(dissenting).
I dissent because I believe that the majority opinion is a serious departure from well-settled rules governing the scope of appellate review. Moreover, I believe that it circumvents the clear intent of the legislature expressed by repeated legislative attempts, including very recent amendments to Minn.Stat. § 518.17, to eliminate inflexible and stereotypical presumptions in child custody cases. In so doing, the majority opinion ignores the proper roles of the legislature and the courts in these matters.
We have frequently been critical of the court of appeals for substituting its own findings of fact for those of the trial court. See, e.g., Sefkow v. Sefkow, 427 N.W.2d 203, 210 (Minn.1988); Pikula v. Pikula, 374 N.W.2d 705, 710 (Minn.1985). Nevertheless, that is precisely what happened in this case. By ignoring the proper scope of appellate review, the majority opinion will encourage more appeals in family law matters. In addition, it will encourage more reversals in a court where the reversal rate is already extraordinarily high.
This court has consistently held that “[ajppellate review of custody determinations is limited to whether the trial court abused its discretion by making findings unsupported by the evidence or by improperly applying the law.” Sefkow, 427 N.W.2d at 210; Pikula, 374 N.W.2d at 710. See also Weatherly v. Weatherly, 330 N.W.2d 890, 891 (Minn.1983); Berndt v. Berndt, 292 N.W.2d 1, 2 (Minn.1980). For almost a century, this court has said that, in determining who should have custody of minor children, the trial court has wide discretion; and we will not reverse unless there is a clear abuse of discretion. LaBelle v. LaBelle, 296 Minn. 173, 175, 207 N.W.2d 291, 292 (1973); Meinhardt v. Meinhardt, 261 Minn. 272, 276, 111 N.W.2d 782, 784 (1961); Wicklem v. Wicklem, 229 Minn. 478, 481, 40 N.W.2d 69, 71 (1949); Menke v. Menke, 213 Minn. 311, 313, 6 N.W.2d 470, 471 (1942); Dacey v. Dacey, 179 Minn. 520, 522, 229 N.W. 868, 869 (1930); Novotny v. Novotny, 152 Minn. 420, 422, 189 N.W. 258, 259 (1922); see Waldref v. Waldref, 135 Minn. 473, 473-74, 159 N.W. 1068, 1068 (1916); Arne v. Holland, 85 Minn. 401, 404, 89 N.W. 3, 4 (1902). Yet, it now appears that this court is saying that, while the findings of the trial court are not erroneous, either the wrong law was applied to the case or the conclusions were invalid.
In May of 1989, the legislature attempted to clarify the law by amending Minn. Stat. § 518.17, subd. 1, effective August 1, 1989, to provide that the child’s primary caretaker, as determined by the Pikula analysis, is only one factor to be considered in awarding custody and that “[t]he court may not use one factor to the exclusion of all others.” See Act of May 25, 1989, ch. 248, § 2,1989 Minn.Laws 834, 835-36. The majority seems to hold that, since that statute was not in effect at the time the trial court decided this case, the pre-1989 status of the law should apply, in which case the presumption of Pikula applies. Even if that were the case, the trial court makes it absolutely crystal clear in its findings that, even applying the Pikula standard, the presumption had been rebutted and the custody of the children in this case ought to go to Steven Maxfield. The majority opinion finds just the opposite result.
Everyone seems to agree that the best interests of the children should prevail. That standard, in my opinion, was not applied to this case. In order to demonstrate this point, I feel compelled to outline the history of the law governing contested child custody cases in Minnesota.
Until about 20 years ago, the law in Minnesota child custody cases was that, in almost all cases, the mother received custody of the parties’ young children unless she was unfit. See, e.g., Bennett v. Bennett, 277 Minn. 227, 152 N.W.2d 187 (1967); In re Alsdurf, 270 Minn. 236, 133 N.W.2d 479 (1965); Eisel v. Eisel, 261 Minn. 1, 110 N.W.2d 881 (1961). The rule in these cases rested on the stereotypical notion that it was automatically in the best interests of a young child to live with the mother. See Meinhardt v. Meinhardt, 261 Minn. 272, 276, 111 N.W.2d 782, 784 (1961) (“[Ojther things being equal, the welfare of children of tender years is best served by their being left in the care of their mother.”). This maternal presumption has become known as the “tender-years doctrine.” See *225Note, A Step Backward: The Minnesota Supreme Court Adopts a “Primary Caretaker" Presumption in Child Custody Cases; Pikula v. Pikula, 70 Minn.L.Rev. 1344, 1348, n. 18 (1986) (hereinafter “A Step Backward ”).
