by The evolutionary psychologist Gad Saad has a favorite classroom ritual. He ran through it again on a podcast not long ago. When he teaches at Concordia University, he asks his students to name the strongest predictor of child abuse in a home. The first answer is usually an absent parent. Then come the rest: alcoholism, poverty, a parent who was beaten as a child. He writes the answers on the board and calls them all reasonable. Then he tells the class that none of them is right. The number one predictor, he says, is a stepparent, and the effect is a hundredfold, a ratio he describes as astronomically larger than anything science ordinarily publishes. He reports getting goosebumps every time he tells it.
The claim deserves better than the performance. It rests on real research, some of the most replicated work in the study of family violence. It has also been compressed past accuracy for the microphone. Both facts matter. The subject is children. Children are owed precision.
The research belongs to Martin Daly and Margo Wilson of McMaster University. Beginning in the late 1970s, they asked a Darwinian question about a domestic crime: whether a child’s risk of abuse depends on who stands in as a parent. They named the pattern they found the Cinderella effect, after the folktale stepmother who starves her husband’s daughter while her own daughters eat.
In the Grimms’ version of Cinderella, the stepmother hands her own daughters a knife so they can cut off pieces of their feet and fit the golden slipper. Her ambition is spent freely on her own blood while the stepchild sleeps in ashes. Cruel stepparents recur in folk traditions around the world. Daly and Wilson read the tales as a cultural memory of a statistical truth, then went looking for the truth in case files.
The word itself already remembered. In Old English, steopcild meant an orphan, a child bereaved by death. The prefix descends from a verb meaning to bereave, and a stepmother was originally a woman who replaced a dead one. Divorce made the word modern. The old meaning kept its grim accuracy about which children history worried about.
In 1985, in Ethology and Sociobiology, Daly and Wilson reported the findings. Preschoolers living with one natural parent and one stepparent were about forty times more likely to become registered abuse cases than children of the same age living with two natural parents. Income could not explain the gap. Family size and the mother’s age at the child’s birth could not explain it either. The stepparent himself was the risk factor.
Combing Canadian records from 1974 to 1990, Daly and Wilson counted 178 children under five killed by genetic fathers and 67 killed by stepfathers. The raw counts favor the fathers until exposure is counted. Fewer than one percent of Canadian preschoolers lived with a stepfather in those years, while about nine in ten lived with their genetic fathers. Per child, per year, the stepfather’s killing rate ran roughly sixty to seventy times higher. American data for children under two put the figure near a hundred. Saad’s number has a source.
One detail ends the suspicion of a just-so story: the weapon. In the Canadian files, stepfathers who killed almost always beat the child. A stepfather was about 120 times more likely than a genetic father to beat a preschooler to death. Genetic fathers who killed more often shot or suffocated, and many of those killings ended in the father’s own suicide, a despairing man destroying what he counted as his. The stepfather’s killings read as rage at something in the way. A theory about evolved motive turned out to predict the means of killing. The killers are nearly always stepfathers or boyfriends. Stepmothers almost never kill.
Missouri supplied the American replication. In 2005, in Pediatrics, Schnitzer and Ewigman analyzed eight years of child fatality reviews covering 149 children under five killed by caregivers. Children living in households with an unrelated adult in residence were nearly fifty times as likely to die of inflicted injury as children living with two biological parents. More than four of five of those households housed the mother’s boyfriend. In about three of four of them, he was the killer.
The same Missouri study buries a lazy phrase. Children living with a single mother and no other adult in the home showed no elevation in the risk of fatal assault.
The hazard was never the broken home.
The hazard was the new man in it.
In a 1984 survey of 930 San Francisco women, published in Child Abuse & Neglect, the sociologist D. E. Russell found that one in six women raised with a stepfather had been sexually abused by him. For biological fathers the figure was one in forty. The stepfathers’ abuse was also more often severe.
The Fourth National Incidence Study of Child Abuse and Neglect is the country’s largest periodic measure. In 2010 it reported that children living with one parent and that parent’s unmarried partner suffered abuse at more than ten times the rate of children living with married biological parents. Children of a single parent living alone were abused at roughly a third of that rate. Method after method, decade after decade, the gradient points one way.
Swedish researchers reported in 2000 that stepchildren in Sweden faced no excess risk of being killed. Daly and Wilson answered that the Swedish team had measured risk against the whole child population. The proper comparison pool was the small number of children living with stepparents. Corrected, the Swedish figures still show an elevated rate, nearer eight times than seventy. Other scholars still dispute how much of the gap is selection, the men and the strain that arrive together in a stephousehold. Selection gives no discount.
The verdict on Saad’s claim is precise. He is right about rank order. Alcoholism matters. Poverty matters. Neither comes close to the presence of an unrelated man in the house as a predictor of abuse. At the lethal end, the gap is the largest well-documented effect in the field. His hundredfold figure measures the rarest and worst outcome at the far end: fatal assaults on the youngest children, counted per year of exposure. His gloss on odds ratios, in which 1.2 means twenty percent more effective, would fail his own first lecture, because an odds ratio of 1.2 describes odds, and odds are not effectiveness. The science is stronger than his delivery of it, and less tidy than his number.
