On July 21, the Wall Street Journal ran a column by Elizabeth Bernstein under the headline “Are Therapists to Blame for the Rise of Adults Cutting Off Their Parents?” The print edition softened it to “Some Therapists Encourage Estrangement to Help Heal.” The piece opens with Shoshana Rubli, who says a therapist told her in only their second session to cut off her mother — an estrangement that lasted ten years and that she now deeply regrets. From there it builds to its central statistic: “Among some 7,000 estranged parents surveyed in April a third believed their adult child’s therapist had influenced or recommended the decision to end the relationship.”
The piece traveled fast. Within days, PJ Media repackaged it as “Is This the Reason Your Leftist Relatives Are Cutting You Off?” and the Daily Wire as “This Insane New Therapy Trend Is Destroying Families.” A reported relationships column became, within one news cycle, a citation for the claim that therapy is dismantling the American family.
Parts of the column describe something real. If a therapist recommended estrangement in a second session, without adequate assessment or any exploration of alternatives, that would be poor practice. Rubli’s account suggests that may have happened — though the column offers only her retrospective telling, with no clinical notes and no response from the therapist. Several therapists responding to Mo News agreed that the approach she described would be wrong. William Doherty, the University of Minnesota family scientist the Journal quotes, has published a small pilot study of a related concern: clients in individual therapy commonly recalled therapists making “relationship-undermining statements” about their romantic partners, and those recollections went along with worse relationship outcomes. The clinical concern deserves the attention it is getting. The broader causal claim has not been established.
I am a data scientist who works on issues of social justice, and much of that work comes down to one habit: before accepting the story a number tells, find out how the number was made. So I went looking for the source behind every empirical claim in the column. Here is what I found. First, Anderson’s public survey archive cannot be reconciled with the column’s central number: the surveys it contains are self-selected, one-sided, and explicitly noncausal, and they do not add up to “some 7,000” parents surveyed in April. Second, the “rise” in the headline rests on no repeated national measure. Third, the evidence that does exist — on why people leave, what leaving costs, and what therapists actually do — is more limited, mixed, and noncausal than the column implies.
A number the column never shows
Start with the 7,000. The column attributes the survey to Peter Anderson, a licensed marriage and family therapist in Beverly, Mass., “who helps families with estrangements and conducted the surveys.” The only public Anderson survey archive I could find, posted to OSF by Anderson and Richard Blake, documents two self-selected surveys run in 2024–25, with roughly 1,027 and 4,226 respondents — about 5,253 in total, not “some 7,000” surveyed in April. The materials describe recruitment through social-media channels for people interested in estrangement and a parent-only sample, and call the work — in the authors’ own words — citizen science rather than peer-reviewed research; they warn that the respondents overrepresent parents who feel wronged, and they state plainly that the data do not show therapy causes estrangement. The authors also disclose a professional interest in criticizing certain therapeutic practices. These materials cannot be matched to the survey as the Journal describes it — a different documented sample size and timing — and the column does not explain whether its figure comes from a later survey, an expanded sample, or some combination of datasets. The Journal should say what survey produced “some 7,000” parents, an April fielding, and a one-third figure, and publish its questionnaire, recruitment, and denominator.
Now look at what the number measures even if the survey were pristine. The load-bearing word is believed. As the column reports the statistic, it records parents’ beliefs about therapy’s role — not therapists’ records or reports collected directly from adult children. The column does not disclose how respondents learned about the therapy, whether the adult child described the therapist’s advice, whether a parent attended any session, or precisely what respondents were asked. The statistic therefore cannot directly measure what therapists did.
That gap between belief and knowledge matters, because parents and adult children explain these ruptures differently. A 2015 study by Kristen Carr and colleagues compared separate, non-matched samples of 546 estranged parents and 352 estranged adult children. Parents most often attributed the rift to their children’s objectionable relationships or sense of entitlement, whereas adult children most often cited their parents’ toxic behavior or feeling unsupported and unaccepted. Joshua Coleman, a prominent advocate of the therapist explanation, summarizes a related asymmetry: parents are more likely to blame “divorce, their child’s marriage or their child’s therapy.” The methodological point is that an attribution survey of parents measures parents’ accounts — not therapist behavior, and not which account is correct.
And even if the one-third figure were accepted, it would estimate parents’ beliefs — not how often therapists recommend estrangement.
Four polls are not a trend line
The headline takes “the Rise” as a given. Inside the column, the only support is one researcher’s stated belief that “more rifts are likely.”
Here is the actual inventory. Karl Pillemer’s national survey of 1,340 American adults, published in 2020, found 27 percent estranged from a family member and 10 percent from a parent or child. A 2022 YouGov poll put estrangement from an immediate family member at 29 percent. A 2024 Harris poll put immediate-family estrangement at 35 percent. A 2025 YouGov poll found 38 percent estranged from at least one of five listed relatives — sibling, parent, child, grandparent, grandchild.
Plot 27, 29, 35, 38 and you see an epidemic. Read the four questionnaires and it dissolves: different polling houses, different definitions, different lists of relatives, different response scales. No comparable repeated national measure — the same instrument, the same population, at two points in time — has been cited to show estrangement rising. (Panels that do follow the same people, like a ten-wave German study, measure how common estrangement is, not whether it is climbing nationally.) That is an open question, not a premise for a headline; even a recent rise-framed article in Knowable Magazine concedes that researchers have not quantified how much estrangement has increased.
