Defendant wife appeals a divorce judgment awarding to plaintiff husband custody of their daughter. The court awarded custody of the couple’s son to wife. She also appeals that part of the judgment relating to visitation between husband and their son. We reverse and remand because the trial court did.not adequately explain application of the primary-care-provider criterion to the facts.

The parties were married in March 1979. During the course of the marriage wife became lonely, in part because husband worked long hours, and she began a relationship with a woman in 1988. Husband filed for divorce in April 1989, after a temporary separation. Following various unsuccessful attempts at reconciliation in 1988 and 1989, the couple separated permanently in October 1989, when wife moved to a nearby town to live with her lover. The couple’s son, Morgan, was five months old when wife left for the last time, and has resided with wife since her final departure. Wife left the couple’s seven-year-old daughter, Ashley, with husband. Before the final separation, wife had agreed in writing that husband would have custody of Ashley and she would have custody of Morgan, whom she was then carrying. She testified that she entered into this agreement only because she was “uninformed about parental rights [and] felt threatened by [husband].”

At trial, both husband and wife testified that husband worked from 50-70 hours per week in the summers and at least 40 hours per week in the winters at his job as director of marketing at a local inn. As of the date of trial, husband had Sundays and Mondays off. Husband retained the same schedule during the six months between separation and trial. Husband brought Ashley to work with him when she was not at school. She stayed in a vacant room, where she sometimes played with another *87child, and was supervised through a glass door by husband or other workers. She often ate meals at the inn restaurant. When Ashley was not at the inn with her father while he worked, she was in the care of a baby-sitter or in school.

As of the date of trial, wife was employed as a teacher, working from 8:00 a.m. to 12:30 p.m. during the school year. Both parties agreed that while the couple lived together, wife had primary responsibility for the shopping, laundering and meal preparation. Wife bathed, dressed and fed Ashley in the mornings, planned or prepared her noon meal at school, and prepared and ate the evening meal with her. She read to the child daily, although husband also did this “once in a while.” She took Ashley on educational excursions to museums and taught her about classical music, activities in which husband did not participate. Wife largely took responsibility for Ashley’s health needs and brought her to the dentist and doctor for regular examinations.

The court found that although wife had been Ashley’s primary-care-provider before the separation, husband had been Ashley’s primary-care-provider after that time. The court concluded it was in Ashley’s best interest to remain in husband’s custody. Morgan, however, had not left his mother’s primary care since he was born in May 1989, and the court concluded wife should have custody of him. Split custody, according to the court, was satisfactory because of the age difference between Ashley and Morgan and because they had not developed a relationship with one another.

Visitation was liberal. Wife was awarded parent-child contact with Ashley, with visitation from Friday through Sunday night every other weekend and Saturday on the intervening weekend. Visitation also included alternate holidays, the first two weeks of July and the first two weeks of August. Husband was awarded parent-child contact with Morgan on alternate Sundays and every Monday, as well as alternate holidays, until Morgan reached age two, after which visitation would include overnight stays and an additional four-week period in the summer. The court made additional accommodation for holiday contact between the two children once Morgan reached school age.

Wife claims that the trial court erred in granting husband sole legal and physical rights and responsibilities for Ashley *88based on its finding that he was the primary care provider. In addition, she argues that the court’s conclusions of law on this issue and the potential effect on Ashley of a change of custodian are not supported by the findings.

Analyzing the evidence on the primary-care-provider factor that wife’s counsel established through the testimony of various witnesses, the court found that “when the parties resided together, [wife] was the primary caretaker of Ashley. She took care of Ashley’s daily needs and was an attentive, loving mother towards her.” The court also found, however, that since Ashley was left in husband’s care when wife left six months earlier, husband had become the child’s primary-care-provider and wife no longer retained that status.

[Husband] has provided for [Ashley’s] physical and emotional needs and has become very involved in her daily activities and schooling. He insures she is fed, clothed, and has appropriate adult supervision. [Husband] has good parenting skills and a good sense of Ashley’s needs. [Husband] loves his daughter very much.

Based on these findings, and on the conclusion that a change would be disruptive to the child’s life, the court granted split custody, thereby maintaining the “status quo.” In so doing, the court stated that “great weight should be given to the child’s relationship with his or her primary caretaker.” This was a reference to 15 V.S.A. § 665(b)(6), the statutory factor that requires the court, in making a custody order, to consider, along with seven other factors, “the quality of the child’s relationship with the primary care provider, if appropriate given the child’s age and development” (criterion six)!

