Factors that affect parental disciplinary practices of children aged 12 to 19 months.

Background: Physicians are encouraged to provide counseling regarding parental discipline. Quality counseling requires knowledge of disciplinary practices and factors that affect these practices.

Methods: One hundred and eighty two parents of 12- to 19-monthold children from general pediatric clinics in North Carolina and Alabama were interviewed regarding discipline using the Discipline Survey. Measures of contextual factors were analyzed to see which predicted disciplinary practices.

Results: Ninety-two percent of the participants were mothers; 6% were fathers; participation rate: 78%. Monitoring was the most common type of discipline used and time out was the least common. Parent, child, and family characteristics were all importantly associated with a broad array of disciplinary practices and modes of administration. However, the situation in which discipline occurred was found to be significant for most disciplinary practices even after controlling for other factors. Our study found that the specific misbehavior was most likely, and the presence of the other parent was least likely, to affect the type of discipline which was utilized.

Conclusions: When counseling families about discipline, practitioners should incorporate the fact that misbehavior happens in various contexts.

Key Words: discipline, corporal punishment, child rearing, socialization 

**********

Disciplinary practices are an important part of parenting, and can affect the child in regards to self-esteem, (1) conscience development, (1-3) aggression, (1,4,5) behavioral problems, (6,7) delinquency, (4,7,8) adult criminal behavior, (4) depression, (9) and alcoholism. (9) Unfortunately, studies regarding discipline have focused on a limited number of disciplinary techniques, and little information has been gathered in regards to administration and context in which discipline occurs. Corporal punishment has been studied extensively, but much less is known about other types of discipline. In addition, the frequency and variety of different types of discipline, such as positive/negative demeanor, consistency, and follow-through, have not been studied.

It has become increasingly clear that the context in which discipline is administered is an important factor in understanding parental discipline. Some studies have shown that the context in which the discipline occurs actually determines the effect of the discipline on the child. (10,11) Contextual factors that may affect disciplinary practices include child, parent, family/community, and situational characteristics (Fig. 1). Previous research has usually studied contextual factors in isolation, so little is known about the relative effects of various contextual factors on discipline practices. What is known is summarized below.

[FIGURE 1 OMITTED]

Child characteristics

Age and gender (1,7,12) have been studied frequently as predictors of spanking. Rates of spanking generally peak for children 2 to 3 years of age. (13) Boys are spanked and subjected to verbal aggression more often than girls. (5,14,15) Aggression has been studied most frequently. (8,16-19) It is examined both as a precursor to, and an outcome of, corporal punishment; and a positive association is often found between aggression and corporal punishment, although whether corporal punishment results in aggression may depend on the age and race of the child. (10,11)

Parental characteristics

Studies have shown that older parents are less likely to use corporal punishment; (5,13) and mothers spank more often than fathers (5) although this may be an effect of the greater amount of time that mothers spend with children. There is conflicting evidence regarding disciplinary styles in black families, but a fair amount of evidence suggests that they tend to be more power-assertive and punitive controlling for the socioeconomic status. (20-25) The issue of race in discipline has been particularly controversial and is a focus of this study, controlling for other factors.

In some studies, parental discipline beliefs correlate strongly with discipline practices, (26) and yet other studies reveal that a belief in the efficacy of hitting did not correlate with actual practices. (27) Abusive parents may either have limited knowledge of child development or may not know how to apply that knowledge to childrearing. (28) However, the relationship of knowledge of child development to nonabusive styles of discipline is not known. Mothers with more education are more likely to use teaching and verbal assertion as opposed to moderate or severe physical force than mothers with less education. (29) Maternal depression is associated with negative and intrusive parent-child interactions. (30)

Family characteristics

Lower socioeconomic status has been associated with more frequent corporal punishment, but parental age and ethnic status may be confounders. (5) Several studies have shown larger families are more likely to use power-assertive methods. (1,29,30) The presence of a father in the family also affects the use of setting limits. (29) Single mothers may be less effective at monitoring their children. (31) There are a number of questions about family composition and its relationship to discipline that have not been resolved, and so marital status is given particular attention in the current study.

