EASTERBROOK, Circuit Judge.

Rodney Scott lost his job as a social worker at Parkview Memorial Hospital when the Hospital reorganized. The number of nonsupervisory social workers dropped from nine to six; the survivors acquired additional responsibilities- — principally dealing with insurers and health maintenance organizations. The Hospital decided to make the transition by requiring its existing staff to apply for positions as if all social work slots were being abolished and new jobs created; the losers at this game of musical chairs would be let go. Scott was not eligible for the new supervisory positions, which required a registered-nurse license, and thus was limited to the six line jobs. The Hospital used a series of interviews to make its decision. Scott lost at the second phase. A panel of interviewers scored the candidates’ answers and passed those who achieved 39 or higher; with a score of 32, Scott was filtered out. Five candidates made it to the third round. Because only four of these passed the third round, the Hospital ended up filling two positions from outside; both of the new hires were women. Scott filed this suit claiming to be a victim of sex and age discrimination. He lost on summary judgment.

Scott’s fundamental problem is that he has no direct evidence that sex or age *525 played a role in the decisions. None of the stated criteria is linked to sex or age. None of the participants in the process said that sex or age mattered. Although younger women seemed to do better in the process — three of the four people who flunked the second round of interviews are men, and all had reached 40 (Scott himself was 46) — too few were involved for the numbers to be significant. The age differences also were slight; the candidates who passed Round Two ranged from 32 to 46, while those who did not pass spanned 42 to 48. These modest differences are not enough to imply that the Hospital was searching for a way to screen out people on the basis of preconceptions about age, especially when we add in the fact that one of the persons hired from outside to make up the complement of six was 51 at the time. Few conceive of people in their 40s as “old” in relation to those in their mid to late 30s. An inference of discrimination is appropriate only when the employer favors “substantially5’ younger persons, a term we have defined operationally as “ten years or more.” See Miller v. Borden, Inc., 168 F.3d 308, 313 (7th Cir.1999); Hartley v. Wisconsin Bell, Inc., 124 F.3d 887, 892-93 (7th Cir.1997). So the age-discrimination claim goes nowhere, and the sex-discrimination claim is only slightly stronger.

Social work traditionally has been a disproportionately female profession. Thus it is not surprising that most members (six of seven) of the interview panel that rated Scott and the other eight incumbents were women. He contends that their questions signaled favoritism toward women. One inquiry was: “Describe a situation where you went beyond your normal responsibilities in order to meet a patient’s needs.” Scott thinks that this indicates a preference for attitudes that he is sure predominate among women. Notes on another person’s interview include “smiles warm,” a notation missing from Scott’s file; he is sure that this likewise shows how the panel favored women. The interview process was subjective, so biases could be both indulged and disguised, Scott insists.

Let us suppose that women are more likely than men to display caring or generally warm-and-fuzzy attitudes— though this may be a stereotype about stereotypes, rather than an accurate description of traits in the population. How would this imply sex discrimination? Scott does not- deny that caring about others’ welfare, and eagerness to assist strangers, are appropriate traits for social workers. They are appropriate even when, as at Parkview Hospital, social work is evolving to include more emphasis on negotiations with third-party payors. Questions about engagement with clients’ needs are no less appropriate for social workers than questions about aggressiveness toward adversaries would be when hiring trial lawyers. Subjective interviews could be smokescreens for bias, but in professions such as social work (or law, medicine, architecture, and many others) they are also necessary; no formulary of approved answers can replace a nuanced evaluation of candidates. Nor does federal law require private employers to behave as if they were running bureaucracies, and to prefer paper-heavy evaluations over contextual assessments by knowledgeable reviewers, or to exalt an assessment of past conduct over a prediction of future performance. Unless the evidence demonstrates that an open-ended process was used to evade statutory anti-discrimination rules, subjectivity cannot be condemned. Diettrich v. Northwest Airlines, Inc., 168 F.3d 961 (7th Cir.1999).

Whatever potential for discrimination interviews create when hiring into a profession is unlikely to continue after a person is in that profession. Suppose supervisory social workers routinely disfavor aggressiveness and prefer warm smiles and going an extra mile. That might lead to unequal representation in the profession, if traits such as empathy are distributed unequally by sex. But it would not cause further disparate effects when selecting from among social workers. Suppose more *526 men than women have aggressive, competitive personas. That might explain why more men than women choose to be professional athletes, but it would not imply that male professional athletes are more competitive than female professional athletes, or that an evaluation of “competitiveness” would disfavor women seeking advancement (or longer careers) in sports. Just so with social work. People enter the field because they want to help others. A selection process limited to social workers therefore can ask about warmth or helpfulness without predictably screening out men, just as law firms can screen their existing trial lawyers for dogged combativeness without discriminating against women.

Discount the imprecations against subjective screening, and Scott has no significant evidence that his sex played any role in the Hospital’s decision. The ultimate decision may be good or bad, but the employment-discrimination laws do not allow judges to second-guess managers. We inquire not whether the decision is right, or was reached after the process judges would use were we running the business, but whether the defendants’ submission that sex (and other forbidden characteristics) played no role is honest. Evidence Scott has assembled could not support a conclusion that the Hospital’s description of the reasons for its selection is dishonest. The judgment therefore is

AFFIRMED.