Petition of RMG

*796 NEWMAN, Chief Judge,

dissenting:

In this proceeding, the trial court considered the competing adoption petitions of a black child’s natural grandparents and her white foster parents. The trial court found that, where the totality of other factors were in equipoise, 1 the interests of the child, D.G., would more likely be better served by an intraracial adoption than by an interracial one. This factual conclusion was more than adequately supported by the unusually thorough testimony addressed to the subject. The ensuing adoption order must therefore be sustained unless the Constitution prohibits taking note of the races of the parties involved even for the limited and important purpose addressed by the trial court: to weigh the possibility that racial differences between adoptive parents and child may adversely affect 2 the latter’s future development and welfare. I think it is clear that the Constitution does not require a court to ignore racial differences between prospective parent and child, to the extent that such differences must be observed in order to judge that best interest.

However, the majority has managed to overturn the trial court’s judgment by taking what I think is an overly stilted view of the relevance of race to adoption, disregarding the weighing process implicit in the court’s expressed reasoning, and thereby concluding that its consideration of race may have been “insubstantial.” The prescribed remedy — remanding for further proceedings — will probably do no harm. It is clear from the opinion and trial record, however, that there was no improper consideration of race to be remedied, and it is unlikely that a wordier opinion will make it easier to detect and overturn abuses if and when they occur in the future.

I. THE NATURE OF THE TRIAL COURT’S DECISION

In the trial court, both sets of prospective adoptive parents had an opportunity to present any and all available evidence bearing on how the child’s best interest would be affected by either of the alternative placements, including potential effects of racial differences, and any factors mitigating those risks. After a thorough hearing, the trial court was able to judge such factors as the families’ respective financial resources, blood relationships with the child, the effect of a shift in custody away from the family with temporary custody, family stability, and the sort of love and care that each family could be expected to provide. 3 *797 The court was in an unusually good position to judge with respect to the H. family, for they had already had custody for some time on a temporary basis. Moreover, they have another adopted child, who, unlike the parents, is black. The court thus had before it evidence of affirmative steps taken by the parents to mitigate some of the potential problems arising in an interracial adoption.

Q. Have you made particular efforts to cultivate his black heritage, and maintain the presence of that in your household?

A. We definitely began an affirmative program when Jeff came to live with us.

Q. How long has Jeff been with you?

A. Jeff came to live with us when he was eight days old, so he has been with us almost six years. He’ll be six in August.

Needless to say, being in the military, we live in a thoroughly integrated society, and much more accepting than you would find in most civilian communities. Through guidance of friends and associates of my husband, the chaplain at the chapel where we go to church — one of our chaplains is black — I have gotten in magazines, and I have some pre-school black history books for my son, a black coloring book for children, put out by Ebony Jr. I make sure he knows that he’s not white. I don’t care how long he lives with us, he’s black, and he’s beautiful, and he’s ours.

Having weighed all this evidence, the court found as a fact that considerations other than the difference in race between the child and one set of adoptive parents were in equipoise. 4

The court then turned to a consideration of the possible effects of racial differences between parents and child, and concluded that the possibility of adverse effects in the H. family tipped the balance in favor of the G. family. This conclusion was based on considerable expert testimony, which the majority opinion recites in part. It is evident that neither the parties nor the court approached this issue lightly. Indeed, when a social worker indicated that her testimony was largely based on the views of Dr. Wels-ing, the agency psychiatrist, the court requested that Dr. Welsing testify in person. After weighing facts and expert opinion on both sides, the court observed that conclusive data on the efficacy of interracial adoption is lacking. 5

*798 While there is a debate among social scientists about the viability of interracial adoption, 6 no one — including the parties herein and their expert witnesses — contends that such adoptions tend to be superior to intraracial adoptions, all other factors equal. Rather, the dispute concerns the degree to which the fact of a racial differ *799 ence presents risks which are not otherwise present, and the extent to which those risks may be reduced by affirmative steps on the part of the parents. 7 As two researchers sympathetic to interracial adoption have said,

it was and still is accepted that the inra-cial adoption, when feasible, is preferable to the transracial adoption. The basic question about transracial adoption was whether it worked at all, and to what extent. [Grow & Shapiro, Not So Bold, supra note 5, at 88.]

Even the most outspoken advocates of interracial adoption, who see it as a means of addressing the fact that the need for placement of homeless minority children has historically outstripped the number of minority adoptive parents, do not contend that there is no risk of problems in social or psychological adjustment. Rather, they contend that the hazards can be reduced to a manageable level, and that racial differences should therefore not stand as an absolute barrier to adoption. 8 However, most authorities agree that intraracial adoption is preferable. 9

The hazards of interracial adoption should not be exaggerated, but neither should they be ignored. An inevitably imprecise prediction about the effects of an interracial placement must be made in the context of all relevant circumstances, including any mitigating efforts by the parents. But when all other factors are in equipoise, the possibility of an adverse effect, no matter how small or how unlikely, would suffice to permit the trial judge to tip the balance in the direction of the in-traracial alternative.

The trial court found as facts that “[a]t a later stage, notwithstanding love and affection, severe questions of identity arising from the adoption and race most probably would evolve. In the world at large, as the circle of contacts and routines widens, there are countless adjustments which must be made.” It concluded that the child’s interests would probably be better served in the G. family. Given the equal balance of other factors, this conclusion is adequately supported if it is permissible to conclude that there is any potential adverse effect from the interracial alternative. That is indisputably the case here. Accordingly, the result below can be upset only if it is constitutionally impermissible to give any weight whatsoever to adverse effects on the child related to racial differences between herself and her parents.