Immediate situation

Previously, we found that the beliefs of parents and physicians in regards to types of discipline depend upon whether the misbehavior is dangerous, or merely annoying. (26) Other studies (32) have also found differences depending on the kind of transgression.

In this study we describe frequencies for a broad array of discipline types and modes of administration. While the context in which discipline occurs has been studied, previous research has mostly focused on limited demographic factors and corporal punishment. In this paper we report on a wider, richer array of contextual factors studied collectively in models rather than just as isolated factors. In addition, a new disciplinary measure (33) is used that was designed to evaluate a broad array of types of discipline, including mode of administration, and contextual factors in relation to disciplinary practices, not just corporal punishment. We hypothesized that even after accounting for child, parental, and family characteristics, the situation in which the discipline occurred would still be an important factor in determining disciplinary practices.

Methods

Participants

Parents involved in the current study were all part of a larger study--the Healthy Steps for Young Children study conducted by Johns Hopkins University (JHU). (34) These parents were consecutively enrolled in the Healthy Steps Study in the general pediatric clinics at the University of North Carolina (UNC) and the University of Alabama (UAB). They were approached in the pediatric clinic during health care visits when their child was 12 to 19 months old, and invited to complete the Discipline Survey as a cross-sectional sample. Families were eligible for the current study if they had not officially withdrawn from the Healthy Steps study and the child had at least one healthcare visit between 6 and 18 months of age. By 12 months of age, when this study began, 155 families were part of the Healthy Steps study at UNC and 131 were eligible at UAB.

Measures

Dependent variables.

The outcome variables for the current study were the types of discipline and the modes of discipline administration as measured by the Discipline Survey. (35,33) The Discipline Survey was designed to be applied to reactive discipline--in other words, what the parent did after a child misbehaved. Items for each subscale of the Discipline Survey were summed and divided by the number of items in the subscale. Subscales measured:

1. Type of discipline: verbal communication (3 items -- talking, explaining, telling), corporal punishment (2 items -- spanking, slapping), monitoring (2 items -- watching, keeping an eye on child), modeling (3 items -- demonstrating good behavior, showing how to behave, providing examples), and ignoring (3 items -- acting as if there was no misbehavior, ignoring, withdrawing attention); and

2. The mode of administration: positive demeanor (3 items -- being loving, respectful, encouraging), negative demeanor (3 items -- being angry, mean, and not calm), consistency (3 items -- child knows what to expect, same discipline is given for the same misbehavior, there is parental advanced planning for misbehavior), and follow through (3 items -- parent does not back down, does not let child misbehave anyway, follows through).

In addition, single items measured distraction, reward, removing privileges, time-out and natural consequences as types of discipline (Fig. 2).

Independent variables.

As part of the larger study conducted by JHU, (34) parents were interviewed in person within one month of the birth of the child and by telephone at 3 to 4 months of age. Data for contextual measures were linked to enable the investigation of factors affecting parental discipline. Specifically, data were available concerning:

Child characteristics. Infant temperament with three subscales (adapted from the National Longitudinal Survey of Youth (36)) including irritable baby (3 items -- how often is baby fussy/irritable/hard to soothe), active baby (2 items -- how often does baby squirm/move), and smiley baby (2 items -- how often does baby smile/laugh).

Parental characteristics. Age, race, education, knowledge of infant development (KIDI (37,38) a 30-item scale about parenting and knowledge of infant development and capabilities), depression (CES-D (39) Center for Epidemiologic Studies Depression Scale a 14-item scale pertaining to symptoms of depression), sense of competence (40) (a 17-item scale pertaining to feelings of competence in parents).

Family characteristics. Income, family stress (Hassles, (41) a 10-item scale about stressors such as work, money, violence, being overwhelmed), marital status, and number of people in the household.

The research reported here was conducted independently of the national evaluation of the Healthy Steps for Young Children Program.

Immediate situation

Family, parent, and child factors have been measured previously, but these are not existing measures or the immediate situation in which the discipline occurred. Thus, 14 items were developed to assess aspects of the situational context that are either known or hypothesized to affect discipline, including the type of misbehavior, temporary life stressors that might affect parent behavior, parent-child interaction at the time of the discipline, and where the discipline took place and who was present. These items were administered at the time of the Discipline Survey (children 12-19 months old). The items had a 6-point Likert response set from 6 = "very much" to 1 = "not at all" (Fig. 2).