I.

We recognize that the trial court has broad discretion in custody matters. Myott v. Myott, 149 Vt. 573, 578, 547 A.2d 1336, 1339 (1988) (“trial court has broad discretion in a custody matter, and we must affirm unless the discretion is ‘erroneously exercised, or was exercised upon unfounded considerations or to an extent clearly unreasonable in light of the evidence’”) (quoting Jensen v. Jensen, 141 Vt. 580, 581-82, 450 A.2d 1155, 1156 (1982)). Findings of fact, from which conclusions of law *89flow, will not be set aside unless clearly erroneous. V.R.C.P. 52(a). When the relevant legal criteria, among them the eight factors set out in 15 V.S.A. § 665(b), are applied to the facts, it is important for the appellate court to know how the trial court weighed the facts and blended the standards to arrive at the conclusion. In other words, we need to review the explanation as well as the found facts to determine if the application of the criteria to the facts is sound. If the explanation is equivocal, indicating that a misapplication of the law may have occurred, the trial court decision is insufficient. See Klein v. Klein, 153 Vt. 551, 558, 572 A.2d 900, 904 (1990) (findings must address reasoning and weight given various factors to support conclusions reached). We conclude in this case that the court did not adequately explain its application of criterion six.

We agree with the trial court that the factor relating to the primary-care-provider under criterion six “should be entitled to great weight unless the primary custodian is unfit.” Harris v. Harris, 149 Vt. 410, 418, 546 A.2d 208, 214 (1988). This does not create a presumption in favor of the primary-care-provider, but instead allows the court to give due consideration to the primary custodian in evaluating the child’s best interests. See id. at 418-19, 546 A.2d at 214 (presumption that the primary custodian will be awarded custody if fit would be inconsistent with the statutory scheme because the court must consider each factor). A person is entitled to this consideration, however, only if the primary-care-provider has been correctly identified.

The factor relating to the primary-care-provider is one consideration in a nonexhaustive list outlining the appropriate considerations in determining the best interests of the child. See 15 V.S.A. § 665(b) (“the court shall be guided by the best interests of the child, and shall consider at least [eight] factors”). We have not enunciated a definitive standard for determining the identity of the primary-care-provider under § 665(b)(6). Nevertheless, mere physical custody by one of two fit parents, during the time the estranged spouses live “apart” to satisfy the no-fault divorce requirements, should not in itself cause a former primary-care-provider to lose that status. See 15 V.S.A. § 551(7) (a divorce may be decreed “[w]hen a married person has lived apart from his or her spouse for six consecu*90tive months and the court finds that the resumption of marital relations is not reasonably probable”); cf. Emmons v. Emmons, 141 Vt. 508, 511, 450 A.2d 1113, 1115 (1982) (“Merely leaving the homestead in and of itself is not sufficient to justify a finding of fault [relevant to distribution of marital property]. Such a result would substantially circumvent the no-fault divorce provision authorized by our statute.”).

A contrary holding may cause a primary-care-provider wishing to leave the home to uproot children from the marital residence solely to remain, in the view of the court, the primary-care-provider. See 15 V.S.A. § 551(7). This kind of parental strategizing is inimical to the best interests of children. Attention should be directed to the needs of the children rather than the actions of the parents. See Bissonette v. Gambrel, 152 Vt. 67, 70, 564 A.2d 600, 602 (1989) (“focus of the court’s decision must be the best interest of the child, not equity between the parties”); Lafko v. Lafko, 127 Vt. 609, 618, 256 A.2d 166, 172 (1969) (in custody dispute “opposing desires of hostile parents ... must yield to the paramount consideration of the children’s well-being”).

In Harris, we underscored the importance of balancing the best interests-of-the-child factors. Agreeing that the primary custodian factor, if that custodian is fit, is entitled to great weight, we stated that “[t]he exact weight cannot be determined unless there is evidence of the likely effect of the change of custodian on the child.” Harris, 149 Vt. at 418-19, 546 A.2d at 214. Presumably, the court followed this analysis. Citing Harris, the court stated, “Ordinarily, a child should not be removed from the care of a primary caretaker if that caretaker is fit.” Based on this statement, together with the rest of its analysis, it appears that the court in concluding that husband was the primary-care-provider may have given Ashley’s present living situation undue weight by not giving any weight to wife’s conceded primary care of Ashley for nearly seven years.