Starting in the late 1960’s, no-fault divorce advocates and various equal rights groups insisted that divorce statutes be amended to discard the tender-years doctrine and codify the rule that the best interests of the child be the test. See A Step Backward, supra at 1349 & n. 21. In 1969, the Minnesota Legislature codified the “best-interests-of-the-child” standard and prohibited consideration of the proposed custodian’s gender. Act of June 6, 1969, ch. 1030, § 1, 1969 Minn.Laws 2081, 2081-82. This court, however, basically ignored the legislature’s first attempt to eliminate the tender-years doctrine. See Reiland v. Reiland, 290 Minn. 497, 500, 185 N.W.2d 879, 881 (1971) (“This amendment does no more than express views contained in prior decisions of this court.”). In 1974, the legislature amended Minn.Stat. § 518.17, subd. 1 in order to provide specific guidelines and require the courts to consider several factors. See Act of Mar. 28, 1974, ch. 330, § 2, 1974 Minn.Laws 555, 555-56; see also A Step Backward, supra at 1351. However, when lower courts actually started to award custody to the father in certain cases, the people who originally favored the objective statutory factors began arguing that the law should move in the opposite direction. See A Step Backward, supra at 1373 n. 142 (noting that, in the early 1980’s, some studies indicated that mothers were losing two-thirds of all contested custody cases). Finally, this court adopted the Pikula rule, which provides that the primary caretaker has a presumption in his or her favor. The Piku-la presumption basically instructs trial courts to substitute a list of homemaking duties for “all relevant factors” expressly provided in Minn.Stat. § 518.17, subd. 1. See id. at 1359. The Pikula presumption, although neutral on its face, has a readily apparent disparate impact on fathers. In other words, we seem to have completed a full cycle and are attempting to get the tender-years doctrine in through the back' door so that, once again, it will be the mother who is invariably granted custody of the children unless she is found to be unfit. Whether we favor such a rule or not,1 that is not the law in Minnesota. See Sefkow v. Sefkow, 427 N.W.2d 203, 212 (Minn.1988). The recent amendments to Minn.Stat. § 518.17, subd. 1 are the legislature’s third attempt to force this court to abandon inflexible presumptions about who is best able to care for a young child. It seems to me that, in refusing to let go of the Pikula presumption, the majority, by judicial fiat, is ignoring the intent of the legislature and reviving the tender-years doctrine. See Op. at 222 (“the aim [of Pikula] was to place the child too young to express a preference where that child might best have the emotional security and nurturing environment so vital to the child’s life and development”).
With respect to the best interests of the children in this case, who can seriously argue against the conclusion that all of the children would be better off raised in Vern-dale, Minnesota, than Wilkes-Barre, Pennsylvania, in a low-cost public housing slum? In addition, the fact that Diane Maxfield took all the children from Verndale to Pennsylvania by deceit is also highly relevant in this case because it undermines the credibility of the rest of her testimony.2 It was only when dissolution proceedings were commenced and the children were returned in 1988 for the purpose of determining custody in connection with that dissolution that Steven was able to see his children on a regular basis. The trial court *226then awarded custody to Steven in August of 1988, and they have been in his custody since that time. The Pikula presumption is supposed to be premised on the importance of stability. Pikula, 374 N.W.2d at 711. Even if it were appropriate for an appellate court to re-weigh the evidence, how can it be seriously argued that it would now be in the children’s best interests to be uprooted for the third time and sent back to Pennsylvania?
The respondent argues that Steven must work during the day and even one weekend a month so that he wouldn’t be able to spend the time with his children that Diane could. That argument is obviously fallacious because the only way that Diane could spend more time with the children than Steven if she received custody would be if she didn’t work at all and was on welfare. If she gets a job, she’ll have to be away from home just as Steven or any other person who must work for a living.
The trial court made 66 separate findings of fact and, from those findings, arrived at 31 conclusions of law. The majority opinion indicates that its concern is not with the accuracy of the findings of fact, but with the conclusions drawn from those findings. The majority opinion summarizes in eight paragraphs the findings of the trial court; however, I do not believe those summaries accurately reflect the entire story. Rather than be accused of omitting facts in my own summaries, I find it necessary to attach all the findings of fact and conclusions of law of the trial court to this dissent as Exhibit A.
As one can readily see from the findings of fact, the trial court has done exactly what this court has said in numerous opinions a trial court should do. See, e.g., Rosenfeld v. Rosenfeld, 311 Minn. 76, 82, 249 N.W.2d 168, 171 (1976) (“court must make written findings which properly reflect its consideration of the factors listed in Minn. Stat. § 518.17, subd. 1”). Mechanically applied presumptions are not useful in child custody cases. Each case is unique. A trial court must consider a great deal of testimony and other evidence, much of it conflicting. The trial court is the appropriate finder of fact and is in the best position to judge the credibility of witnesses. I doubt whether one could find a case where a trial court has spent more time and made more detailed and careful findings of fact than has been done in this case. Moreover, I submit that the trial court’s conclusions of law were justified by those facts. I find it simply extraordinary that either the court of appeals or this court would take it upon itself to reverse a trial court when it can’t point to one of the findings as being erroneous. In other words, we have now reverted to a rule of law in Minnesota where an appellate court will dictate a result-orientated decision rather than a principled decision based on fact or law.
For all the above reasons, I would reverse the court of appeals and reinstate the decision made by the trial court.
. At least one commentator has argued for an overt resurrection of the tender-years doctrine on the grounds that, “in a society which continues to discriminate against women in every other aspect of life, a mother’s preferred status regarding custody of her children is not unlike the ‘reverse discrimination’ and preferred status urged on behalf of historically disadvantaged groups in a variety of other contexts.” Uviller, Fathers' Rights and Feminism: The Maternal Presumption Revisted, 1 Harv.Women’s LJ. 107, 130 (1978). While I do not support the idea that "reverse discrimination” is the best way to correct past wrongs, at least this approach would have the benefit of candor.
. In June 1987, respondent told her husband that she was going to go home to visit her parents in Pennsylvania, but would be back with the children and they would attempt to *226work out their marital differences. All along, she was contemplating terminating the marriage and removing the children to Pennsylvania, which she did. Thereafter, she even denied Steven’s attempts to visit the children or talk with them on the telephone. He then had to get a court order to see his children. At trial, Diane testified that she intended to marry an older man whom she met through newspaper advertisements. Since trial, those plans have been abandoned. Surely, the trial court is in the best position to determine Diane’s credibility and weigh that assessment along with other factors relevant to the best interests of the children.