The study was approved by the UNC and UAB institutional review boards, and signed consent was sought from each parent to complete the Discipline Survey as well as for permission to transfer data collected by the JHU evaluation.

Data analysis

Immediate situation factors.

To determine factors for immediate situation, the items were entered into a principal components analysis and subscales were created based on factor loadings. The number of factors was determined using a combination of eigenvalues greater than one and the location of the "elbow" of the eigenvalue graph. Items with factor loadings of [greater than or equal to] 0.5 with no cross loadings of [greater than or equal to] 0.3 were grouped as factors. Internal reliability analysis was then done to maximize internal reliability. Items from each factor were summed and divided by the number of items in the factor to create scales (Table 1).

Regression modeling.

Previous models about the relationship between the context of misbehavior, type of discipline, mode of administration, and outcomes from parental discipline, (33,35) were adapted using the data available in this study to guide the statistical analysis. The purpose of the statistical analysis was to find factors to explain the observed sample variation for each Discipline Survey subscale (ie, for each type of discipline and the mode of administration). Thus, analysis was geared toward the bolded arrows in Figure 1. The challenge of the analysis was to develop valid models given our modest sample size and dozens of correlated potential predictors. To this end, we first classified each potential predictor into domains as a "parent," "child," "family," or "situation" predictor. Then, to reduce the number of predictors within each category, we selected an optimal subset of variables using multiple [R.sup.2] and Mallow's Cp statistics as selection criteria. (42) The optimal subsets differ among the subscales, but race and marital status were of key interest and hence, were always included. Our purpose in this step was to capture the bulk of the sample variance explained by each domain (ie, "parent," "child," "family," "situation"). Respondents were included in regression analysis only if there was no missing data for all variables examined.

We used the reduced domains obtained by the above analyses to fit a hierarchical regression model for each subscale. After including "site" as a blocking variable, we added variables from "parent," "child," "family," and "situation" domains sequentially into the model and tested domain effects using Type 1 tests. Our reasoning was that (1) situation factors were of interest only after accounting for differences among parent, child, and family factors; (2) family effects should be adjusted for differences in child and parent characteristics; and (3) child factors were adjusted for parent characteristics. Thus, models were created such that for model I where I = 1[right arrow]4. Model 1 = site; Model 2 = Model 1 + parent variables; Model 3 = Model 2 + child variables; Model 4 = Model 3 + family variables; Final Model = model 4 + situation variables.

Finally, we wanted to be able to say something about individual characteristics that affect discipline, rather than just the domain groups of variables such as "child," "parent," "family," and "situation." So, for the most robust models where the multiple [R.sup.2] exceeded 0.25, we identified the individual predictors within each domain from the final models.

Results

Participation and descriptive characteristics

Of the 286 eligible families, we were unable to contact 52, leaving 254 families that were contacted. One hundred eighty-two parents were interviewed, for a participation rate of 78%. Reasons for the additional 52 not completing the discipline survey were: 31 were contacted but never completed a survey and 21 refused. Almost all of the respondents were mothers, about half were black, and about half had completed some education beyond high school (Table 2). Data were available from JHU for 177 of the children as newborns and 167 of the families for the phone interview at 3 to 4 months of age. Twenty-two percent of respondents were clinically depressed as measured by modified CES-D, using a standard cut point. The vast majority of parents underestimated their baby's developmental abilities as measured by the KIDI (Table 2).

Discipline frequencies

Monitoring was the most common type of discipline and time-out was the least common for these 12- to 19-month-old children. Positive demeanor was the most common mode of administration and belittling was the least common (Table 3). For individual items, parental demeanor was most often self described as respectful and least often as mean. When used, the average length of time-out was 1.4 minutes (range: 1-3 minutes); average number of spanks was 1.6 (range: 1-4); average time spent talking 2 minutes (range: < 1 minute to > 1 hour). The most common privilege taken away was a toy. The most common time for which a privilege was taken away was more than 5 minutes but less than one hour.