II. THE CONSTITUTIONALITY OF THE TRIAL COURT’S DECISION

A. The Level of Scrutiny

A majority of the United States. Supreme Court has not agreed on the standard of scrutiny to be applied to “benign” as op *800 posed to “invidious” racial classifications. 10 However, at least four justices are of the view that racial classifications that attempt to remedy the effects of past discrimination and do not stigmatize a particular racial group are subject to less exacting scrutiny than other racial classifications. Thus, a plurality of the Supreme Court subjected a medical school program designed to assure the admission of a specified number of minority students, to an intermediate level of scrutiny, instead of the strict scrutiny test. Regents of the University of California v. Bakke, 438 U.S. 265, 359-61 , 98 S.Ct. 2733, 2783-2784 , 57 L.Ed.2d 750 (1978) (Brennan, J., White, J., Marshall, J., and Blackmun, J., concurring in the judgment in part and dissenting in part). See also Fullilove v. Klutznick, 448 U.S. 448, 517-19 , 100 S.Ct. 2758, 2794-2795 , 65 L.Ed.2d 902 (1980) (Marshall, J., with whom Brennan, J. and Blackmun, J., join, concurring in the judgment). While these cases involved affirmative action, the authors of the cited opinions neither stated nor implied that only affirmative action could qualify as a benign use of a racial classification.

Employing the same analysis, the use of race in this case does not require application of the strict scrutiny standard. As in Bakke and Fullilove , the normal indicia of racial benignancy are present. First, the use of the racial factor does not stigmatize a particular racial group. In other words, the use of race in this adoption proceeding was not based on the presumption that one race is inferior to another, nor does it place the weight of the court behind racial bigotry and separatism. The trial judge used race as one of many factors to be weighed in calculating the child’s best interests, not as a means of insuring racial purity or separation.

Secondly, though the consideration of race was not purposefully remedial, neither were its purpose and effect pernicious with respect to the distribution of burdens and benefits among racial groups. Indeed, the use made of race in this case should be even less open to objection than that in Bakke and Fullilove . There, the racial classification provided benefits to an identifiable class, to the detriment of another class, for the purpose of remedying the present effects of past discrimination. Here, the consideration of race seeks not to improve the position of any particular racial group, but simply to protect the best interests of the child, of whatever race.

However, the majority quickly dismisses the possibility that the consideration of racial differences made in this case could qualify for intermediate scrutiny as a benign use of race. It does so on the ground that, in the adoption context, “racial classifications over the years have resulted in particularly vivid examples of invidious discrimination.” Ante at 786. But this distinction is clearly inconsistent with the position taken by the Supreme Court plurality favoring intermediate scrutiny. They proposed intermediate scrutiny in a case involving alleged race discrimination in admission to a public institution of higher education. No one who recalls the last twenty-five years of the history of this country could possibly believe that university admissions is not an area “where racial classifications over the years have resulted in particularly vivid examples of invidious discrimination.” The fact that invidious uses of race have occurred, and continue to arise, is not a persuasive reason for rejecting intermediate scrutiny in a clearly benign case. If the possibility of invidious discrimination in the same context is all that is necessary to bar intermediate scrutiny-regardless of the fact that the actual consideration of race was clearly benign— then such scrutiny is never possible. That approach would place too extreme an impediment to benign and important uses of *801 color consciousness. Unless and until the promised land of racial equality is achieved, it is unrealistic to blind ourselves to color in such instances.

Moreover, the consideration of race in this case differs from the prejudiced and separationist approaches taken in cases referred to in the majority’s “historical perspective” in at least two fundamental respects. Since race is only one factor among many, many interracial adoptions will be permitted. This is not a classification entailing the separation of races, which would be “invalid without more.” Regents of the University of California v. Bakke, supra, 438 U.S. at 358 , 98 S.Ct. at 2782 (Brennan, J., concurring). Second, any tendency to reduce the number of interracial families is purely incidental to a non-racial, non-sepa-rationist purpose: to promote the best interest of each adopted child.

The difference between cases in which race-consciousness required the application of strict scrutiny and the present case is well illustrated by Loving v. Virginia, 388 U.S. 1 , 87 S.Ct. 1817 , 18 L.Ed.2d 1010 (1967). In Loving , the Supreme Court struck down Virginia’s anti-miscegenation statute, which made it a crime for a white and non-white person to marry. The state contended that the statute should be regarded as non-invidious because equal penalties were placed on any person engaging in an interracial marriage, rather than only the non-white spouse. The Court looked beyond this facade of neutrality and recognized that the sole purpose and function of the statute was to separate the races. Since the state had no non-racist and compelling purpose in preventing, consenting adults of different races from marrying, the statute violated equal protection.

Here, in contrast, the purpose and predominant effect of the court’s action is legitimate and compelling: protection of the child’s best interests. It is also significant that only one side of the proposed interracial adoption consists of consenting parties: the H. family. The child is not of consenting age, and therefore it is the court which is in the difficult position of standing in her shoes and judging what is in her interest. If the child had capacity to decide for herself, she would certainly not be prevented from taking note of the different race of those proposing to adopt her. When the court does the same on her behalf, there is no indication of invidiousness that necessitates the strictest level of scrutiny.

This situation should also be distinguished from one in which a standard that makes no ostensibly improper use of race is being subverted in a race-conscious way by those who apply it. Yick Wo v. Hopkins, 118 U.S. 356 , 6 S.Ct. 1064 , 30 L.Ed. 220 (1886), for example, provides an apt illustration. That case involved applications for licenses to operate commercial laundries. While the official criteria for the granting of license applications made no reference to race, in practice, the authorities denied licenses to Orientals while granting them to white owners of similar facilities. If it appeared in an adoption case that the trial court was manipulating non-racial criteria, or benign and appropriate race-conscious factors, so as to achieve an invidious goal, strict scrutiny should be applied and the ultimate result would no doubt be reversal.

I conclude that, under the view espoused by at least four Supreme Court justices, the consideration of race in this case is “benign” rather than invidious. Therefore, it may be justified by showing that it serves “important,” rather than “compelling” governmental objectives, and is “substantially related to,” rather than “necessary to,” the achievement of those objectives.

B. Constitutional Scrutiny of the Court’s Decision

The selection of the appropriate level of scrutiny is not determinative in this case, since the consideration of race is necessary to the achievement of a compelling governmental interest and thus comports with equal protection even under the strict scrutiny standard. See, e.g., In re Griffiths, 413 U.S. 717, 721-22 , 93 S.Ct. 2851, 2854-2855 , 37 L.Ed.2d 910 (1973); Graham v. Richard *802 son, 403 U.S. 365, 375 , 91 S.Ct. 1848, 1853 , 29 L.Ed.2d 534 (1971); McLaughlin v. Florida, 379 U.S. 184, 196 , 85 S.Ct. 283, 290 , 13 L.Ed.2d 222 (1964). The interest at stake — protection of the prospective adop-tee’s best interest — is indeed compelling, a fact which the majority does not dispute. Ante at 786. It is, in fact, the only state interest that can justify the granting of an adoption petition.