Longitudinal data also suggest estrangement is anything but new. Analyzing the National Longitudinal Survey of Youth — a federal study that has followed the same families for decades — Rin Reczek and colleagues found that 26 percent of adult children experienced a period of estrangement from a father and 6 percent from a mother, with first estrangements at average ages of 23 and 26. Reczek defines estrangement broadly — little or no contact, or limited contact paired with low closeness — not only a declared, permanent “no contact.” The panel detected such periods as early as 1994, long before estrangement TikTok, “no contact” hashtags, and Instagram therapists. Pillemer’s own 2020 framing was not that estrangement was new. It was that it had been “hiding in plain sight.”
What the estranged actually say
Now weigh the evidence that does exist. The column’s cultural theory — “everything becomes trauma,” minor grievances recast as abuse — is checkable against the reasons estranged people themselves give. In the 2025 YouGov poll, adults estranged from a parent most often cited manipulative behavior (34 percent), physical, emotional, or sexual abuse (34 percent), lies or betrayal (31 percent), and abandonment or neglect (29 percent). Political disagreement was cited by 7 percent.
Are those self-reports? Of course. But the column cannot have it both ways: it treats adult children’s accounts as social-media contagion while treating parents’ beliefs about unseen therapy sessions as data. Pick an evidentiary standard and hold it for both sides of the rupture.
The qualitative research undercuts the fragility story from another angle. Kristina Scharp, who has studied estrangement for over a decade, describes it as a slow, costly process that people struggle to achieve and struggle to maintain: “Typically something very bad has happened for someone to decide they are completely done with their family. There’s definitely a misconception that adult children are overreacting, but I haven’t seen that.”
The cost no one can find
The column’s most quotable research claim is that people who initiate cutoffs “experience significantly higher rates of chronic unhappiness and depression over time,” a finding attributed to “research from Cornell’s Pillemer and colleagues” and given a name: the “exiter cost.”
I could not find it. The phrase “exiter cost” appears in the column and in pieces quoting the column; I found no identifiable Pillemer publication or public report supporting this specific claim about initiators, statistical differences, and change over time — not in his published studies, not in the public materials around his book, which describe distress on all sides of a rift, not on Cornell’s own pages about his work. It may derive from unpublished analysis. But a claim this specific deserves a citation, and the column supplies none.
The peer-reviewed outcome literature is more careful. A German panel study of 5,245 adults found estrangement — from parents or siblings — correlated with lower well-being, a correlation that, by the study’s own design, cannot say whether estrangement causes distress or whether preexisting conflict or other factors contribute to both. And in a large U.S. analysis of maternal ties, Reczek and colleagues compared estranged adult children with those in warm relationships and those in high-contact but socially negative ones. The estranged children’s depressive symptoms were statistically indistinguishable from those with warm ties; their self-rated health resembled that of people in the high-contact, low-quality relationships, which carry costs of their own. For many people, in other words, the relevant comparison may not be estrangement versus a happy family but estrangement versus remaining in a low-quality relationship — and staying is not free.
Estrangement is also not always permanent: in the NLSY data, among adult children observed after an estranged period, 81 percent of maternal and 69 percent of paternal relationships later no longer met the study’s estrangement threshold. These figures do not establish full or durable reconciliation, but they show that many episodes meeting this study’s definition were impermanent.
The evidence on therapists, stated by its champion
Which brings the argument back to where the column started: the therapists. What does research say about therapy’s role in estrangement? Almost nothing estimates how often therapists steer clients toward estrangement, or whether such advice causes it. The closest study analyzed open-text answers from 209 members of a UK estrangement-support charity, who described counseling as most helpful when clinicians knew the estrangement literature and supported clients’ own decisions rather than steering them; it cannot tell us how often steering happens. Doherty’s undermining-statements study, the strongest card in the column’s hand, concerns romantic partners, not parents, and rests on 101 clients’ recollections.
The only quantitative evidence the column offers for a common causal role is an unexplained survey statistic whose nearest public counterpart is self-selected, parent-only, and explicitly noncausal. Beyond that is expert opinion, including the sentence Coleman gave Knowable Magazine this year: “I don’t believe there’s research on this topic, but I do think that therapists are a big agent of estrangements.”
Read that twice. A prominent advocate of the therapist explanation says he knows of no research establishing therapists’ role, then offers his own judgment anyway. He may turn out to be right — but a belief is not a finding, and no amount of quotation turns one into the other.
The tragic exception
The column’s central number cannot be reconciled with Anderson’s public archive. Its headline trend is not established by any comparable repeated national measure. Its “exiter cost” lacks an identifiable citation. And its causal implication remains unsupported by representative or causal evidence.
One more fact belongs next to those. This was not an op-ed. Bernstein writes a reported relationships column, outside the Opinion section, translating social science for millions of readers. Journalism of that kind owes readers what science owes everyone: sources that can be checked. The fix is cheap, and it is the Journal’s to make. Say which survey produced the “some 7,000” figure, and publish its methodology. Cite the study behind the “exiter cost.” Show the trend line, or stop implying one exists.
The stakes are not academic. Estranged parents are in real pain, and treating a therapist as the sole cause risks becoming a counsel of helplessness. Among the 100 reconcilers Pillemer interviewed, almost all reported abandoning the need for the relative to accept their version of the past or to apologize. That descriptive finding does not establish what causes reconciliation, but it emphasizes the agency the participants themselves retained. Adult children are owed something too. In the 2025 YouGov poll, 34 percent of respondents estranged from a parent cited physical, emotional, or sexual abuse, and they deserve not to have their hardest decision recast as a wellness trend.
Doherty told the Journal that cutting off a parent “should be the tragic exception… not business as usual in therapy.” Whether cutting off a parent is business as usual in therapy is exactly what has not been established with representative or causal evidence. In journalism, though, the standard is not in dispute: a causal-epidemic framing whose central evidence readers cannot check should be the tragic exception. Not business as usual.
This is the kind of data-driven justice work I do in my book Unlocking Justice, now available from Princeton University Press.