Immediate situation description

Of the 14 items regarding the situation and its effect on parental discipline, parents reported that the items measuring what the child did to misbehave (mean Likert score = 4.1) most affected the way they disciplined, whereas whether the other parent was there (mean Likert score = 1.9) least affected the way parents disciplined. The situation factors that resulted from principal component analysis retained 11 of the items and included: the type of misbehavior, temporary stressors, the location of the discipline, and parent-child interaction (Table 1). The situation factor that affected parental discipline the most was type of misbehavior (mean Likert score = 4.0), and the situation factor that affected their discipline the least was temporary stressors (mean Likert score = 2.3).

Factors affecting disciplinary practice

Results from regression analysis indicate that (Table 4 and Table 5) parent, child, family, and site were each important for predicting disciplinary subscales. These results are summarized below.

Influence of parental characteristics

For the 16 disciplinary subscales/items examined as outcomes, parent characteristics were significantly associated with all but six. Where examined, individual predictors showed that more knowledge of infant development (KIDI) was positively associated with consistency. Younger and white mothers were associated with more modeling.

Influence of child characteristics

After adjusting for differences among parent characteristics, child characteristics were only significantly associated with two disciplinary subscales/items: negative demeanor and modeling. A more "smiley" temperament was associated with more consistency and modeling. Negative demeanor increased with increasing child's age.

Influence of family characteristics

After controlling for parent and child differences, family characteristics were only significantly associated with one disciplinary subscale: corporal punishment. Interestingly, long-term stressors (eg, worries about food, shelter, money, work) were positively associated with use of corporal punishment, while short-term stressors (eg, feeling tired or having a bad day at the time the discipline occurred) were negatively associated with use of corporal punishment.

Influence of situational characteristics

Even after accounting for differences in parent, child, and family characteristics, the situation was an important predictor in 7 of the disciplinary subscale models. Among the models with at least 25% explained variance, modeling was the only subscale where situation was not important. The type of misbehavior was found to affect the use of consistency and corporal punishment.

Differences between sites

The only discipline practice for which site remained significant after controlling for the other variables in the model was corporal punishment. Parents in Alabama report more use of corporal punishment than do parents in North Carolina in our sample of one-year olds. In bivariate analysis, 71% of parents in North Carolina and 23% of parents in Alabama reported they had never spanked the child over the previous three months. Of those who spanked, the average number of times the child was spanked for each misbehavior was 1.3 in North Carolina and 1.7 in Alabama.

In contrast to some previous evidence (1,5,26) income, education, maternal depression, child's gender, marital status, and the number of people in the household were not found to be predictive of any disciplinary subscale after controlling for other variables.

Discussion

This study affirms, more comprehensively than in the past, that context matters when describing a broad array of disciplinary practices. The parent, the child, the family and the situation are all important in the type of discipline that is used and how it is administered, but parental characteristics and the immediate situation were especially strong determinants. Our hypothesis that the situation in which the discipline occurred would be an important predictor of disciplinary practices even after controlling for other contextual factors was confirmed for most disciplinary practices. This study suggests that the situation in which the discipline happened was a significant determinant for the majority of types of discipline and modes of administration used by parents and that the situational factors have not received enough attention in the empirical literature to date.

The findings related to individual variables should be considered preliminary due to the small sample size, but are nonetheless interesting. Some of these results confirm other research. For example, in a previous study (26) mothers of 1 to 3 year olds were more often found to believe in a negative approach more commonly for older children, and this practice is confirmed in the current study, even after controlling for other factors. The fact that corporal punishment was more common in Alabama than North Carolina may confirm earlier research that showed corporal punishment to be most common in the deep south. (43)

Some findings from previous research are not confirmed. For example, we did not find that younger parental age or family size was predictive of corporal punishment either before or after controlling for other variables. However, two recent large studies have found either no association between maternal age and corporal punishment (44) or an unexpected positive correlation. (27) Perhaps this represents a new trend away from the previously accepted correlation between young mothers and the use of corporal punishment. In addition, marital status, income and education were not predictive of any subscales, and race was only significantly predictive for the modeling subscale. Though not addressing discipline practice directly, some previous work suggests that depressed mothers have inaccurate expectations of child development and more negative or intrusive parent-child interactions. (45, 46) In this study, depression was not related to any disciplinary practice after controlling for other variables that were retained in final models.

