The answer to at least some of these questions must be yes, if the child is a growing human being and not a robot. Experimenting with autonomy and observing how his parents react is part of the job of a child. Setting age-appropriate boundaries is the role of the adult.

The dynamics become even more complex when a child is defiant or impulsive by nature, when a parent is under inordinate pressure, or all of the above. That is perhaps one reason why two-thirds of American parents, when asked by the federally funded General Social Survey in 2016, agreed with the statement, “Sometimes a child just needs a good, hard spanking.” (The number has dropped by about 15 points in the past three decades.)

A host of studies link spanking to later behavior problems. A 2016 meta-analysis of five decades of research on the topic suggests that spanking a young child is not only an ineffective form of discipline but a catalyst for more serious acting out and mental health problems in the future. Indeed, corporal punishment of children is now illegal in 53 countries, and banning any kind of hitting of children—with a hand or an object—is a growing international movement.

Whether striking a preschooler’s bottom with an open hand discourages or exacerbates misbehavior remains a controversial topic in the U.S. Adding grist to the debate: The studies that have been conducted are observational—that is, they show that spanking and future behavior problems are tightly linked but not that the former definitively causes the latter. Children can’t be randomly assigned, for experimental purposes, to spanked and not spanked groups, so it’s hard to discern whether later behavior problems can be attributed to that one factor.

A new study led by Elizabeth Gershoff, a professor of human development at the University of Texas at Austin, aims to settle this dispute. Published last month in the journal Psychological Science, the study statistically controlled for children’s initial behavior problems and the characteristics of their parents. More than 12,000 American families were surveyed, from their children’s kindergarten year through eighth grade, as part of the nationally representative Early Childhood Longitudinal Study.

The researchers paired subjects who had and had not been spanked at 5 years old but were equivalent on 38 other factors. Those included the child’s initial level of behavior problems as rated by the teacher, and the parents’ marital status, mental health, stress levels and parenting style as defined by their answers to interview questions.

The researchers found that a child who was spanked at age 5 was far more likely to have behavior problems at age 6, and more serious ones again at age 8, according to teachers’ ratings. The relationship between corporal punishment and later acting out was even stronger if parents said that they had spanked the 5-year-olds the week before the survey, an indication that spanking may have been relatively frequent.

“This is the closest we can get, outside of an experiment, to say that spanking causes negative changes in children’s behavior. I can’t think of another way to explain our results,” Ms. Gershoff told me.

The American Pediatric Society advises parents to avoid spanking, and the American Psychological Association cautions against the practice. American parents seem to be left with a choice: To use a form of physical discipline that gambles with the future of their children or to find other ways to help them learn self-control.