Prevalence of Spanking in US National Samples of 35-Year-Old Parents From 1993 to 2017

Abstract

This survey study examines the prevalence of spanking among US parents aged 35 years with children aged 2 to 12 years from 1993 to 2017.

Spanking has been the subject of considerable research and discussion in popular culture. Reviews and state-of-the-art analyses support an association between corporal punishment (including spanking) and negative outcomes for children.¹ Professional organizations (eg, American Academy of Pediatrics) have recently issued statements conveying unequivocal opposition to corporal punishment and support for alternative means of discipline.²

Sporadically, studies have presented prevalence estimates of corporal punishment from one-time national samples^3,4; together, they suggest a decrease over recent decades. However, to our knowledge, no repeated surveys have documented trends in the prevalence of spanking in the US.

Methods

We used national panel data from the Monitoring the Future (MTF) study on 25 consecutive cohorts (graduating high school seniors in 1976-2000) assessed 17 years later (data at modal age 35 years collected from 1993-2017).⁵ This study was approved by the University of Michigan institutional review board and informed consent (either passive consent or active [ie, written] consent, per school policy) was obtained at baseline from parents for students younger than 18 years and from students 18 years or older. Each year, MTF surveys a nationally representative sample of 12th graders (modal age 18 years). A subsample is randomly selected, oversampling for those with substance use, and periodically surveyed through adulthood; 29 604 participants (48.5%) of those eligible were retained at the modal age 35 years follow-up survey. The analytic sample included those reporting at least 1 child age 2 to 12 years living at home part-time or full-time (including biological, adopted, or stepchildren); after excluding 11 777 participants (39.8%) without children aged 2 to 12 years and 1437 (4.9%) with missing data on child age or spanking, the analytic sample included 16 390 participants (Table). Parents in the analytic sample reported a mean (SD) of 1.8 (0.96) children aged 2 to 12 years (range, 1.7-1.9 mean number of children). The mean (SD) age of respondents’ youngest child aged 2 to 12 years was 5.2 (3.3) years, which decreased over time, ranging from 5.8 (1993) to 4.8 (2013) years.

Table. Descriptive Characteristics of 16 390 Study Participants.

Characteristic	No. unweighted (% weighted)a
Marital status
Married or engaged	14 273 (86.2)
Separated or divorced	1331 (8.3)
Single or widowed	702 (5.5)
Sex
Women	9981 (56.4)
Men	6409 (43.6)
Race/ethnicityb
White	13 953 (76.2)
Black	941 (11.3)
Hispanic	669 (6.9)
Asian	202 (1.7)
Other	488 (3.9)
Completed bachelor’s degree or higher
No	9146 (56.8)
Yes	7044 (41.7)

^aTotals may not equal 16 390 because of missing data.

^bRespondents were asked to select 1 of 9 race/ethnicity descriptions, which were collapsed in the Table.

Respondents with children were asked, “How often do you spank your child(ren)?” Seven response options ranged from never or almost never to every day; responses were dichotomized to represent 0 (never/almost never) vs 1 (more often [ie, several times or more during the past year]).

We calculated prevalence of spanking for each cohort at modal age 35 years. We used JoinPoint (National Cancer Institute) to analyze the trend (ie, whether the slope significantly differed from 0) and the best-fitting model in terms of the number of changes in the trend (joinpoints). Results were considered statistically significant at P < .05. We constructed weights to adjust for attrition (calculated as the inverse of the probability of participating at modal age 35 years based on age 18 years demographic characteristics) and an oversampling of those with substance use.⁶

Results

In a sample of parents at modal age 35 years with children aged 2 to 12 years, spanking declined over 25 years. The modeled prevalence of spanking decreased (slope = −0.005; SE = 0.001; P < .001) from 50% in 1993 to 35% in 2017 (Figure). A zero-joinpoint model (ie, a single decreasing trend) provided the best fit.

Figure. Trend in Prevalence of Spanking Within the Past Year Among US Parents at Modal Age 35 Years From 1993 to 2017.

In post hoc analyses, the modeled prevalence for men decreased from 52% to 36% and for women from 48% to 35%. The modeled prevalence among a subset of parents with a child aged 2 to 4 years living at home (8203 [50.0%]) decreased from 60% to 39%.

Discussion

To our knowledge, this study is the first to demonstrate a decreasing trend in US spanking prevalence in a repeated survey. The results correspond with single-year prevalence rates in other studies.^3,4 Spanking has been reported to be highest with children age 2 to 4 years⁴; this study’s results show a decrease across cohorts despite parents having slightly younger children in later cohorts. Spanking decreased among mothers and fathers from 1993 to 2017.

This study makes a substantial contribution by describing spanking epidemiology over 25 consecutive cohorts of age-homogenous parents. The data reflect parents’ use of spanking rather than whether a specific child was spanked. The primary limitation of these data was the single-item measure of spanking with limited response options. Although a downward trend was observed, there is a clear need for ongoing education about alternative discipline strategies.