Other findings are new. For example, knowledge of infant development was found to be a significant predictor of consistency. Previous evidence suggests that lack of knowledge of infant development may be a predictor of harsh and abusive discipline, (47) and the present study provides new evidence that knowledge of infant development is associated with other aspects of discipline as well. This is particularly exciting in that knowledge of infant development may be addressed through intervention.

Some findings confirm previous research but raise new questions. For example, the finding that long- and short-term stressors behave oppositely in their association with corporal punishment. Previous research has briefly explored the association between stressful life events, unemployment, poverty, and social isolation and child maltreatment as well as some types of discipline. When an association has been found it has been a positive one between these long-term stressors and child maltreatment. (28) The relationship between short-term stressors and child maltreatment or discipline has not been previously explored. Our finding that long-term stressors are positively associated with corporal punishment is consistent with previous literature. The fact that we found short-term stressors to be negatively correlated with corporal punishment is a new unexpected finding and needs to be explored further.

Other findings are notable. Even though the age range of the children involved in the study was 7 months, and these children were very young, there was a significant increase in negative demeanor of parents over this age range. It will be useful to learn the reasons for, and implications of, this increased negative demeanor over time in parents disciplining very young children. Understanding more about a negative approach toward children and how it evolves may be especially important in understanding discipline and intervening in families with child maltreatment. In this sample of very young children (12-19 months), 57% of parents reported using corporal punishment, and 11% said they did this often, almost always, or always. This is disturbing. In addition, for children of 12 to 19 months, only 41% of parents ever used time out. It would be helpful to learn more about the reasons for this limited use of time out.

This study has several limitations. First, the sample size is small for analyzing the number of independent variables of interest. With these analyses, very strong predictors of a given dependent variable can be found but the procedures used will not identify all of the important predictors. Second, disciplinary practices were assessed by self report which may lack accuracy, due to parental desire to respond in a socially acceptable manner. We need to confirm the validity of this self-report survey with observational data, as is planned. Third, 92% of participants in this study were mothers. It would be useful to include more fathers and other major caregivers in the home and compare the disciplinary practices of mothers and others in future research. Fourth, most of the parent and family characteristics were measured at 3 to 4 months of age and some may have changed by 12 to 19 months. Finally, the Discipline Survey did not measure discipline in a specifically identified discipline encounter. The Discipline Survey was designed to measure actual parent behavior and so hypothetical situations were not employed. Thus the Discipline Survey does not allow comparison among parents for given defined problem behaviors.

In previous research concerning the context in which discipline occurs, the situation in which the discipline occurred has largely been ignored. It is heartening to know that the parents in this study understand that the situation makes a difference in the type of discipline used and the mode of administration.

As parents seek to understand how to discipline children, it is important to include the context in the discussion. It doesn't make sense to address how to best discipline children in any circumstance, but rather how to discipline for a given child, parent, family, and situation. Current recommendations for effective discipline (48) do not adequately include discussion of the context in which discipline occurs and particularly not the immediate situation such as type of misbehavior, parentchild interaction, location of misbehavior, and temporary stressors. Many aspects of the context cannot be changed through intervention--eg, child/parent age--but some aspects might be amenable to intervention such as knowledge of infant development or a better understanding of how the situation should affect disciplinary practice. As more is learned about the effects of types of discipline and modes of administration, it may be possible to affect parental discipline practices, and even to prevent physical abuse, by addressing some aspects of the context (eg, knowledge of infant development) in which discipline is practiced.

Conclusion

When counseling families about discipline it is useful to know common practices and to incorporate counseling that acknowledges that misbehavior happens in various situations. Monitoring and verbal communication were the most common types of discipline for these very young children. Ignoring and time out were the least common types of discipline. The context in which the discipline occurred-including parent, child, and family characteristics and the immediate situation-affected the type of discipline used and the way it was administered. The situation in which the discipline occurred was still an important determinant of discipline after taking the other factors into account.

