BRICKLEY, J.

This is a child custody case in which the circuit court awarded the parties joint legal custody of their three minor children, with physical custody to the plaintiff father and liberal visitation to the defendant mother. The Court of Appeals reversed and awarded physical custody to the defendant. Because it did not correctly review this matter, we reverse the judgment of the Court of Appeals insofar as it relates to the standards of review applicable to child custody cases, and remand to circuit court for further proceedings.

I. FACTS AND PROCEEDINGS

The parties were married in 1975 and have three children. The plaintiff filed for divorce in 1990. Following extended hearings incident to the divorce proceeding, a referee recommended that physical custody of the children be awarded to the defendant.

The referee reached this conclusion after reviewing the statutory factors for determining the best interests of the children. See MCL 722.23; MSA 25.312(3). Although the referee found that most of the factors were of neutral weight, he did find that factors b and c weighed slightly in favor of the defendant and factor f weighed slightly in favor of the plaintiff. The referee also indicated that he considered what the children had told him during private and separate interviews.

At the request of the plaintiff, the circuit court held a de novo hearing. After hearing the testimony of the parties, reviewing the evidentiary record made before the referee, and interviewing the two older children, the circuit court granted physical custody to the plaintiff. Although the circuit court judge found that most of the statutory factors were of neutral weight, he found that factors b, e, and f weighed slightly in favor of the plaintiff. The circuit court issued a December 6, 1991, custody order granting physical custody to the plaintiff and awarding the defendant liberal visitation rights, as well as joint legal custody.

The defendant appealed, and the Court of Appeals reversed the custody determination, concluding that the circuit court had made clearly erroneous findings of fact with regard to factor f and had committed clear legal errors concerning factors b, e, and f. 200 Mich. App. 505; 504 N.W.2d 684 (1993). Without discussing the remaining factors, the Court of Appeals further stated that upon de novo review it found the parties to be equal on all the statutory factors except factor i, which relates to the child's reasonable preference. On the basis of that factor, the Court of Appeals awarded custody to the defendant.

The plaintiff appealed to this Court.

II. STANDARD OF REVIEW

We are asked to clarify the proper standard of review in child custody cases. We conclude that such cases are to be reviewed in accordance with § 8 of the Child Custody Act, which provides:

To expedite the resolution of a child custody dispute by prompt and final adjudication, all orders and judgments of the circuit court shall be affirmed on appeal unless the trial judge made findings of fact against the great weight of evidence or committed a palpable abuse of discretion or a clear legal error on a major issue. [MCL 722.28; MSA 25.312(8).]

By its terms, § 8 distinguishes among three types of findings and assigns standards of review to each. Findings of fact are to be reviewed under the "great weight" standard, discretionary rulings are to be reviewed for "abuse of discretion," and questions of law for "clear legal error."

As noted by the Court of Appeals, the review standards of the Child Custody Act differ from the standards that this Court has applied in alimony disputes and property disputes incident to divorce. However, unlike the Court of Appeals, we do not attempt to import or superimpose the standards applicable to property and alimony disputes into the legislatively prescribed standards applicable to custody disputes. We believe that the difference in standards comports with the fact that the interests at stake in a child custody dispute differ from those involved in a dispute over money or property.

The three standards of review enumerated in § 8 of the custody act are part of the Legislature's comprehensive effort to promote the best interests and welfare of children. By incorporating standards of review into the act, the Legislature apparently recognized that in custody cases the proceedings themselves may jeopardize a child's welfare. Because we believe that standards set forth in § 8 reasonably minimize the possibility of unwarranted and disruptive changes in custody, while at the same time enabling courts to pursue suitable custody arrangements, we adopt those standards as explained more fully herein.

A

The Child Custody Act provides that findings of fact in child custody cases are reviewed under the "great weight of evidence" standard. MCL 722.28; MSA 25.312(8). When the Legislature enacted the custody act it presumably used the phrase, "against the great weight of evidence," with knowledge of its existing meaning and with intent that the phrase maintain its existing meaning. In re Chamberlain Estate, 298 Mich. 278, 284; 299 N.W. 82 (1941). "Against the great weight of evidence" was defined by Murchie v Standard Oil Co, 355 Mich. 550, 558; 94 N.W.2d 799 (1959). The Court explained that a reviewing court should not substitute its judgment on questions of fact unless they "`clearly preponderate in the opposite direction.'" The court should review "the record in order to determine whether the verdict is so contrary to the great weight of the evidence as to disclose an unwarranted finding, or whether the verdict is so plainly a miscarriage of justice as to call for a new trial. . . ." Id.