A degree of race-consciousness is permissible to the achievement of that interest, because certain potential future hardships to the child arise when the parents are of a different race. The trial court’s finding that such risks exist is more than amply supported by trial testimony. In turn, that conclusion, and the testimony addressed to it, find virtually unanimous support in the research and social science literature on the subject.

Some of the risks of interracial adoption involve the child’s development of identity (including racial and cultural identity), self-esteem, and a sense of belonging in the family — relevant considerations which the majority recognizes as legitimate and important. One problem is the possibility that the child may not perceive herself as black or develop an identity as a black person. 11 There was evidence at trial that the foster parents would make efforts to alleviate this possibility. Even if the child is made av/are of her black identity, however, other problems might develop. The child would then have to cope with the fact that she is different from her parents. As one of the expert witnesses in this case testified:

If a child has to start out, “I’m different. I’m special,” that child is being deprived of a strong and healthy concept of self that they could have in an environment where they do not have to deal with that kind of division twenty-four hours a day.

Another aspect of the identity problem is the possibility that the child may experience a “conflict of loyalties” as she grows older. If she identifies with the culture of her heritage, she may feel isolated from her family. If she identifies with her family, her skin color will always be there to remind her and others of her origins. 12 In other words, the child may be caught between two cultures and accepted by neither. 13 Even if the white foster parents were able to nurture the child’s ability to perceive herself as black, there is little they could do to prevent the ambivalent feelings and rejection she may experience later in life.

In addition to a strong sense of identity, the black child must learn to develop certain survival skills. Regardless of how she is identified by herself or her family, she will be identified as a black person by society and will inevitably experience racism. Blacks and other minorities develop survival skills for coping with such racism, which they can pass to their children expressly, or more importantly, by unconscious example. E.g., Chestang, supra note 6, at 101-03. Parents of interracial families may attempt to learn these lessons and then teach them, but most authorities recognize that this is an inferior substitute for learning directly from minority role models. J. Ladner, supra note 5, at 255; Chestang, supra note 6, at 102-04. Few white parents even claim they can teach such skills. J. Ladner, supra note 5, at 115. In one study, a third of interracial parents did not undertake affirmative efforts to teach them. The two thirds who did, did so predominately through secondary materials like books, *803 rather than by example, which is the normal method of socialization. Simon & Altstein II, supra note 5, at 18. See also J. Ladner, supra note 5, at 115-16.

The racism experienced by blacks in this society may be encountered even more often by blacks in interracial families. Those with racist attitudes are undoubtedly opposed to interracial families. Acts of bigotry may be visited upon a child of a different race than his adoptive parents, regardless of the child’s own race. Jones, supra note 5, at 163. Other children in the family may also become the targets of slurs. Chestang, supra note 6, at 103. Certainly the possibility of such traumatic experiences, with their attendant psychological effects, bears on the child’s best interest. However, it cannot be emphasized too strongly that the desires of those opposing racial mixing are not to be weighed, in the balance. The existence of such desires is significant only in that they may lead to acts directed against the child that adversely affect his or her interest. The majority fails to attach any significance to this distinction, stating flatly that “a well-intentioned effort to protect a child against community prejudice is [not] a proper justification, in itself, for an adoption decision.” Ante at 789-790. That is so only if the “effort” consists of barring interracial adoption. But there is no reason for failing to weigh the effect of prejudice on the child’s best interest as one factor among a multitude.

Since black children in interracial families may be even more exposed to racist attitudes than other blacks, their need for survival skills is more acute. White parents, however, tend to be less equipped to pass on those skills. See discussion at 53 supra. It is ironic and unfortunate that those black children most in need of survival skills are in environments which are the least able to provide them.

Racial slurs are not the only kinds of public reaction that pose problems for the child and his family. The range of responses is wide, and serves as a recurrent reminder that the child is “different” from his family. See B. Jackson, supra note 5, at 13, 16-17; J. Ladner, supra note 5, at 209-15; C. Zastrow, supra note 5, at 39-43. When some people see a child whose race is different from tht of his parents, they assume he is an illegitimate child or the product of a multi-racial marriage — circumstances they may disapprove of. Other people overreact in a well-meaning way, commenting on how wonderful it is to adopt a minority child. But however well-intentioned, such reactions have the effect of emphasizing to the child that he is “different”, and can lead to a sense of isolation. B. Jackson, supra note 5, at 13-14. Thus, while all adopted children have to cope with the fact that they are adopted, the interracial adoptee may have an even more difficult experience since his status is evident to the world at large.

The child may at times feel that no-one understands his confused feelings about his natural parents or his pain at being called names.... Of course, every child at times suffers growing pains and feels no-one understands. But the black adopted child is visibly different from those he loves and this can give him a more complex situation to cope with. [Id at 14.]

In sum, consciousness of a racial difference between the adoptive parents and the child is necessitated by society’s compelling interest in fostering the child’s best interest. The trial court took note of race only so far as required to ascertain the child’s interest, explaining its concern both about problems related to identity, and to relations with the world outside the family. Its fact findings that such problems exist and that the child’s best interest would therefore — all else being equal — be furthered in the G. home, are well supported by evidence of record.

III. FLAWS IN THE MAJORITY’S APPROACH

A. An Overly Narrow View of the Relevance of Racial Differences to the Child’s Best Interest

As already set forth, there are a number of respects in which racial differences be *804 tween parents and child may affect the latter’s interest. However, the majority treats all factors it considers relevant as subsumed by the concept of identity. Moreover, there is reason to wonder how much the majority’s race factor actually has to do with race. In enumerating the questions that should be explored, the focus is on the attitudes and behavior of the parents, and on the environment in which they live (e.g., whether the local schools are segregated). These were addressed at trial, and no one doubts that they are relevant. But the question that social scientists, the expert witnesses, and ultimately the courts have to grapple with is whether, notwithstanding the very best of intentions and efforts on the part of the parents, there are potential problems in the interracial context of a kind not expected in the intraracial case. The court should be permitted to conclude, as it did in this case, that such risks persist to some degree even when parental attitudes and the external environment are favorable. In other words, the issue is not exhausted simply by consideration of attitudes and environment; a difference in race between parents and child is of independent significance.