Acknowledgments

This study was carried out with funding from the Ambulatory Pediatric Association Special Project Award, The Duke Endowment, and the Commonwealth Fund. The Commonwealth Fund, a New York-based private independent foundation, undertakes independent research on health and social issues. The views presented here are those of the authors and not necessarily those of The Commonwealth Fund, its directors, officers, or staff. The authors acknowledge the contribution of the Department of Population and Family Health Sciences at Johns Hopkins University (JHU), the site of national evaluation for Healthy Steps, for providing the design and instruments for the overall evaluation. This project was carried out under a collaborative agreement with JHU. Special thanks also to Sandy Fuller, Carla Fenson, Karen Wysocki, Joy Stewart, Leslie Spooner, and Leslie Harrington, and to all the parents who participated in the study.

References

1. Becker WC. Consequences of different kinds of parental discipline. In: Hoffman ML and Hoffman LW, eds. Review of Child Development Research, New York, Russell Sage Foundation, 1964, pp 169-208.

2. Steinmetz SK. Disciplinary techniques and their relationship to aggressiveness, dependency, and conscience. In Contemporary Theories about the Family. New York, Macmillan Publishing Co., 1979, pp 405-438.

3. Kochanska G, Tjebkes TL, Forman DR. Children's emerging regulation of conduct: restraint, compliance, and internalization form infancy to the second year. Child Dev 1998;69:1378-1389.

4. McCord J. Some childrearing antecedents of criminal behavior in adult men. J Pers Soc Psychol 1979;37:1477-86.

5. Straus MA, Donnelly DA. Beating the Devil Out of Them: Corporal Punishment in American Families. New York, Lexington Books, 1994.

6. Forehand RL, Mcmahon RJ. Helping the Noncompliant Child: A Clinician's Guide to Parent Training. New York, The Guilford Press, 1981.

7. Patterson GR, Reid JB, Dishion TJ. Antisocial Boys. Eugene, OR: Castalia, 1992.

8. Sampson RJ. The family context of juvenile delinquency. In Sampson RJ and Laub JH, eds: Crime in the Making: Pathways and Turning Points Through Life. Cambridge, Harvard University Press, 1993.

9. Holmes SJ, Robins LN. The role of parental disciplinary practices in the development of depression and alcoholism. Psychiatry 1988;51:24-35.

10. Gunnoe ML, Mariner CL. Toward a developmental contextual model of the effects of parental spanking on children's aggression. Arch Pediatr Adolesc Med 1997;151:768-775.

11. Deater-Deckard K, Dodge KA, Bates JE, et al. Physical discipline among African American and European American mothers: links to children's externalizing behaviors. Dev Psychol 1996;32:1065-1072.

12. Wauchope BA, Straus MA. Physical punishment and physical abuse of American children: incidence rates by age, gender and occupational class. In Straus MA, Gelles RJ, eds: Physical Violence in American Families. New Brunswick, Transaction Publishers, 1990.

13. Socolar RRS, Stein REK. Spanking infants and toddlers: maternal belief and practice. Pediatrics 1995;95:105-111.

14. Vissing YM, Straus MA, Gelles RJ, et al. Verbal aggression by parents and psychosocial problems of children. Child Abuse Negl 1991;15:223-238.

15. Wissow LS. Ethnicity, income, and parenting contexts of physical punishment in a national sample of families with young children. Child Maltreat 2001;6:118-129.

16. McCord J. Parental Behavior in the Cycle of Aggression. Psychiatry 1988;51:14-23.

17. Gershoff ET. Corporal punishment by parents and associated child behaviors and experiences: a meta-analytic and theoretical review. Psychol Bull 2002;128:539-579.

18. Benjet C, Kazdin AE. Spanking children: the controversies, findings, and new directions. Clin Psychol Rev 2003;23:197-224.

19. Weiss B, Dodge KA, Bates JE, et al. Some consequences of early harsh discipline: child aggression and a maladaptive social informational processing style. Child Dev 1992;63:1321-1335.

