A national study of American parents indicates that at age 35, routine disciplinary spanking and hitting a child out of anger are linked to largely different personal and household characteristics. In an overlapping sample followed over five years, parents who spanked at age 35 had over twice the odds of losing their temper and hitting or throwing things at their children at age 40. Published in Family Process, the findings suggest everyday spanking and anger-driven hitting may have distinct roots even as milder discipline predicts harsher aggression later on.
Corporal punishment refers to using physical force to correct or control a child’s behavior, from open-handed spanking on the buttocks to washing a mouth out with soap or hitting with an object. Over one in three U.S. parents use physical discipline, despite opposition from groups like the American Academy of Pediatrics citing harm to children. Although some analyses argue mild, infrequent spanking has minimal negative effects, harsher forms can cross into physical abuse.
Studying physical discipline is complicated because surveys often group very different behaviors together. Prior work links physical punishment to personal history, beliefs, mental health, and stress; for instance, an earlier study of parents found that adults with difficult childhoods, especially those spanked growing up, were over twice as likely to spank their own children. Yet routine discipline and harsher aggression can have different predictors. In a 2019 study of 310 Black parents, higher income within a low-income sample predicted less frequent physical discipline, yet higher odds of physical maltreatment leaving bruises or drawing authorities’ attention.
Researchers have also asked whether milder spanking increases the risk of harsher aggression later on. Most studies linking spanking to abuse rely on snapshots or childhood memories rather than tracking parents forward. One exception was a 2012 study that followed mothers and their children from ages 6 to 8, finding that 17 percent of mothers who mildly spanked at age 6 spanked more often or with an object at 7, versus 12 percent who did not spank at 6. Still, that study used a relatively small mother-only sample, spanned one year, and omitted other personal or demographic factors.
Don’t miss the next big psychology finding
Every morning, PsyPost sends a short briefing on the most interesting new peer-reviewed research in psychology and neuroscience.
To examine both questions nationally, Christopher J. Mehus of the University of Minnesota and colleagues conducted an exploratory study of two overlapping samples of parents surveyed at ages 35 and 40. Their work builds on a 2020 study by two of the same authors using this shared dataset, showing over a third of 35-year-old parents spanked from 1993 to 2017. The team compared predictors of spanking and hitting in anger at 35, and tested whether spanking at 35 predicted hitting in anger at 40.
The team analyzed licensed data (available with survey program permission) from the Monitoring the Future Panel Study, which surveys U.S. high school seniors annually and randomly selects 2,450 per class for adult follow-ups. Wave-to-wave response rates averaged 89.5 percent at age 35 and 93.0 percent at age 40, while overall retention from 12th grade through age 40 averaged 48.4 percent. Statistical weights accounted for dropout so analyses continued to reflect the original 12th-grade classes.
Two items asked parents how often they spanked their children and how often they lost their temper and hit or threw things at them. Seven options from “never or almost never” to “every day” were grouped into “never or almost never” versus several times a year or more.
Other measures included sex, living with a partner, holding a bachelor’s degree, conservative political views, attending religious services at least monthly, number of children under 18 at home, and graduating class year. Race or ethnicity, recorded in 12th grade, served not as an inherent trait, but as a rough stand-in for shared cultural values and experiences of systemic oppression.
Parents also completed a four-item depressive symptoms scale rating agreement with statements like life often seeming meaningless and the future seeming hopeless. Drinking items reflected diagnostic criteria for alcohol use disorder, capturing standard signs of problem drinking like trouble limiting intake or daily-life harm, coded as non-disordered drinking (zero or one criterion), disordered drinking (two or more), or no recent drinking.
The first analysis covered 18,432 parents from 28 classes surveyed at age 35 from 1993 to 2020 with children aged 2 to 12 at home (84 percent lived with a partner; 76 percent were white). At age 35, 52 percent reported neither behavior, and 34 percent reported spanking without hitting in anger. Another 9 percent reported hitting or throwing things out of anger, about two-thirds of whom also spanked.
Comparing each discipline group with parents using neither practice indicated that spanking and hitting in anger had mostly distinct associated factors at age 35. Living with a partner and conservative views predicted higher odds of spanking alone, while a bachelor’s degree predicted modestly lower odds. For hitting in anger, associations with all three factors were not statistically significant.