We decline to follow the unyielding approach of some courts which bifurcate the pre- and post-separation periods in determining the primary-care-provider. See, e.g., Efaw v. Efaw, 400 S.E.2d 599, 602 (W. Va. 1990) (“‘[t]he primary caretaker is that natural or adoptive parent who, until the initiation of the divorce proceedings, has been primarily responsible for the car*91ing and nurturing of the child’”) (quoting Garska v. McCoy, 167 W. Va. 59, 278 S.E.2d 357 (1981)) (emphasis added). Instead, the inquiry should focus on all relevant periods of the child’s life, rather than exclusively on the period immediately preceding trial. See Draper v. Draper, 556 So. 2d 210, 214 (La. Ct. App. 1990) (child’s residence with mother for approximately one year prior to custody trial had been unstable and thus did not mandate a finding that his best interest required him to remain with her where stability and continuity had not played a significant role in his life while he lived with her); In re Thompson, 103 Or. App. 458, 460-61, 797 P.2d 1077, 1078 (1990) (wife, who moved out of family home and left eighteen-month-old child to minimize emotional impact of separation, was properly deemed primary parent; fact that husband provided primary care for year since separation was not determinative). We decline to excerpt any period from the child’s life in determining who, out of two competing parents, has provided the nurturing that would make the parent the primary-care-provider. The trial court, on the other hand, apparently accepted a per se rule that the parent with physical custody at the time of the divorce hearing is the primary-care-provider.

In sum, we cannot afford meaningful review because we cannot tell from the court’s decision if husband was the overall primary-care-provider under criterion six in light of all the circumstances, including the fact that wife had provided the primary care before the separation. On this record, we are left to speculate as to the reasons the court favored husband over wife with respect to Ashley’s placement. This is especially so because post-separation “primary care” of Ashley was relatively minimal in comparison to that before separation. At the time of trial, husband worked a schedule which allowed him to spend at most one full day — Sunday—with Ashley. Although husband could spend Mondays with Ashley during the summer, her school schedule made this impossible during the months school was in session. Upon separation, Ashley was transferred from the almost full-time care of her mother to the full-time care of baby-sitters, husband’s co-workers, and husband, who often took her to work. Ashley’s school, extracurricular activities and community would not have changed had wife been granted custody. When Ashley lived with husband, much of the *92care provided to her by others replaced care she had formerly received from a parent.

We realize that a parent should not be penalized for time at work spent away from a child, but consideration of the child’s best interests “requires comparison of the attributes of each parent.” Bissonette, 152 Vt. at 69, 564 A.2d at 601. Concededly, husband provided care after wife left the family home, but this was a function of wife’s absence, which made him the only parent at home. The issue cannot be decided solely on the performance of one parent. Id. The husband’s expert, who testified that husband was a good father and custody of Ashley should remain with him absent “compelling reason,” had investigated only husband’s relationship with Ashley. The court failed to justify its conclusion by an analysis of the facts over the entire relevant time period.

II.

If a different custody disposition is ordered on remand, the court may be required to alter the visitation order. Still, wife’s challenge to the original visitation arrangement is unfounded. Wife argues that the trial court’s order respecting visitation between husband and Morgan allowed for excessive parent-child contact. Even though the testimony indicated that husband worked more during the summer, the trial court did not abuse its discretion in awarding husband four weeks of visitation with Morgan during the summer. See Palmer v. Palmer, 138 Vt. 412, 414-15, 416 A.2d 143, 144 (1980) (visitation arrangement is for trial court’s discretion). Considering the facts as they existed at the time of trial, the court’s order was appropriate.

Wife’s contention that she was denied a fair trial based on remarks by the trial judge is also without merit.

On remand, the court is to apply the factors of § 665(b) to the situation at the time of hearing on remand, not solely to the situation as it existed at the time of the final divorce hearing. Cf. Klein, 153 Vt. at 557, 572 A.2d at 904 (court must rely on “most recent information available” in fashioning maintenance on remand).

Reversed in part and affirmed in part.

*93Dooley, J.,

dissenting. Anybody who reads the majority opinion carefully will recognize that despite the remand, the Court has ruled that custody of Ashley Nickerson must be awarded to defendant as a matter of law. Although the opinion suggests that the defect in the trial court decision is the failure to explain its rationale, I think the opinion can be understood only as a holding that defendant remained the primary-care-provider. Once the rule of law is extracted from the majority opinion, it is impossible for plaintiff to prevail.