Although the great weight standard traditionally is applied in the context of granting and denying new trials, the Legislature's adoption of that standard for appellate review of child custody orders reasonably furthers its express intent to: "expedite the resolution of a child custody dispute by prompt and final adjudication. . . ." MCL 722.28; MSA 25.312(8). In custody trials, the trial judge is the finder of fact. Thus, when an appellate court reviews a trial judge's findings, it acts as the functional equivalent of a trial judge reviewing the findings of a jury. Therefore, the great weight standard, which is the standard by which jury findings are reviewed, is also a logical standard by which to review findings of a trial judge. The great weight standard of review allows a meaningful yet deferential review by the Court of Appeals. A more deferential standard, such as "insufficient evidence" or "supported by competent, material, and substantial evidence," could effectively immunize the trial judge's fact finding in contravention of a child's best interests.

In the context of child custody cases, there are findings of ultimate facts, i.e., a finding regarding each factor, and there are findings of ordinary or evidentiary facts. The great weight of the evidence standard applies to all findings of fact. Thus, a trial court's findings on each factor should be affirmed unless the evidence "`clearly preponderate[s] in the opposite direction.'" Murchie, supra at 558.

B

Under the Child Custody Act, discretionary rulings are reviewed under a "palpable abuse of discretion" standard. MCL 722.28; MSA 25.312(8). Abuse of discretion was defined in Spalding v Spalding, 355 Mich. 382, 384-385; 94 N.W.2d 810 (1959), as follows:

Where, as here, the exercise of discretion turns upon a factual determination made by the trier of the facts, an abuse of discretion involves far more than a difference in judicial opinion between the trial and appellate courts. . . . In order to have an "abuse" in reaching such determination, the result must be so palpably and grossly violative of fact and logic that it evidences not the exercise of will but perversity of will, not the exercise of judgment but defiance thereof, not the exercise of reason but rather of passion or bias.

Although defendant and amici curiae have suggested that this Court abandon the Spalding definition, at least in the context of custody disputes, the disposition of this case does not require such an outright rejection. By using the word "palpable," the Legislature clearly contemplated a strict standard of review. Nevertheless, we do not agree with plaintiff that the Legislature's use of the word "palpable" should be construed as requiring an even stricter standard than Spalding — in essence, plaintiff proposes: palpability plus Spalding. A standard stricter than Spalding, as suggested by plaintiff, would effectively grant trial courts unfettered discretion over child custody matters and potentially jeopardize a child's best interests. Because we construe statutes in harmony with the spirit and intent of the act and in a manner that does not achieve absurd or unreasonable results, In re State Hwy Comm, 383 Mich. 709, 714-715; 178 N.W.2d 923 (1970), we instead conclude that the Legislature's reference to "palpable," the same word this Court used in Spalding, manifests its intent to adopt a high standard of review not significantly unlike the Spalding standard.

To whom custody is granted is a discretionary dispositional ruling. Cooley v Cooley, 329 Mich. 91, 95; 44 N.W.2d 884 (1950). Therefore, a custody award should be affirmed unless it represents an abuse of discretion. While the abuse of discretion standard is strict, it does not afford trial courts unfettered discretion in awarding custody. The court's exercise of that discretion is already limited by the statutory best interest factors, MCL 722.23; MSA 25.312(3), each of which must be carefully evaluated by the trial judge. Wolfe v Howatt, 119 Mich. App. 109; 326 N.W.2d 442 (1982). Further, as noted above, a court's ultimate finding regarding a particular factor is a factual finding that can be set aside if it is against the great weight of the evidence. Therefore, a court's discretion in weighing the evidence regarding a particular factor is not unlimited; rather, it must be supported by the weight of the evidence. See Baker v Baker, 411 Mich. 567, 584-585; 309 N.W.2d 532 (1981).

C

Lastly, the custody act provides that questions of law are reviewed for "clear legal error." MCL 722.28; MSA 25.312(8). Because we are unable to discern any substantial distinction between "clear legal error" and ordinary "legal error" we regard the word "clear" as a legislative caution against result-oriented jurisprudence that might operate to the detriment of a child's interest in maintaining a stable custodial environment. When a court incorrectly chooses, interprets, or applies the law, it commits legal error that the appellate court is bound to correct.

D

To alleviate any confusion that prior decisions of this Court may well have engendered, we expressly recognize that the standards of review described in this section preclude de novo review of custody cases by appellate courts. In the context of appellate review, a truly de novo review would be a review of the record without deference to prior proceedings; on the basis of the record, the appellate court would make its own findings of fact and conclusions of law. Although, courts have recited that review of custody cases is "de novo limited by the statute," they have not actually accorded a truly de novo review. See Baker, supra. Indeed, the extent to which the statutory standards "limit" de novo review is tantamount to abolishing de novo review. Today, we recognize as much by holding that review of custody orders is not de novo.