Mental health and drinking showed the reverse split: depressive symptoms, disordered drinking, and no recent drinking (versus non-disordered drinking) each predicted higher odds of hitting in anger, while their links to spanking alone were not statistically significant. Fathers had modestly higher odds than mothers of spanking alone, yet modestly lower odds of hitting in anger. Frequent religious attendance and more children predicted higher odds of both behaviors.
Racial and ethnic patterns at age 35 also varied compared with white parents. Black parents had roughly twice the odds of spanking alone and nearly three times the odds of hitting in anger. Hispanic parents had about two-and-a-half times the odds of hitting in anger, while their odds of spanking alone were not statistically significant. Asian parents had lower odds of spanking alone and higher odds of hitting in anger, though Asian respondents made up just 1.7 percent of the sample (249 parents), making those estimates far less precise.
The second analysis examined an overlapping sample of 11,460 parents from 23 classes surveyed at age 35 (children aged 2 to 12) and age 40 from 1998 to 2020 (children aged 7 to 17). Instead of a three-group outcome, it tested whether parents hit or threw things in anger at 40, endorsed by 8.6 percent (versus 91.4 percent who did not). Narrowing the outcome focused on escalation from spanking and reflected that routine spanking declines by adolescence.
Adjusting for age 40 factors, parents who spanked without hitting in anger at 35 had over twice the odds of hitting or throwing things in anger at 40, versus parents using neither behavior at 35. Parents who already hit in anger at 35 had over eight times the odds of doing so at 40, though against the 8.6 percent overall rate at age 40, the behavior remained far from universal.
Because the model accounted for age 35 discipline, remaining age 40 predictors reflect incremental risk beyond what those factors contributed at 35. Disordered drinking, depressive symptoms, more children, and identifying as Black or Hispanic (versus white) at 40 each predicted higher incremental odds of hitting in anger. The estimate for the 130 Asian parents (1.4 percent) pointed the same way, though it was imprecise and not statistically significant. Rerunning the model with age 35 drinking data, addressing a post-2016 shift from a five-year to a 12-month alcohol timeframe, yielded no substantive difference.
Other age 40 measures, including sex, partner status, education, conservative views, religious attendance, and abstaining from alcohol, were not statistically significant. Graduating class year also showed a notable asymmetry: both spanking and hitting in anger declined slightly across historical classes at age 35, yet class year was not statistically significant at age 40 once age 35 behavior was controlled, indicating no further historical decline over those five years.
The authors offer two competing explanations for the distinct age 35 profiles. Spanking and hitting in anger may have genuinely different roots, with spanking tied to traditional norms and demographics and hitting in anger tied to emotional distress, pointing to different prevention approaches. Alternatively, the split could reflect how parents interpret the words “spanking” and “hitting” on a survey, whether or not their physical actions differ.
The authors also suggest higher odds of hitting in anger among age 35 alcohol abstainers could reflect health or social challenges in that small group, while sex differences might reflect caregiving time. Citing prior scholarship, they propose higher rates in some minoritized groups may reflect community norms, beliefs rooted in historical oppression that strict discipline keeps children safe, or strain from systemic racism.
As with all research, there are caveats to consider. Because the study used observational surveys rather than an experiment, it cannot establish that spanking at 35 causes hitting in anger at 40; unmeasured stress or child behavior challenges could contribute to both. Physical punishment was also measured by two brief items without a timeframe or definitions, allowing varied interpretations, and surveys did not record the disciplined child’s exact age.
Because participants were recruited in 12th grade and first asked about discipline at 35, the findings omit high school dropouts and may not generalize to younger parents. Losing roughly half the initial samples by age 40 may also slightly underestimate physical punishment despite weighting. The authors call for long-term studies with detailed parenting measures to test whether spanking and anger-driven hitting have separate causes and how mental health and substance use shape discipline over time.
The study, “Exploring Factors Associated With Self-Reported Parental Spanking and Hitting Out of Anger: Cross-Sectional and Longitudinal Analyses Using National Panel Data,” was authored by Christopher J. Mehus, Ana Mireya Díaz-Howard, Yunqi He, Jaime Ballard, Yvonne Terry-McElrath, Mona Mittal, and Megan E. Patrick.