraph examinations. Jana passed her exam but Tim failed his exam with deception noted. The parties were asked the same questions. Detective Kvande did not find Tim credible.
Based upon consideration of all the evidence the Court does find that the child suffered serious bodily injury while in the care of his father, Tim. The Court is unsure whether the injuries were intentional or unintentional, but notes that Tim did not provide any plausible explanation regarding unintentional injury to the child. The results of his polygraph while not admissible in a criminal case are distributing [sic] and can be considered in a custody proceeding. If Tim did not intent [sic] to harm the child there would be no deception indicated.
The Court believes Tim is not being truthful. Because of this, the polygraph results, the medical evidence which indicates the injuries occurred during Tim’s visitation, and the child received serious bodily injury, the Court makes a finding of domestic violence against Tim.
[¶21] In this case, the district court did not ignore the incompetent and inadmissible evidence, but expressly relied upon it as the basis for the crucial findings that Sorenson had been untruthful about the broken clavicle incident and had committed domestic violence upon the child. Under the circumstances, the admission of the testimony of the polygraph results and the court’s express reliance upon that evidence as the basis for critical findings of fact was reversible error.
[¶ 22] We conclude that the district court erred in admitting into evidence testimony of the polygraph results and in relying upon the inadmissible evidence when making its findings of fact. Accordingly, we reverse the amended judgment and remand for further proceedings. We note that the district judge who presided over the original trial and the proceedings on the prior remand has retired, and the case must be assigned to a different district judge. The new judge must review the existing record and has discretion in determining whether to allow an evidentia-ry hearing to hear additional testimony.
IV
[¶ 23] The amended judgment is reversed and the case is remanded for further proceedings in accordance with this opinion.
[¶ 24] GERALD W. VANDE WALLE, C.J., CAROL RONNING KAPSNER, and DALE V. SANDSTROM, JJ„ concur. DANIEL J. CROTHERS, J., concurs in the result.