Normative Trends in Physically Aggressive Behavior: Age-Aggression Curves from 6 to 24 Months

Abstract

Objective: To investigate age-related trends in physically aggressive behaviors in children before age 2 years.

Study design: A normative US sample of 477 mothers of 6- to 24-month-old children reported on the frequency of 9 interpersonally directed aggressive child behaviors, and hurting animals, in the past month.

Results: Almost all (94%) of the children were reported to have engaged in physically aggressive behavior in the past month. Based on 2-part regression models, the prevalences of kicking (OR, 1.70; P = .023), pushing (OR, 3.22; P < .001), and swiping (OR, 1.78; P = .018) increased with years of age, but the prevalence of hair pulling decreased with age (OR, 0.55; P = .020). The prevalences of hitting and throwing increased initially, then plateaued at age 18-20 months, and then decreased (quadratic aOR, 0.13 and 0.16; P < .001 and .010, respectively). The frequencies of hitting (R² = .05; P < .001) and throwing (R² = .03; P = .030) increased, and the frequencies of hair pulling (R² = .07; P < .001) and scratching (R² = .02; P = .042) decreased with age (P values adjusted for false discovery rate).

Conclusions: Physically aggressive behavior in the 6- to 24-month age range appears to be nearly ubiquitous. Most, but not all, forms of physical aggression increase with age. These results can guide pediatricians as they educate and counsel parents about their child's behavior in the first 2 years of life.

Keywords: aggression; child behavior; child development; infancy; toddlerhood.

Figure 1.

Statistically Significant Age Trends for Individual Aggressive Behaviors. Predicted probabilities (top panel) and mean frequencies (bottom panel) for behaviors with statistically significant age trends; means restricted to cases with non-zero values. Prevalence based on behavior present in the past month; frequency $\overline{\mathrm{x}}$, SD, minimum, and maximum based on non-zero values only; frequency response choices are 0 (Never), 1 (Rarely [less than once a week]), 2 (Some [1–3] days of the week), 3 (Most [4–6] days of the week), 4 (Every day of the week), and 5 (Many times each day). Because a natural log transformation is applied to continuous values in two-part regression models, plotted predicted values for the continuous variables were transformed into their original scaling by multiplying the exponentiated predicted values by the Duan smearing factor.