Spanking infants and toddlers: maternal belief and practice.

OBJECTIVE

To describe maternal beliefs and practices of spanking infants and toddlers and the relations between factors affecting these beliefs and practices.

DESIGN

Cross-sectional survey.

SETTING

Site 1 was an inner-city teaching hospital pediatric clinic. Site 2 was a private pediatrician's office in a nearby suburban neighborhood.

PARTICIPANTS

Mothers of children less than 4 years old in the waiting area. Site 1: n = 104; site 2: n = 100.

SELECTION

Systematic sample of convenience.

METHODS

Mothers were interviewed using a 20-minute structured questionnaire. Measures were constructed to assess beliefs (Cronbach's alpha = .90) and practices about spanking and approach to discipline (alpha > .71).

RESULTS

Belief in spanking correlated significantly (P < .001) with belief in negative approach (r = .52) and with the practice (r = .46) and severity (r = .34) of spanking. Nineteen percent of the mothers believed that there are times when it is appropriate to spank a child less than 1 year old, and 74% believed this about children 1 to 3 years old. Forty-two percent reported that they had spanked their own child in the past week. Mothers believed more strongly in spanking for dangerous misbehaviors than for annoying ones (P < .001). Belief in spanking and negative approach were stronger for older toddlers (P < .001). Belief in and frequency of spanking were more common at the inner-city site (P < .001).

CONCLUSIONS

The belief in and practice of spanking are prevalent in disciplining very young children. The context of the spanking affects beliefs and practices. The finding that belief and practice of spanking are highly correlated suggests that belief rather than impulse largely explains spanking of children less than 4 years old. The high correlation between spanking and negative approach toward discipline raises questions about whether negative consequences of spanking are the result of spanking per se, the negative approach toward the child, or both.