While I believe this result is unjustified in this case, it is a hallmark of a greater error in the majority opinion. The majority opinion is in fact a trial court opinion. Most of the “facts” stated in the first few pages were not found by the trial court, and some are disputed. The opinion narrowly construes the trial court decision in order to criticize it. Little discretion is accorded to the trial judge in her evaluation of the evidence. Moreover, the “error” found by the majority involves an issue not raised below. In short, this Court has now “tried” this case on theories different from those presented to the trial court and awarded custody of the child based on its fact-finding and evaluation of the evidence.

I do not believe it just to abandon the proper role of appellate review even where the hindsight of the Justices of this Court shows we would have reached a different result. Such appellate decisionmaking of this kind inevitably makes bad law that haunts us in the future when we try to stay within our proper role.

It is helpful at the outset to examine the procedural posture of the case. Not only did defendant leave the child with plaintiff and agree that he should have custody, she did not contest custody when plaintiff filed for divorce in 1989. In January of 1990, she filed a counterclaim requesting that plaintiff receive custody of Ashley and she have visitation rights. It was not until March, shortly before the divorce hearing, that she changed her mind and sought custody.

Both parties submitted trial memoranda of law dedicated almost entirely to the effect of defendant’s lesbian relationship on the custody question. Plaintiff briefly stated that the primary-care-provider factor, as specified in 15 V.S.A. § 665(b)(6), was in his favor. Even more briefly, defendant stated that the evidence *94would show that she was the children’s primary-care-provider. During the trial, the trial judge gave a “weather report” based on the evidence to that point. She indicated that she was leaning to keeping the status quo because of the disruption of changing custody. Despite this statement, defendant never addressed the primary-care-provider question in her closing argument. Neither party submitted requests for findings of fact. There were no relevant post-trial motions.

Except in very exceptional cases, we have insisted that issues of law be raised in the trial court before they will be considered by this Court. I have no doubt that if defendant had raised the application of the primary-care-provider factor to a situation where the custodian of the child at the time of trial is not the person who was the primary-care-provider prior to separation, the trial court would have resolved it directly. Instead, this issue is resolved for the first time on appeal. It is particularly ironic that the majority faults the trial court for not fully explaining its decision when defendant failed to take the minimal steps necessary to obtain an explanation. Without preservation and any help from the lawyers, the trial judge is left with the unmeetable burden of explaining everything an appellate court might want to know in response to creative new arguments raised by the losing party for the first time on appeal.

In Varnum v. Varnum, 155 Vt. 376, 382-83, 586 A.2d 1107, 1110-11 (1990), a custody case where the claim on appeal was improper consideration of the mother’s religious beliefs in violation of her constitutional rights, we emphasized the critical interest of the children in family stability that is undermined by treating issues for the first time on appeal. After balancing the mother’s constitutional rights against the interests of the children, we held: “Even with the important rights and interests defendant seeks to vindicate, we think the balance tips decidedly in favor of enforcing rules of preservation to avoid the impact of lengthy delays on the well-being of the children.” Id. at 383, 586 A.2d at 1111. This is not a case where appellant is excused from preservation because she was unaware of the trial court’s error until the decision was rendered. The court’s position was crystal clear in its “weather report,” and appellant failed to respond to that position with the arguments she makes here. As a result, we are abandoning the requirement of preser*95vation. Almost two years after the trial court’s decision, this Court will overturn the custody award based on an issue never presented to the trial court. I cannot believe that the trauma to the child is outweighed by the possible gain of a better custodian.

Even if there had been preservation, I could not accept the majority opinion. While couched in language that criticizes the trial court for failure to explain its decision, the majority’s reasoning on the primary-care-provider argument is really that the trial court improperly found plaintiff to be the primary-care-provider and, thus, improperly weighed that factor in his favor. Instead, the majority concludes that defendant was the primary-care-provider and never lost that status. Thus, in the majority’s view, this very important factor should weigh heavily in defendant’s favor.

I find the majority’s analysis of both the facts and the law to be in error. The determination of which parent is the primary-care-provider is primarily one of fact subject to the clearly erroneous test for setting it aside on appeal. See Bissonette v. Gambrel, 152 Vt. 67, 70, 564 A.2d 600, 601 (1989). As the majority recognizes, we have not accorded a custodial presumption in favor of the primary-care-provider. Instead, we accord great weight to this factor with the exact weight to be determined based on the likely effect of a change of custodian on the child. See Harris v. Harris, 149 Vt. 410, 418-19, 546 A.2d 208, 214 (1988).

