Earlier research had produced similar results, but most had not taken into account how aggressive kids were to begin with, and other factors could have biased the results.
Although the new study doesn't prove that corporal punishment causes aggression by itself, it shows that the link remains even after excluding a broad range of possible explanations.
"That is really a key point that sets the study apart," said Catherine A. Taylor, of Tulane University School of Public Health and Tropical Medicine in New Orleans, who led the research, published in the journal Paediatrics.
"Causality is extremely difficult to prove," Taylor told Reuters Health. Still, she added, "the evidence is at a point where we want to encourage parents to use techniques other than spanking that can actually lower children's risk for being more aggressive."
Taylor and colleagues examined data from an earlier, population-based study of families from 20 large cities in the US. For that study, researchers had interviewed mothers when their children were three years old and again when they were five. Based on the children's behaviours, roughly half were categorized as "higher aggression," and roughly half as "lower aggression."
More than half of the nearly 2,500 kids had been spanked in the month before the interview. And those who had been swatted more than twice at age three had twice as high odds of being highly aggressive at age five. Even after accounting for baseline differences in aggression and other factors - for instance, psychological maltreatment, maternal depression and substance abuse - the odds remained increased.
Although the researchers based their findings on what mothers told them, they echo the data behind anti-spanking recommendations by several professional societies, including the American Academy of Paediatrics and the American Psychological Association.
"The evidence is clear that spanking does lead to aggression," psychologist Sandra A. Graham-Bermann, who was not involved in the new study, told Reuters Health in an e-mail.
Graham-Bermann, of the University of Michigan at Ann Arbor, recently chaired an American Psychological Association division task force that reviewed the research on corporal punishment.
She said spanking - defined as open-handed hitting that does not injure the child - makes children do what they're told in the short term, but doesn't work in the long term and may in fact be harmful.
Instead, many psychologists recommend time-outs and other types of non-physical punishment. If that doesn't work, Graham-Bermann said a parent might want to wait until his or her anger has blown over before talking to the child about the problem.
Despite the opinion of professional societies, surveys show that as many as 90% of parents spank their children. Taylor encourages parents to talk to a paediatrician about how to better control their toddlers if they use this type of punishment.
"Children need guidance and discipline," said Taylor. "However, parents should focus on positive, non-physical forms of discipline and avoid the use of spanking."