Yet the majority comes close to saying that differences in the relative viability of interracial and uniracial families are only coincidentally or indirectly related to racial differences, due to a possible statistical correlation between the race and favorable attitudes or external environment. See ante at 792, 793. However, there is no basis for excluding from consideration the more direct influence of race. While the majority de-emphasizes that influence, I do not believe its opinion can be read as eliminating it from evaluation.

If instead the majority opinion were read as reducing the race factor to attitudes about race, and environmental conditions bearing on race relations, paradoxical results would follow. The only reason this court is faced with a constitutional issue is that the trial court directly took the race of the parties into account. That is why there is a racial classification calling for more than minimal scrutiny under the Equal Protection Clause. If, for example, all that the court had addressed were attitudes about race, without direct reference to the race of the parties in relation to the race of the child, there would be no ground for the majority to engage in strict scrutiny. In finding strict scrutiny satisfied, the court’s opinion cannot be read as limiting the process to factors other than, albeit related to, race itself.

A second paradoxical result of an overly narrow view of the race factor is that it would require attention to the effect of race on identity even when there is no petition by prospective parents of a race different than the child’s. If attitudes and the extra-familial environment were to dominate the analysis, the racial factor could not be overlooked in any adoption. See Johnson v. United States, D.C.App., 398 A.2d 354, 365 (1979) (In ascertaining whether an abuse of discretion has been committed, one criterion is whether all relevant factors have been considered). Yet neither prospective parents nor the court raise race as an issue when all involved are of the same race. The reason is simple: common sense and experience lead us not to expect race-related identity problems in the single-'race context, and the expert evidence supports that view. Accordingly, when competing interracial and intraracial petitions are presented, it is hardly surprising if the focus of the court’s discussion regarding race is on the interracial family. But for that family’s involvement, there would be no issue. In this context, the absence of specific mention of the intraracial family in the trial court’s discussion of race obviously reflects a finding that the potential problems identified by the court do not pertain to any significant degree to that family.

In this ease, there was evidence regarding the love and concern each set of parents would have for the child. The H. family emphasized their positive attitudes toward the child’s race and culture, which had already been exhibited in the raising of an *805 other black child in their family. 14 Both families had the opportunity to present any and all evidence regarding their respective abilities to raise the child, including attitudes related to race. But having considered all non-race facts, including the treatment that each family would give the child, the court concluded that the petitions had equal merit. 15 Thus the trial court concluded that the race of the parties tipped the balance, however slightly.

If the race factor really has little to do with the race of the parties, as the majority appears to believe, then it would not be unusual to find that a child’s identity, including racial identity, would be more secure in an interracial placement than ah intraracial one. The majority thus proposes what appears, at least superficially, to be a fair and neutral approach, in which neither interracial nor intraracial adoption is preferred. The majority says “the court cannot properly weight [the race factor], either automatically or presumptively — i.e., without regard to evidence — for or against cross-racial adoption.” Ante at 787. To do otherwise would “add a racially discriminatory policy” and give the intraracial adoption a “head start.” Id. If such language simply means that the courtijs not to rule out interracial adoption, inject a personal disapproval of interracial adoption, or give racial differences an undue weight as compared with other factors, it is unobjectionable. But the prescription is clearly incorrect if it means that the court cannot give weight to evidence that, other things equal, intraracial adoption tends to further a child’s interest morey than interracial adoption. To put such relevant facts into the balance is not evidence of bias. To the contrary, their exclusion would give a “head start” to the interracial alternative. 16

Thus the more evenhanded approach is to take note of the fact-that there are some respects in which interracial adoption presents potential problems that are not significant in the intraracial context. See section II B supra. To reiterate just one example, in the uniracial family, race does not serve as a constant reminder to the world at large and the child that he is adopted. Thus while the positive attitudes and environment of the interracial family are relevant, by way of mitigating some of the problems, it would be difficult to regard them as eliminating all risks entirely, much less making the interracial alternative superior as to this factor. This approach is supported by the record and was evidently adopted by the trial court, which said,

[N Notwithstanding love and affection, severe questions of identity arising from the adoption and race most probably would evolve. In the world at large ... there are countless adjustments which must be made.

As discussed above, the social researchers— even those who advocate interracial alternatives — generally accept the premise that where an intraracial placement is available and other factors are equal, it is preferable. Not even appellee’s expert witness contended that interracial placements in general, or the H. family in particular, would be superi- or as far as race-related considerations are concerned.

If an approach taking account of potential problems associated with interracial adoption amounts to a “preferénce” for in-traracial adoption, it is an entirely permissible one. Such a preference is no less legitimate than a preference for a ytwo parent family or for parents with adequate resources to raise the child. If, in the guise of neutrality, the majority’s framework rules out of order a preference for intrara- *806 cial adoption that is supported by evidence, it introduces a bias against adoption by the black intraracial family. It is just as prejudicial to eliminate a legitimate consideration as it is to introduce an illegitimate one.

B. The “Historical Perspective”

The majority gets off the track in suggesting that the possibility of a prejudiced or bad faith use of race bears on either the result or the remedy in this case. The majority reviews statutory and case law that today is rightly recognized as founded on unjust racial attitudes or inaccurate assumptions. Ante at 788. Indeed, historically the law has reflected the racial attitudes of those in power. See generally A. Higginbotham, In the MatteR of Color (1978). For example, there is one court’s conclusion that mixed race children

... will have a much better opportunity to take their rightful place in society if they are brought up among their own people. [Ward v. Ward, 36 Wash.2d 143, 145 , 216 P.2d 755, 756 (1950).]