III

Having considered the proper standards of review in custody cases, we next turn to the Court of Appeals opinion. The Court held that the trial court committed clear legal error with regard to factors b, e, and f. As explained below, we agree with the Court of Appeals finding of legal error with respect to factors e and f, but we do not find error in the trial court's analysis of factor b. Further, in accord with Dempsy v Dempsy, 409 Mich. 495; 296 N.W.2d 813 (1980), we hold that upon a finding of error, appellate courts should remand to the trial court unless the error was harmless.

A

The Court of Appeals held that it was clear legal error for the trial judge to provide only a one-sided account of the facts when considering factor b. We disagree. While we believe that the most effective appellate review obviously would result from a more thorough account of the court's balancing, we disagree with the Court of Appeals finding of clear legal error on this issue. The trial court's findings in this case minimally satisfy the requirements of the court rules and case law.

MCR 2.517, which governs findings of fact by the trial court, provides in part:

(A) Requirements.

(1) In actions tried on the facts without a jury or with an advisory jury, the court shall find the facts specially, state separately its conclusions of law, and direct entry of the appropriate judgment.

(2) Brief, definite, and pertinent findings and conclusions on the contested matters are sufficient, without over elaboration of detail or particularization of facts. [Emphasis added.]

In the context of custody cases, this Court has refused to interpret this rule as requiring a recitation of all the evidence considered. In Baker, supra at 583, the Court explained:

Neither the mandate of the Child Custody Act nor our requirement that in non-jury cases the "court shall find the facts specially and state separately its conclusions of law thereon" requires in cases involving child custody decisions that the court comment upon every matter in evidence or declare acceptance or rejection of every proposition argued.

The Baker Court explained that the trial court's failure to address the myriad facts pertaining to a factor does not suggest that the relevant among them were overlooked. Id.

In the instant case, the factors not discussed by the trial court in ruling on factor b were: the defendant's involvement in the academic affairs and extracurricular activities of the children and the children's reliance on the defendant to answer questions about sexual maturation. The record does not suggest that those relevant facts were overlooked by the court. Instead, because the court only found "some" difference between the parents, rather than a "significant" or "substantial" difference, it would seem that the court did consider those facts.

Because the trial court complied with the court rules and case law, the "one-sided account" is not "clear legal error."

B

Factor e requires the trial court to consider "[t]he permanence, as a family unit, of the existing or proposed custodial home or homes." MCL 722.23(e); MSA 25.312(3)(e). In the instant case, the trial court focused on the " acceptability of the custodial home," as opposed to its permanence. It stated its finding as follows:

The Court is satisfied that either parent would provide permanence, as a family unit, and would offer acceptable custodial homes. It is undisputed that plaintiff as a father is accustomed to and willing to preform [sic] the day to day jobs to maintain a household. The Court feels that some weight should be assigned in favor of plaintiff because the evidence shows that defendant had been out of the home in the evening many times and thus not caring for the family while plaintiff has been present on those occasions.

Because acceptability of the home is not pertinent to factor e, the panel found that it was legal error for the trial court to consider it. We agree with the Court of Appeals. The facts relied upon and expressed by the judge relate to acceptability, rather than permanence, of the custodial unit. Therefore, the trial court's error seems to go beyond mere word choice.

C

Standing alone, the trial court's error on factor e might be deemed harmless; however, we also agree with the Court of Appeals that there was legal error in the trial court's evaluation of factor f, which requires the court to consider the parties' "moral fitness." MCL 722.23; MSA 25.312(3). The trial court found this factor favored the plaintiff and supported its conclusion with the fact that the defendant had engaged in extramarital affairs. It stated:

The Court is of the opinion that the moral fitness factor favors plaintiff under the circumstances. Defendant does not dispute that she engaged in extra-marital affairs during the past few years. The Court is of the opinion that this demonstrates a poor moral example for the children and thus finds this factor in favor of plaintiff.

In finding that the trial court committed clear legal error in its analysis of factor f, the Court of Appeals explained:

Although the record reveals that defendant did not admit that she had engaged in extramarital affairs, defendant apparently admitted one affair during the friend of the court hearings and there was some evidence of another. Even assuming that there was evidence supporting the existence of two extramarital affairs, the evidence was clear that the children had no knowledge of them. Given that the children had no knowledge of defendant's affairs, there was no reason for finding that defendant's affairs established a poor moral example for the children. Under the circumstances, the affairs could establish a poor moral example only if the children knew about them. Because there were no findings of fact demonstrating a poor moral example for the children, there was no reason for finding this factor in favor of plaintiff.

