This appeal involves a dispute over whether an award to defendant of $25,000 as a property distribution, plus $5,000 for attorney’s fees, is an abuse of discretion. Defendant-wife claims that the total amount of money awarded to her represents only 20% of the marital assets, a decidedly unjust result given the evidence of battering she endured over the course of this four year marriage.
We need not reach the issue of abuse of discretion, because the oral findings of fact issued extemporaneously at the conclusion of the trial are sufficiently ambiguous and incomplete to warrant another hearing on property distribution. In particular, two evidentiary areas cause us to reverse and remand.
I.
Plaintiff-husband acquired the marital residence before the marriage. At trial, he evaluated its worth at about $65,000. The court did not believe this testimony, stating in findings issued orally from the bench, “[Plaintiff] tried to tell us that the house was worth sixty or sixty-five thousand dollars which is ... ridiculous. His own expert witness reported to us a value of $125,000, and that is what he thinks it’s worth.” The mortgage balance of about $15,000 on the house was undisputed. Therefore, the equity in the house was somewhere between $50,000 and $110,000.
Putting aside the value of the considerable personal property awarded to the husband, an award of $30,000 (including $5,000 in attorney’s fees) to the wife was about one quarter of the equity in the house if it is valued at the high end. We assume the court — as evidenced by the above remarks — was persuaded more by the $125,000 opinion, but it made no finding concerning the value of the house. Sullivan v. Sullivan, 147 Vt. 407, 408, 518 A.2d 33, 34 (1986). It also made no finding about the probable increase in the fair market value of the house over the course of the marriage. On remand, the court should clarify any increase in the value of the house during the marriage.
*203II.
At another point in the findings, the court said in reference to the $25,000 property distribution award: “That’s a number that at times during the trial was as big at one point as $50,000, and at another point it turned out to be $25,000.” The findings do not explain what caused the award to plunge from $50,000 to $25,000.
The court observed:
During the four years of the marriage, Karen worked consistently or persistently at her job at Castleton State College, 12 months a year with her vacations and home days off. Stan, on the other hand, was casual about his work. He more or less worked when he had to and didn’t work when he could get away with not working.
. . . Parental responsibility was awarded to Stan’s previous wife and after two years of that situation, Stan was successful in getting custody or parental responsibility of the children. Karen helped substantially in making that a reality in helping to provide a normal and comfortable home with both the mother and father when the children returned in June of ’81 even though Karen and Stan were not married yet.
This reflects favorably on the wife and fails to support the court’s rumination in reducing an award as “big as $50,000” to $25,000.
The following findings make us wonder just how the court viewed defendant’s claims of abuse by plaintiff.
We mentioned earlier the alcohol and we’d say more on it later. I don’t think it is important for the record to show you how we view it. Karen’s version of it is much more accurate. ... The marital misdeeds that have been attributed to Stanley, most of them, we don’t believe. We do recognize that there was a certain amount of misbehavior; that there may be these temper tantrums and items of misbehavior, but the strangling with the hands and violence and threats that were described by Karen have been blown way out of pro*204portion as evidenced by the fact that she stayed throughout the four years of marriage. The children were in the house and don’t relate any corroboration at all regarding these things that she told us happened. I think that they’re blown up by her own hurt with what happened to the marriage.
The court’s remark about staying in the relationship marked by violence manifests the “popular misconception[ ]... that women who remain in battering relationships are free to leave their abusers at any time.” State v. Kelly, 97 N.J. 178, 192, 478 A.2d 364, 370 (1984); see State v. Anaya, 438 A.2d 892, 894 (Me. 1981) (“abused women often continue to live with their abusers even though beatings continue”); State v. Hennum, 441 N.W.2d 793, 798 (Minn. 1989) (“common misconception that a normal or reasonable person would not remain in such an abusive relationship”); People v. Torres, 128 Misc. 2d 129, 133, 488 N.Y.S.2d 358, 361 (Sup. Ct. 1985) (battered women “[n]umbed by a dread of imminent aggression . . . are unable to think clearly about the means of escape”); State v. Ciskie, 110 Wash. 2d 263, 276, 751 P.2d 1165, 1172 (1988) (en banc) (characteristic of battered women’s syndrome is the “unlimited patience that these battered women have with regard to hoping ... ‘that everything will be all right’”).
We are mindful that the court is the ultimate judge of credibility, which can be based on the demeanor of a witness and how the witness articulates an account. Belanger v. Belanger, 148 Vt. 202, 204, 531 A.2d 912, 913 (1987). Here, however, the court’s reasons for discrediting the defendant’s testimony about domestic violence are ones rendered suspect in the literature on the subject. See State v. Kelly, 97 N.J. at 190-97, 478 A.2d at 369-73 (discussing battered women’s syndrome with citations to the literature); see also L. Walker, Terrifying Love: Why Battered Women Kill and How Society Responds (1989); Mather, The Skeleton in the Closet: The Battered Woman Syndrome, Self-Defense, and Expert Testimony, 39 Mercer L. Rev. 545, 550-55 (1988). We think the findings were inadequate to afford meaningful review, Klein v. Klein, 153 Vt. 551, 558, 572 A.2d 900, 904 (1990), given defendant’s extensive testimony which dovetailed the profile of a battered woman, compare L. *205Walker, supra, at 27, with transcript at 115 and 143; compare L. Walker, supra, at 48-49, with transcript at 111; compare L. Walker, supra, at 50-51, with transcript at 118-19; and compare L. Walker, supra, at 137, with transcript at 105-06.
This case illustrates the dangers of dictating extemporaneous findings on the record. When this is done, there is no opportunity to review and edit the decision. Read in hindsight, oral findings may appear misspoken, ill advised, or contrary to the true intent of their message.
We doubt that the record of the court’s findings and conclusions would appear so confusing, enigmatic, and incomplete if it had been drafted, reviewed, and edited. Time constraints and work pressures in our trial courts are great and make certain short cuts both tempting and useful. However, a court should exercise caution before making extemporaneous findings. See Bucholt v. Bucholt, 152 Vt. 238, 241 n.4, 566 A.2d 409, 411 n.4 (1989).
We reverse the property distribution and remand for a new trial on that issue.