Parental effort is precious, and natural selection builds animals that spend it where it buys descendants. A new man in a house gains nothing genetically from another man’s child, while that child absorbs years of the mother’s care and resources that could have gone to his own future children. Sarah Blaffer Hrdy documented the extreme version among langur monkeys in the 1970s, where males taking over a group killed the nursing infants so the mothers would breed with them, and lion researchers have recorded the same pattern. Human fathers mostly refuse the arithmetic.
The great majority of stepfathers never harm a child, and many are the steadiest parent in the house. And absolute risk stays small. Even multiplied many times, fatal abuse remains rare. In the most recent federal count, for fiscal 2024, 1,773 American children died of abuse and neglect of every kind, and infants carried the largest share of that toll. The gradient is a map for policy and vigilance.
Adoptive parents share no genes with their children, and their maltreatment rates run low, because adoption screens for adults who actively chose the child. The danger zone is a specific configuration of circumstance. An unrelated man arrives after the child’s birth, a man who did not choose the child, often young, often strained, treating the child’s claims on the mother as a tax on his own prospects.
Prevention starts with fathers. In the Missouri data, a resident biological father cut a child’s risk of fatal assault by more than half. Courts and policies that keep a safe father engaged after separation are abuse prevention, whatever else they are called.
Money is prevention too. In 2017, Kerri Raissian and Lindsey Rose Bullinger showed that a one-dollar increase in a state’s minimum wage produced a nearly ten percent drop in child neglect reports, with the largest effect among children under five. Poverty does not cause abuse, and most poor parents are loving.
The strongest single program on record is a nurse on a doorstep. In the Elmira trials begun in the late 1970s, David Olds and his colleagues sent nurses into the homes of poor first-time mothers, from pregnancy through the child’s second birthday. By the time those children were fifteen, state-verified abuse and neglect in the highest-risk families had fallen by seventy-nine percent. Nothing else in the prevention literature has matched it.
Abuse is a crime of privacy, and its natural enemy is routine contact between children and adults outside the house. Education personnel filed about one in five of the maltreatment reports in the federal count for 2024, among the largest sources of all. The pandemic ran the experiment no one wanted. When schools closed in March 2020, Florida’s abuse allegations ran twenty-seven percent below expectation, a gap the economist E. Jason Baron and his colleagues scaled to more than two hundred thousand missing allegations nationwide in two months. Over the same weeks, a Johns Hopkins pediatric trauma center treated a significantly higher share of children injured by abuse. The children were still being hurt. The watchers had been sent home.
The podcast never reaches it: the parent who stays, and who defends him. Clinicians call her the nonoffending parent. The research on what follows is blunt. In the literature on child sexual abuse, reviewed most comprehensively by Elliott and Carnes in 2001, the strongest consistent predictor of a child’s recovery is the belief and support of that parent. Children whose mothers believe them and act recover better on every measure the studies use. The mother who defends the man doubles the wound. Her child loses safety twice, once to the blows and once to the defense of them. The child learns to stop telling.
Most states criminalize a parent’s failure to protect. Oklahoma’s statute punishes enabling child abuse in the same range as abuse itself, up to life in prison. The principle is old and sound: protection is the parent’s first duty, and choosing the partner over the child is an act with a charge attached.
In Texas in 2006, Arlena Lindley tried to pull her boyfriend off her three-year-old son, Titches, and was shoved to the floor and forced from the apartment. The boy was beaten to death. The boyfriend took a life sentence. Lindley, a battered woman, took forty-five years for failing to protect him. Reporters at BuzzFeed later counted twenty-eight mothers in eleven states serving a decade or more under such laws, many of them beaten themselves. In Oklahoma, Tondalo Hall drew thirty years while the man who broke her children’s bones drew two.
A mother who shields the abuser abandons her first duty, and the law is right to reach her. A mother who is herself terrorized is a second victim, and a statute that cannot tell terror from allegiance is a bad instrument. Prosecutors need the knowledge to separate the two, and courts need to treat protective capacity as a custody question long before it becomes a felony question. The child cannot wait for a body to make the argument.
A grandparent or an aunt keeps showing up. Teachers notice the flinch and ask twice. The pediatrician charts the bruise pattern. Someone on the street makes the report that every state permits and that professionals are bound to make. The resilience research keeps returning to one factor above the rest: a single steady adult, outside the trouble, who believes the child the first time.
One more lesson belongs to the child in that house. Watching the defense, he learns what no later explanation erases: the adult who was supposed to be safest chose someone else.
Saad gets goosebumps from a ratio. The ratio was earned in small bodies, in the one American child lost to abuse or neglect roughly every five hours. Nearly half are babies under one. It maps where watchfulness pays: the wage floor, the nurse on the doorstep, the teacher who asks twice, the statute that names the defender, the grandmother who keeps coming.
The old word is still the truest one. Steopcild, the bereaved child. It remembers that a chair was emptied before anyone filled it again. When a new man sits down at that table, the kin and the institutions around that house owe the child one thing above all: eyes that stay open, and one adult who believes him the first time he tells.
Leave a Reply