Needless to say, I concur in the proposition that appellate courts must be vigilant in correcting such instances whenever they occur. But there is not the slightest indication that improper attitudes or preconceived factual assumptions influenced the decision below; indeed, every indication is squarely to the contrary. The fact that discriminatory uses of race have occurred is not a reason to eliminate entirely legitimate uses. As explained in section II A, supra, this history does not justify a refusal to apply intermediate scrutiny to a benign consideration of race. Neither should it bear on the application of the standard that is chosen.

C. Substantiality of the Trial Court’s Consideration of Racial Differences Between Parents and Child

Although the majority does not acknowledge some of the reasons that it is appropriate to take note of the race of the petitioners, it nonetheless decides that compelling interests justify attention to race. Accordingly, the statutory authorization of the consideration of race is upheld. In examining the trial court’s exercise of discretion, the majority finds that no proper factor was ignored, and no improper factor was considered. How then does it arrive at the conclusion that the trial court’s consideration of the “racial factor” requires reversal? It does so by indicating that the court’s reasoning may not have been “substantial,” and “based on a firm factual foundation.”

In my view, the suggestion that the court’s reasoning regarding the use of face was anything but substantial is implausible, and clearly belied by the record. Although the parties presented considerable lay and expert testimony addressing the merits of an interracial adoption at their own initiative, the trial court requested and received additional expert testimony. Rather than granting an immediate ruling, the court reserved decision until it had had a further opportunity to weigh the issues. Its written order devoted more discussion to race than to any other factor. It took great care to point out that it considered the possible effects of an interracial adoption to be only one factor among many, not the basis of a per se rule. It identified the concerns that led to the conclusion that adoption by the G. family would be preferable, i.e., that problems of psychological and social adjustment might arise in the adolescent years if the interracial alternative were adopted. This explanation clearly indicated that expert testimony to that effect was credited. In short, every indication is that the trial court’s reasoning, particularly as regards the race issue, was thoughtful, restrained, and serious: in a word, substantial. The “firm factual foundation” for the court’s reasoning and conclusions is clearly present in extensive record testimony. That the court’s reasoning was in fact based on that record is amply indicated by the court’s reference to the risks of interracial adoption that come to fruition in adolescence, and to relations with the world outside the family. There is no basis for supposing otherwise.

D.Seeing Less in the Trial Court’s Reasoning Than Was Actually There

As the majority apparently recognizes, its conclusion can only be reached, if at all, by *807 departing from the normal procedure for the review of discretionary trial court rulings. In Johnson v. United States, supra, we drew a fundamental distinction between the degree of explanation ordinarily required of a trial court and that expected of administrative agencies for review of their respective exercises of discretion. As aptly explained in Washington Public Interest Organization v. Public Service Commission, D.C.App., 393 A.2d 71 (1978), cert. denied sub nom., Potomac Electric Power Co. v. Public Service Commission, 444 U.S. 926 , 100 S.Ct. 265 , 62 L.Ed.2d 182 (1979), review of agency action under the District of Columbia Administrative Procedure Act and similar statutes providing for review of agency action requires a rather detailed statement of findings and reasoning. For a number of reasons, several of which are discussed in that case, it is both feasible and necessary to require extensive written explanations from agencies. Agencies exist in large part because they can develop the sort of specialized expertise in a narrow policy area that the courts and the legislative branch are unable to do. The records in such cases are also often voluminous and highly technical. Without rather detailed explanation, a reviewing court may be at a loss to understand the decision and the significance of the data in the record.

As we explained in Johnson , it is usually neither necessary nor feasible to require such detail in the case of review of judicial decisions. By definition, such matters are within the judicial sphere, rather than confined to specialized expertise. Indeed, the reviewing court is the source of the authoritative rulings that guide the trial court’s discretion. Trial judges, unlike agencies, deliver final judgments and numerous interlocutory rulings on a daily basis. We thus concluded in Johnson that a trial court need not make a lengthy written exposition of its reasoning. Accordingly, if the trial court opinion alone will not suffice for purposes of judicial review, this court can search the record and infer the court’s reasoning based on the applicable legal standards and the results reached. Id.

But the majority treats the trial court’s opinion as if it constituted the court’s reasoning process rather than merely post hoc evidence of it. Inexplicably, the majority cites Johnson as support for the proposition that the court’s attention to the race issue must be presumed to be limited to the words contained in its order. It relies on a quoted observation that it is impractical to expect a detailed explanation of decisions rendered during trial. But this fact is explicitly cited as an example of why one may flesh out a court’s reasoning with an eye to the record. Johnson neither stated nor implied that looking to the record to fill the interstices in the court’s expressed reasoning is precluded except in the case of decisions in the midst of trial.

What the majority seems to suggest is that trial court decisions are presumptively to be treated in the same manner as administrative decisions: unless there are exigent circumstances making it impractical for the trial court to deliver a lengthier opinion, a reviewing court must read the court’s words in pristine isolation from the record on which they are based. According to the majority “Because the trial court’s decision here was rendered comfortably after trial ... the general justification for requiring us to ‘examine the record and infer the reasoning’ the trial court used ... is not present here.” Ante at 791. To my mind, the fact that the decision in this case was deferred until after a period of reflection is not dispositive of whether we may look to the record. The trial court could have rendered its decision orally at the conclusion of the evidence. See Super.Ct.Civ.R. 52(a). The paradoxical result of the majority’s approach would be to make reversal of a decision — as embodied in the court’s opinion — more likely if it is made in written form after a period of reflection than when it is rendered on the spot.

The majority also states that the Constitution prevents this court from looking to the trial record as an aid in understanding the court’s reasoning. To my knowledge *808 there is no judicial authority whatsoever for this proposition. The cited cases, Personnel Administrator of Massachusetts v. Feeney, 442 U.S. 256 , 99 S.Ct. 2282 , 60 L.Ed.2d 870 (1979), and Compos v. McKeithen, 341 F.Supp. 264 (E.D.La.1972) (three judge court), at most support the proposition that the party seeking to uphold the constitutionality of state action bears the burden during subsequent judicial review of that action of demonstrating that the action satisfies the appropriate level of constitutional scrutiny. That is not in dispute. Nowhere do the cited passages state or imply that a reviewing court is ever constitutionally precluded from examining a trial court’s reasoning in light of the trial record.