Moreover, the trial court's conclusion that defendant's affairs demonstrated a poor moral example for the children evinced a clear legal error on a major issue. In Truitt [ v Truitt, 172 Mich. App. 38, 46; 431 N.W.2d 454 (1988)], this Court stated that "[s]tanding alone, unmarried cohabitation is not enough to constitute immorality under the Child Custody Act." Previously, a panel of this Court also observed that a mother's acts of adultery do not necessarily preclude her from having custody of her children. Williamson v Williamson, 122 Mich. App. 667, 673; 333 N.W.2d 6 (1982). Even if defendant engaged in extramarital affairs, this finding, standing alone, is not sufficient to support the trial court's conclusion that defendant behaved immorally. This finding is a clear legal error that requires reversal. [ 200 Mich. App. 514-515.]

Thus, the Court of Appeals concluded that extramarital conduct of the plaintiff, which was unknown to the children, did not, as a matter of law, bear on the defendant's moral fitness. While, for reasons explained below, we agree with the Court of Appeals, we also note that by focusing exclusively on the parent's "moral example" the Court of Appeals, like the trial court, overlooked the critical words in factor f: "fitness of the parties involved."

Factor f (moral fitness), like all the other statutory factors, relates to a person's fitness as a parent. To evaluate parental fitness, courts must look to the parent-child relationship and the effect that the conduct at issue will have on that relationship. Thus, the question under factor f is not "who is the morally superior adult"; the question concerns the parties' relative fitness to provide for their child, given the moral disposition of each party as demonstrated by individual conduct. We hold that in making that finding, questionable conduct is relevant to factor f only if it is a type of conduct that necessarily has a significant influence on how one will function as a parent.

Extramarital relations are not necessarily a reliable indicator of how one will function within the parent-child relationship. While such conduct certainly has a bearing on one's spousal fitness, it need not be probative of how one will interact with or raise a child. Because of its limited probative value and the significant potential for prejudicially ascribing disproportionate weight to that fact, extramarital conduct, in and of itself, may not be relevant to factor f. To the extent that one's marital misconduct actually does have an identifiable adverse effect on a particular person's ability or disposition to raise a child, those parental shortcomings often may be reflected in other relevant statutory factors.

Unfortunately, extramarital relations are not uncommon incidents in divorces. Indeed, they are often the impetus for a marital breakdown. We do not take lightly the moral or destructive consequences of extramarital conduct on a marriage. Nonetheless, to punish infidelity at the risk of jeopardizing a child's best interests simply contravenes the overriding purpose of the Child Custody Act. Accordingly, we agree with the Court of Appeals to the extent that it found that the trial court committed legal error when considering the extramarital conduct under factor f.

IV

Finally, we address the proper disposition of custody cases following appellate review. In its strained effort to reconcile de novo review with the otherwise deferential standards of review in custody cases, the Court of Appeals concluded that upon finding error, it was required to engage in a de novo review to determine the proper custody arrangement.

[W]e conclude that this Court may review de novo the dispositional ruling of a trial court, but only after we have concluded, as we do here, that the trial court has either made findings of fact against the great weight of the evidence or committed a palpable abuse of discretion or a clear legal error on a major issue. In this case, the trial court committed the first and third of these prohibitions and we are, therefore, free to review de novo the custody issue. [ 200 Mich. App. 522.]

We decline to adopt the dispositional framework offered by the Court of Appeals. De novo review of the ultimate custodial disposition is inappropriate.

In accord with Dempsy, supra, upon a finding of error an appellate court should remand the case for reevaluation, unless the error was harmless. We further hold that on remand, the court should consider up-to-date information, including the children's current and reasonable preferences, as well as the fact that the children have been living with the plaintiff during the appeal and any other changes in circumstances arising since the trial court's original custody order.

We are aware that an appellate court's peremptory change of custody would be more expeditious than remanding to the trial court. However, we do not believe that the Legislature's express intent to expedite proceedings was to take precedence over the primary goal of the Child Custody Act, which is to secure custody decisions that are in the best interests of the child. The state's interest in child protection cannot be reduced merely to providing expedient proceedings. As amici curiae have pointed out: "It is a false efficiency to put `expediency' over accuracy."

Our disposition of this case recognizes that trial courts are in a superior position to make accurate decisions concerning the custody arrangement that will be in a child's best interests. Although not infallible, trial courts are more experienced and better situated to weigh evidence and assess credibility. Trial courts not only hear testimony and observe witnesses, but also may elicit testimony, interview children, and invoke other judicial resources to assure a thorough and careful evaluation of the child's best interests.

V. CONCLUSION

For the reasons set forth herein, we reverse that portion of the Court of Appeals decision that posits a standard of review different than that prescribed in § 8 of the Child Custody Act. Further, we reverse the Court of Appeals decision with respect to factor b, affirm its decision with respect to factors e and f, and remand to the trial court for reevaluation of its custody award and further disposition consistent with this opinion.

CAVANAGH, C.J., and BOYLE, J., concurred with BRICKLEY, J.