20. Reis J, Barbera-Stein L, Bennett S. Ecological determinants of parenting. Fam Relat 1986;35:547-554.

21. Kelley ML, Power TG, Wimbush DD. Determinants of disciplinary practices in low-income black mothers. Child Dev 1992;63:573-582.

22. Regalado M, Sareen H, Inkelas M, et al. Parents' discipline of young children: results from the National Survey of Early Childhood Health. Pediatrics 2004;113 (Suppl 6):1952-1958.

23. Patterson G, DeBarsyshe B, Ramsey E. A developmental perspective on antisocial behavior. Am Psychol 1989;44:329-335.

24. McLoyd VC, Jayaratne TE, Ceballo R, et al. Unemployment and work interruption among African American single mothers: effects on parenting and adolescent socioemotional functioning. Child Dev 1994;65:562-589.

25. McLoyd VC. The impact of economic hardship on black families and children: psychological distress, parenting, and socioemotional development. Child Dev 1990;61:311-346.

26. Socolar RRS, Stein REK. Maternal discipline of young children: context, belief, and practice. J Dev Behav Pediatr 1996;17:1-8.

27. Gallup Organization. Disciplining Children in America. Princeton, Gallup Organization, 1995.

28. Panel of Research on Child Abuse and Neglect of the National Research Council. Understanding Child Abuse and Neglect. Washington, D.C., National Academy Press, 1993.

29. Socolar RRS, Winsor J, Hunter WM, et al. Maternal disciplinary practices in an at-risk population. Arch Pediatr Adolesc Med 1999;153:927-934.

30. Eamon MK. Antecedents and socioemotional consequences of physical punishment on children in two-parent families. Child Abuse Negl 2001;25:787-802.

31. Florsheim P, Tolan P, Gorman-Smith D. Family relationships, parenting practices, the availability of male family members, and the behavior of inner-city boys in single-mother and two-parent families. Child Dev 1998;69:1437-1447.

32. Catron TF. Masters JC. Mothers' and children's conceptualizations of corporal punishment. Child Dev 1993;64:1815-1828.

33. Socolar RRS, Savage E, Devellis RF, et al. The Discipline Survey: a new measure of parental discipline. Ambul Pediatr 2004;4:166-173.

34. Minkovitz CS, Hughart N, Strobino D, et al. A practice-based intervention to enhance quality of care in the first 3 years of life: the Healthy Steps for Young Children Program. JAMA 2003;290:3081-3091.

35. Socolar RRS. A classification scheme for discipline: type, mode of administration, context. Aggress Violent Behav 1997;2:355-364.

36. Rothbart MK. Measurement of temperament in infancy. Child Dev 1981;52:569-578.

37. MacPhee D. Knowledge of Infant Development Inventory. [Unpublished questionnaire and manual.] University of N. Carolina, Department of Psychology, Chapel Hill, NC, 1981.

38. Conrad B, Gross D, Fogg L, et al. Maternal confidence, knowledge, and quality of mother-toddler interactions: a preliminary study. Infant Ment Health J 1992;13:353-362.

39. Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Applied Psychological Measurement 1977;3:385-401.

40. Griband-Wallston J, Wandersman L. Development and Utility of the Parenting Sense of Competence Scale. Presented at American Psychological Association meeting, Toronto, 1978.

41. Kanner AD, Coyne JC, Schaefer C, et al. Comparison of two modes of stress measurement: daily hassles and uplift versus major life events. J Behav Med 1981;4:1-39.

42. Kleibaum David G, Kupper Lawrence L, Muller Keith E. Chapter 16: Selecting the best regression equation. In: Applied Regression Analysis and Other Multivariable Methods, 2nd Ed. PW-KENT Publishing Company, 1988, pp 318-324.

43. Straus MA, Mathur AK. Social change trends in approval of corporal punishment by parents from 1968 to 1994. In Freshee, D. Horn W, Bussman K, eds: Violence Against Children. Berlin and New York, Walter de Gruyter, 1996, pp 91-105.

44. Straus MA, Stewart JH. Corporal punishment by American parents: national data on prevalence, chronicity, severity, and duration, in relation to child and family characteristics. Clin Child Fam Psychol Rev 1999;2:55-70.