It is important to bear in mind that this is not a case in which the trial court simply announced the result of its deliberation. An economical but clear explanation was given. If, as the majority contends, resort to the record is necessary in order to confirm that the court rendered a proper decision, it is only by way of fleshing out the explanation that was given: (viz., that certain problems which would not be expected in the intraracial context might arise during adolescence in the interracial adoption under consideration).

I agree with the majority that there is a limit on how far we should go in inferring reasoning from the record. But to push this constraint as far as is necessary to overturn the trial court’s decision in this case strikes me as absurd. It requires us to tear the court’s words from the evidentiary context from which they grow. We should not be surprised that, by uprooting the court’s decision from the ground which gives it life, we prevent it from sustaining itself.

E. Implementation of the Majority’s Three Step Procedure

However, even accepting the majority’s approach to judicial review and its three step procedure for considering race in the adoption context, it is difficult to see why reversal is thought to be required in this case. The majority concedes that the first step — evaluating “how race is likely to affect the child’s development of a sense of identity” — was satisfied. Because of the unique posture of this case, the third step— judging “how significant the differences between the families [with respect to racial identity] are when all the factors relevant to adoption are considered together” — also cannot serve as a basis for reversal. The trial court found that the race factor favored the G. family, and that all other factors were in equipoise. Therefore in this case the magnitude of the differential as to the race factor is of absolutely no consequence; any weight, no matter how slight, could suffice to tip the balance.

Thus, according to the majority’s framework, the need for reversal should depend on whether the second step is satisfied. This step requires that the court compare how race will affect identity in each family. I am at a loss to understand how the majority can suggest that the court’s conclusions may not have been based on a comparison of the parties with respect to this factor, and that its reasoning may not have been “substantial.” In searching for a particular tree — an explicit statement that the potential problems which could arise in the H. family were not expected in the uniracial G. family — the majority seems to have lost sight of the fact that it is standing in the midst of a forest. The entire enterprise of judging the competing adoption petitions is a matter of comparing how each family rates with respect to each factor and then determining which placement is preferable overall. I would have thought — and I daresay the trial judge would have as well— that this was so elementary and taken as given by all involved that it was unnecessary for the court to state the proposition in its opinion. But if an explicit statement was required, it was supplied by the trial court:

The Court does not conclude [an interracial family such as the H.’s] could not sustain itself. Rather the question is, is there not a better alternative?

*809 The court then concluded that adoption by the G. family was a “better alternative.” It is difficult to imagine words more clearly conveying a comparison of the two families as to the racial factor.

A line by line reading of the opinion reveals that the comparison of the families with respect to race was both substantial and based on the record. The court began its discussion by referring to the expert testimony and joining in the experts’ dismay that there is not more conclusive research on the relation of race to adoption. It is evident that the discussion followed was founded on such predictions as the experts could give based on what is known about interracial adoption and on their knowledge of the particular parties involved. The court noted that “no conclusive absolutes are to be drawn on the basis of race. It would seem, however, entirely reasonable that as the child grows older the ramifications of this problem would increase.” It is clear that “the ramifications of this problem” refers to social and psychological risks associated with interracial adoption, above and beyond those that inhere in adoptions in general. The succeeding sentence explicitly indicates that the court took into account the positive attitudes and efforts which would tend to mitigate those problems: “At a later stage, notwithstanding love and affection, severe questions of identity arising from the adoption and race most probably would evolve.” Having stressed the identity issue, the court then referred to relations with people outside the family, saying, “In the world at large, as the circle of contacts and routines widens, there are countless adjustments which must be made.” Both types of considerations were clearly founded upon more than ample record evidence. I find implausible the majority’s assertion that, while ample record support is present, the trial court may not have intended to rest its reasoning on that evidence. It was not necessary for the court to recite the testimony, at the risk of having it disregarded by this court.

The majority, however, finds the lack of “specific findings ... as to how race would be likely to affect this particular black child” in the H. & G. families to be fatal. 17 Ante at 793. However, the effort to distinguish a “generalized” judgment from a “particularized” inquiry concerning the individual family before it is misleading. This approach seems quite plausible on the surface, but the supposed distinction blurs considerably on closer examination. Of course, the statement that “this family is of a different race than D.G.” would be true of a great many potential adoptive families. But the same is true of virtually any other relevant fact including those emphasized by the majority (e.g., living in an integrated neighborhood, attending an integrated school). That such factors may have positive effects is no less a “generalization” than the type of inference that the majority criticizes. Indeed, where the goal is prediction of future conditions such as how a child’s interests will be served by a given adoption, no piece of evidence has any probative value unless it is possible to relate it to the past experience of other adoptive families. Only then can one infer (generalize) therefrom that the experience of this family is likely to be similar. Thus there is nothing suspect about reliance on testimony based on “generalization” from the experience of other families who are similarly situated in relevant ways, including racial differences between parent and child.

The fact that the adoption decision is inevitably based on prediction and generalization is important in another way: it places limits on the extent to which specific findings can be meaningful. Of course, one could recite each piece of evidence individually, but it would be an exercise in futility to attempt to attach a discrete value to each bit thereof. .Rather, what a court must do in such situations is to look at all the evidence and judge the interrelated whole, arriving at the broad conclusion that, on balance, one alternative seems pre *810 ferable to the others. The majority’s discussion of the factors to be considered when interracial adoption is proposed will no doubt provide useful guidance to trial judges. But I question whether finding an opinion such as that of the trial judge in this case to be insufficiently detailed will have a further and independent effect in improving the quality of adoption decision-making. As a practical matter, the most we can do is to provide guidelines for the court’s discretion, and then place our reliance on the trial court to arrive at fair judgments. In this case, the court devoted more space to the discussion of the race issue than to any other factor. Of course, it could have said more. But what it did say succeeded in demonstrating that race was considered only in a limited and entirely proper way. There was no indication that countervailing evidence favoring the losing party was excluded from consideration.