Both the plaintiff and defendant used a child psychologist as an expert witness. Each respective psychologist interviewed the child and the parent who hired the psychologist, and observed the interaction between the parent and child. Only plaintiff’s expert, however, directly responded to the primary-care-provider issue. He found a “mutual bond” between plaintiff and the child and that plaintiff “is a real psychological parent.” He went on to find that the child was flourishing in plaintiff’s care, and that in the absence of a compelling circumstance, he would not recommend changing custodian. He found no compelling circumstance for a change of custodian. In addition to the expert’s testimony, plaintiff testified that he performed the day-to-day functions of the care provider. Other witnesses who observed plaintiff and the child together provided similar testimony.

*96Defendant’s expert evidence was in some ways comparable. The psychologist found defendant to be a fit parent with a good relationship to the child. The expert offered no opinion, however, on who should have custody and no opinion on the effect of a change of custodian at the time of trial.

It is clear that much of the majority’s criticism of the trial court is based on the majority’s reevaluation of the evidence and separate fact-finding. Relying primarily on the expert evidence, the trial court found that plaintiff had become the primary-care-provider and gave that factor the weight recommended by plaintiff’s expert witness.

The majority is giving different weight to the testimony of plaintiff’s expert, as shown by its comments on page 92 of the majority opinion.* It is, of course, settled law that the weight to be assigned to evidence is for the trial court, not this Court. The evaluation of the expert testimony is not the only area where the majority has reweighed the evidence. Another critical example is the statement of the majority that the child “was transferred from the almost full-time care of her mother to the full-time care of baby-sitters, husband’s co-workers, and husband, who often took her to work.” This statement is wildly exaggerated. It fails to take into account that plaintiff’s work schedule allowed for two full days of availability, Sunday and Monday, separate from the day, Saturday, that the child is with defendant. It further ignores that defendant works part time and while plaintiff and defendant were together, the child was often with the same baby-sitter she is with now. Thus, the “full-time care of baby-sitters” is actually three or four afternoons a week, at least part of which also occurred before the separation. Obviously, the trial court did not see a flexible forty-hour work schedule as the same impediment to good parenting as the majority sees it.

*97The majority opinion presents three interrelated criticisms of the trial court’s application of the law. None are warranted. The first is that the trial court “apparently accepted a per se rule that the parent with physical custody at the time of the divorce hearing is the primary-care-provider.” The heart of the trial court’s analysis was:

Both Plaintiff and Defendant point to minor shortcomings in one another’s parenting skills. However, neither party has given a compelling reason to alter the present custodial arrangement. Plaintiff and Defendant are fit parents who can provide for their children’s needs. The children have adjusted to their parents’ separation and more change in their lives would be disruptive. Ashley, in particular, has strong ties to the area in which she presently resides.

As noted above, the “compelling reason” analysis was urged by plaintiff’s expert witness. There is nothing of a per se rule in this analysis. If there had been no expert testimony in the trial court, or if defendant had seriously contested the primary-care-provider issue in the terms it is raised here, one could understand the appearance of a per se rule. In context, it is a mischaracterization of the trial court conclusion.

The second asserted error, and the one that the majority identifies as the ground for its reversal, is that the trial court failed to indicate whether the proper analysis was used and thus “we are left to speculate as to the reasons the court favored husband over wife with respect to Ashley’s placement.” We have occasionally used this rationale where we cannot discern why the trial court rendered its decision or there is an internal inconsistency in the explanation. See, e.g., Klein v. Klein, 150 Vt. 466, 472, 555 A.2d 382, 386 (1988); McCormick v. McCormick, 150 Vt. 431, 438, 553 A.2d 1098, 1103 (1988). It is a misuse of this rationale to apply it to a case where the trial court is crystal clear on what was decided, and why, but the majority disagrees with the decision. The only thing we are left to speculate on in this case is why the trial court didn’t anticipate the majority’s view of the law.