CONCLUSION

The Constitution does not require a court to blind itself to realities affecting a child’s best interest, even when those facts depend on the respective races of the parents and child. What is in the best interest of a child is a factual question to be resolved by the trial court. We cannot require that courts, even indirectly, find facts contrary to their view of the evidence. That a majority of a panel of this court might have drawn different conclusions from the evidence concerning interracial adoption in general and the G. and H. families in particular is not a basis for reversal. While ostensibly accepting these propositions, the majority’s opinion may be read as circumscribing the consideration to be given to race, to an extent which I think is unjustifiable. Probably more significantly, the majority has narrowed its examination of the court’s reasoning so as to see less than is actually there. The result is to lead the majority astray from the relatively straightforward issue posed by this case: whether it is constitutionally permissible and appropriate for a court to give any weight to evidence that an interracial adoption presents certain risks to a child’s welfare, which may be mitigated though not eliminated by special efforts on the part of the parents. I think that this factor may and should be included among the many others relevant to the child’s best interest. It follows that, in a case such as this, where other factors are in equipoise, the interracial factor may sway the result. Since race-consciousness in this limited context is permissible, and the court’s judgment as to the child’s best interest is supported by the evidence and not clearly erroneous, the result should be affirmed.

We must live in the world as it is while we strive to make it as it should be.

.The majority has stated that the trial court’s examination of all factors, other than race, “slightly favored” the white foster family and that the race factor “tipped the decision in favor of adoption by D.’s black grandparents.” Ante at 793 & n. 40. I disagree with that characterization. The trial court found, as is shown by these portions of its opinion quoted by the majority, ante at 781-782, that all relevant factors, other than race, were in equipoise. Even if the majority’s reading of the record were correct, I would reach the same analytical results, and I do here.

. Advantages that interracial adoptions might conceivably have over intraracial adoptions in promoting the child’s best interest, if raised by the parties, would also be relevant. However, it is not contended that interracial adoptions are, other things equal, superior to intraracial ones. See infra at 798.

. Race was considered by the trial court as merely one of numerous relevant factors and not as a determinative or conclusive factor. There is not the least intimation in the trial judge’s opinion that an otherwise fit family of a different race can never be granted adoption as against a family of the same race, regardless of its fitness. Thus, the court’s decision is not inconsistent with cases prohibiting state action which gives race automatically controlling dimension in custody or adoption decisions. See *797 In re Adoption of a Minor, 97 U.S.App.D.C. 99 , 228 F.2d 446 (1955); Compos v. McKeithin, 341 F.Supp. 264 (E.D.La.1972); In re Gomez, 424 S.W.2d 656 (Tex.Civ.App. 1967). While some courts have held that race can never be a relevant factor in a custody or adoption proceeding, Be azley v. Davis, 92 Nev. 81, 83 , 545 P.2d 206, 208 (1976); DeLander v. DeLander, 37 U.S.L.W. 2139 (Cal.Super.Ct., Aug. 14, 1968), the better view, and the view adopted by this court in this case, is that race may be considered a relevant factor in such proceedings. See Drummond v. Fulton County Dept. of Family and Children’s Services, 563 F.2d 1200, 1205-06 (5th Cir.1977) (en banc), cert. denied, 437 U.S. 910 , 98 S.Ct. 3103 , 57 L.Ed.2d 1141 (1978). Farmer v. Farmer, 109 Misc.2d 137, 147 , 439 N.Y.S.2d 584, 590 (Sup.Ct.1981).

. As I have noted, the majority does not agree that all considerations except the race factor were in substantial balance. Ante at 793 n. 40. Evidently it assumes that the evidence regarding a healthy racial and cultural identity in the interracial context was ignored by the court in considering the love and care that would be provided by each family. The majority thus considers that the balance favored the white foster family without consideration of the family’s testimony. This interpretation is at best debatable. But even if one assumes that such a balance exists before consideration of mitigating efforts, and those efforts are then considered separately, along with the impact of the racial difference between parents and child, the proper result would be the same. Since even the best efforts of parents cannot totally eliminate the potential for future problems, the race factor could still tip the balance. See infra at 799.

. The lack of empirical data is especially acute as regards the crucial adolescent years, which typically strain the relations between parents and children and are important in the formation of a healthy self-concept. To date, all of the major studies have focused on pre-adoles-cents, and thus even those researchers most sympathetic to interracial adoption concede that the jury is still out regarding the ultimate viability of such placements. J. Ladner, Mixed Families 109, 249 (1977). The following comment is typical: “In view of the age of the *798 children in the study, it is not possible to give a definitive answer to the question whether white parents can deal successfully with the problem of racial identity.” Grow & Shapiro, Adoption of Black Children by White Parents, 54 Child Welfare 57, 59 (1975) [hereinafter cited as Grow & Shapiro, Adoption ]; see also L. Grow & D. Shapiro, Black Children White Parents 233-34 (1974) [hereinafter cited as Grow & Shapiro I]; B. Jackson, Family Experiences of Inter-Racial Adoption 19 (1976); R. Simon & H. Altstein, Transracial Adoption 49, 187 (1977) [hereinafter cited as Simon & Altstein I]; R. Simon & H. Altstein, Transracial Adoption: A Follow-Up 115 (1981) [hereinafter cited as Simon & Altstein II]; Jones, On Transracial Adoption of Black Children, 51 Child Welfare 156 (1972).

However, the age of the subjects is not the only limitation on the value of the studies in resolving the issue at hand. Most are based largely or exclusively on interviews of the parents rather than direct observation or questioning of the children. See Grow & Shapiro I, supra; L. Grow & D. Shapiro, Transracial Adoption Today (1975) [hereinafter cited as Grow & Shapiro II]; B. Jackson, supra; Simon & Altstein II, supra; C. Zastrow, Outcome of Black Children-White Parents Transracial Adoptions (1977); Grow & Shapiro, Adoption, supra. As a consequence, the data tend to reflect the parents’ experience rather than the child’s, whereas it is the latter’s best interest that must govern the placement decision.