If we use the “anti-speculation” rationale in a case like this, we will inevitably be left to speculate on what the law is after this opinion. At one point, the majority opinion is clear that it is irrelevant that one spouse becomes the primary-care-provider *98after the other spouse, who was the primary-care-provider, leaves the home. Later in the opinion, the trial court is directed to engage in some sort of weighing process, balancing the preseparation care-giving against the post-separation care-giving to determine the “overall primary-care-provider.” I doubt that trial courts would be able to figure out how to award points to implement this legalistic process. It certainly will not be obvious how this process is better in determining the best interest of the child than the approach used by the trial court in this case. None of the justifications for a rule giving special weight to the primary-care-provider are advanced by this process. See Crippen, Stumbling Beyond Best Interests of the Child: Reexamining Child Custody Standard-Setting in the Wake of Minnesota’s Four Year Experiment with the Primary Caretaker Preference, 75 Minn. L. Rev. 427, 440-52 (1990) (purpose of primary caretaker preference is to protect parent-child bonding, promote certainty in custody adjudication and preserve gender-neutrality and the general best interests of the child). While I strongly disagree with a rule that ignores post-separation care provision, I find it far preferable to the mire of weighing unlike things with the risk that carefully considered custody determinations will be overturned because the legalistic equation was not accurate. I fear that the real result of this opinion, and its determination that the trial court failed to explain its decision, will be a plethora of appeals to make us explain our decision. See id. at 452 (primary caretaker preference in Minnesota, since abandoned by legislative action, “caused an explosion of litigation”).

The third error involves the legal principles controlling the determination of the primary-care-provider in a case like this. The real holding of this case is that a parent who leaves the home without notice and without continuing to be the primary-care-provider is, as a matter of law, entitled to primary-care-provider status in the custody analysis. I believe this position is rigid and inappropriate.

Defendant not only voluntarily relinquished custody to plaintiff, she failed to contest custody until shortly before the divorce hearing. She gave no reason for failing to take Ashley with her to her new home. As a result, the primary-care-provider role shifted to plaintiff, and the child adjusted to that situ*99ation to the point that it would be traumatic to make a further change.

The majority gives one reason why it is important to continue the legal status of the primary-care-provider with the parent who leaves the home despite the change in the status in fact. It should not be the controlling consideration. In determining custody, we are governed by the best interest of the child and must look at the custody determination from the child’s perspective. See Price v. Price, 149 Vt. 118, 125, 541 A.2d 79, 83 (1987). In a battle between fit and loving parents, the child’s interest is in stability and continuity of relationships and surroundings. Any rule that allows a child, having been left in the custody of one parent, to be reclaimed at will by the other parent despite the bonding that has occurred and the psychological damage to the child is a bad rule.

Ironically, the case that the majority criticizes as “rigid,” Efaw v. Efaw, 400 S.E.2d 599 (W. Va. 1990), is an example of the kind of flexible approach that is far preferable to the majority’s approach. Although West Virginia is now the only state that gives custodial preference to the primary-care-provider who is fit, irrespective of other factors, it is flexible in determining whether either parent achieves the preference. The facts in Efaw are similar to those here in that the primary-care-provider, who was the mother, left the children when she moved in with another man. The father then became the primary-care-provider and remained so until the divorce hearing. The court found that neither parent was entitled to the preference under the primary-care-provider rule and that custody must be determined based on the best interest of the children. It awarded custody to the father, in large part because “[t]he children have developed a stable relationship with their father and grandparents .... To remove the children from such an established environment would jeopardize their emotional stability . . . .” Id. at 603. See also Mills v. Gorrick, 381 S.E.2d 273, 276-77 (W. Va. 1989) (trial court has discretion to award custody to father where evidence supports conclusion that wife intentionally relinquished custody to him in order to maintain a relationship with another man); Dempsey v. Dempsey, 306 S.E.2d 230, 231-32 (W. Va. 1983) (where mother had relinquished care to father *100shortly before divorce action, court could find that no primary-care-provider presumption arose and award custody to husband based on the best interest of the child).

Efaw is consistent with how other courts have handled a shift of primary-care-provider. See, e.g., Davis v. Davis, 749 P.2d 647, 648-49 (Utah 1988). It is also how the trial court handled this case. A fair evaluation of the trial court’s decision shows that it is not based on a rigid application of the primary-care-provider rule. Instead, it is based on the expert testimony of plaintiff’s psychologist, which the court chose to believe and follow. The decision should be affirmed on that basis.

Virtually all custody decisions emphasize that the trial court has broad discretion in custody cases and its custody award cannot be overturned unless its “discretion was erroneously exercised, or exercised upon unfounded considerations, or to an extent clearly unreasonable in light of the evidence.” Peckham v. Peckham, 149 Vt. 388, 389, 543 A.2d 267, 268 (1988). The majority states the standard of review on appeal, but emphasizes all the exceptions to the point where they justify what is essentially de novo review of both the facts and the law. If the majority followed the well-established standard of review, it would have to affirm the custody award in this case as within the trial court’s broad discretion. Accordingly, I dissent.

dissenting. I join Justice Dooley’s dissent, except that I believe the primary-care-provider issue was adequately preserved.