These studies also raise the spectre of observer bias. Most parents who have chosen to adopt a child of another race undoubtedly would not have done so absent a prior belief in the viability of interracial adoption. Moreover, their responses reflect on their own efficacy as parents. For these reasons, their responses are likely to paint an overly rosy picture of the experience and the effects on the child. See Chimezie, Bold But Irrelevant: Grow & Shapiro on Transracial Adoption, 56 Child Welfare 75, 86 (1977) [hereinafter cited as Chimezie, Bold But Irrelevant],

The measures of “success” in some studies are too broad to shed much light on the specific aspects of child development and socialization that are likely to be adversely affected in an interracial adoption, such as learning of “survival skills” for coping with racism. See Grow & Shapiro I, supra; Grow & Shapiro II, supra; B. Jackson, supra; Simon & Altstein II, supra; C. Zastrow, supra; Grow & Shapiro, Adoption, supra. Moreover, the category of “successful” placements covers a broad range from the borderline to the mediocre to the exceptional. Id. Thus there is little indication of the relative success of interracial, as compared to intraracial adoptions. This is a major shortcoming in the context of the instant case, where the court found that either set of prospective parents would provide a good home for the child. As the court observed, the issue in such a case is not whether an interracial adoption would be “successful” but whether there is not a better alternative.

Finally, none of the studies has successfully controlled for non-racial variables. It has often been observed that parents adopting interra-cially tend to rate higher than others on factors such as financial resources, age and experience in parenting, religious conviction, and education. See Simon & Altstein I, supra; D. Robertson, Parental Socialization Patterns in Interracial Adoption, (1974), abstracted in 35A Dissertation Abstracts Int’l 5553 (1975); G. St. Denis, Interracial Adoptions in Minnesota; Self-concept and Child Rearing Attitudes of Caucasian Parents Who Have Adopted Negro Children, (1969) abstracted in 30A Dissertation Abstracts Int’l 2633 (1969). The special problems of racial identity and coping skills may thus be underestimated or obscured due to the effect of other factors which result in a favorable overall assessment. Such findings are misleading when applied to a situation where, as here, the interracial alternative is found not superior, but equal, as to nonrace factors.

. See Grow & Shapiro I, supra note 5; Grow & Shapiro II, supra note 5; B. Jackson, supra note 5, at 115-26; S. Klibanoff & E. Klibanoff, Let’s Talk About Adoption 115-26 (1973); J. Ladner, supra note 5; Simon & Altstein I, supra note 5; Simon & Altstein II, supra note 5; A. Sorosky, A. Baran, & R. Pannor, The Adoption Triangle 202-04 (1978); C. Zastrow, supra note 5; Chestang, The Dilemma of Biracial Adoption, 17 Soc. Work, 100 (1972); Chimezie, Transracial Adoption of Black Children, 20 Soc. Work 296 (1975); Chimezie, Bold but Irrelevant, supra note 5; Grow & Shapiro, Adoption, supra note 5; Grow & Shapiro, Not So Bold and Not So Irrelevant, 56 Child Welfare 86 (1975) [hereinafter cited as Grow & Shapiro, Not So Bold]; Jones, supra note 5; Jones & Else, Racial and Cultural Issues in Adoption, 58 Child Welfare 373 (1979); Simon, An Assessment of Racial Awareness, Preference, and Self-identity Among White and Adopted Non-White Children, 22 Soc. Problems, Oct. 1974, at 43-57; W. Allen, The Formation of Racial Identity in Black Children Adopted By White Parents (1976), abstracted in 37B Dissertation Abstracts Int’l 1888 (1976); D. Robertson, supra note 5; G. St. Denis, supra note 5.

. There is a broad consensus among professionals that interracial adoptions pose problems not present in the intraracial context. In one survey of social workers involved in child placement, 79 percent found that interracial adoptions are riskier. Grow & Shapiro II, supra note 5, at 50. Sixty-nine percent disagreed with the statement that “Growing up in our current-day society is as difficult for a black child in a black home as in a white home.” Id at 51.

. See Grow & Shapiro I, supra note 5; Grow & Shapiro II, supra note 5; B. Jackson, supra note 5; S. Klibanoff & E. Klibanoff, supra note 6; J. Ladner, supra note 5; Simon & Altstein I, supra note 5; Simon & Altstein II, supra note 5; Chestang, supra note 6; Grow & Shapiro, Adoption, supra note 5; Grow & Shapiro, Not So Bold, supra note 6. The position that intraracial adoption is preferable when other factors are equal is generally accepted even among organizations among whose primary functions it is to promote interracial adoption as an alternative for hard to place children. See Simon & Altstein II, supra note 5, at 70-74.

In contrast, the point of view adopted by the majority opinion — that intraracial adoptions should not be regarded as generally preferable with respect to identity formation and socialization of the child — is a comparative rarity among social researchers and commentators. See Jones & Else, supra note 5.

. J. Ladner, supra note 5, at 124 — 25. See also the other authorities cited at note 6 supra.

. The majority of this court prefers to state that “[a] majority of the [Supreme] Court has not accepted this intermediate standard for benign racial classifications to which one must add that neither has the Court rejected intermediate scrutiny in that context. Until the Court as a whole addresses the issue, one cannot simply presume that intermediate scrutiny will be rejected (or adopted). Indeed, more Justices are on record in favor than opposed.

. The importance of such development has been emphasized by sociologist Joyce Ladner: “A healthy integrated personality involves one’s having a stable concept of self as an individual as well as a group (black) identity.” J. Ladner, supra note 5, at 104.

. B. Jackson, supra note 5, at 19.

. One 27-year-old black woman who was adopted by a white woman when she was five, describes the ambivalence that comes from being caught between two cultures: “I certainly couldn’t have been accepted by white people, and I couldn’t fit easily into a black world because, even though I was trying my best, things still didn’t work out.” Ladner, supra note 4, at 158. See also Jones & Else, supra note 5, at 378. In the present case, the trial court heard expert testimony on the importance to the child of acceptance by the larger community.

. See supra at 40.

. See note 3 supra.

. The majority later states that the race factor must be considered “without preference for the race of any party.” Ante at 787. I concur in this conclusion. Certainly there is no justification for a policy favoring black petitioners to white ones or vice versa. However, if the majority’s statement meant that a court cannot take note of whether the petitioners’ race is the same as that of the child, regardless of what that particular race happens to be, I would disagree for the reasons elaborated in the text.

. Emphasis supplied.