45. Reis J. Correlates of depression according to maternal age. J Genet Psychol 1988;149:535-545.

46. Cummings E, Davies P. Maternal depression and child development. J Child Psychol Psychiatry 1994;35:73-112.

47. Erlanger HS. Social class and corporal punishment in child rearing: a reassessment. Amer Sociolog Review 1974;39:68-85.

48. Committee on Psychosocial Aspects of Child and Family Health of the American Academy of Pediatrics. Guidance for effective discipline. Pediatrics 1998;101:723-728.

We consider blessed those who have persevered ... The Lord is full of
compassion and mercy.
--The Bible, James 5:11 NIV

Rebecca R. S. Socolar, MD, MPH, Eric Savage, MA, Lynette Keyes-Elstein, DRPH, MPH, Hughes Evans, MD, PHD

From the Division of General Pediatrics and Adolescent Medicine, Department of Pediatrics and Department of Social Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC; Design and Statistical Computing Unit, Frank Porter Graham Child Development Center, University of North Carolina at Chapel Hill, Chapel Hill, NC; and the Division of General Pediatrics and Adolescent Medicine, Department of Pediatrics, University of Alabama at Birmingham, Birmingham, AL This study was carried out with funding from the Ambulatory Pediatric Association Special Project Award, The Duke Endowment, and the Commonwealth Fund.

Reprint requests to Rebecca R. S. Socolar, MD, MPH, C.B. #7220, General Pediatrics, University of North Carolina, Chapel Hill, NC 27599-7225. Email: [e-mail]

Accepted September 9, 2005.

RELATED ARTICLE: Key Points

* Monitoring and verbal communication were the most common types of discipline in this age group. Ignoring the child and time outs were the least common types of discipline.

* The context in which the discipline occurred, including parent, child, family characteristics and the immediate situation, affected the type of discipline used and the way it was administered.

* The situation in which the discipline occurred was still an important determinant of discipline after taking other factors into account.

* When counseling families about discipline it is useful to be aware of common practices and to acknowledge that misbehavior occurs in various situations.

SAMPLE OF ITEMS FROM DISCIPLINE MEASURE

All parents handle their problems with their children's behavior in
different ways. We want to learn about how you have handled problem
behavior in the past three months with NAME. We want you to think about
both what you did when your child misbehaved and also how you got your
child to behave in a way that will help him/her get along in the world.
You may notice that some questions are similar. This helps us to learn
more accurately what you do when you discipline your child.

For all of these questions, we want you to think about how you have
handled problems with NAME in the past 3 months.

Examples of:    In the past three months when you had a problem with the
                way NAME behaved,

TYPE OF
DISCIPLINE
Verbal          did you talk with NAME  never  rarely  sometimes  often
communication   about the problem?        1      2         3        4
Ignoring        did you ignore the      never  rarely  sometimes  often
                behavior?                 1      2         3        4

MODE OF
ADMINISTRATION
Negative        would you say you       never  rarely  sometimes  often
demeanor        were angry?               1      2         3        4
Positive        were you respectful     never  rarely  sometimes  often
demeanor        of NAME?                  1      2         3        4

TYPE OF
DISCIPLINE
Verbal          did you talk with NAME  almost always  always
communication   about the problem?            5          6
Ignoring        did you ignore the      almost always  always
                behavior?                     5          6

MODE OF
ADMINISTRATION
Negative         would you say you      almost always  always
demeanor         were angry?                  5          6
Positive         were you respectful    almost always  always
demeanor         of NAME?                     5          6

SAMPLE ITEMS TO MEASURE IMMEDIATE SITUATION

Many things can affect the way a parent disciplines his/her child. Think
about the things that have affected the way you disciplined NAME in the
past 3 months. How much would you say each of the following affected the
way you disciplined NAME?

The way I disciplined NAME depended on:

Whether I was having a bad day
Very much  Quite a bit  Some  Not really  Not much  Not at all
    6           5        4        3          2          1

The particular misbehavior that occurred
Very much  Quite a bit  Some  Not really  Not much  Not at all
    6           5        4        3          2          1

Fig. 2 Sample of items from discipline measure.

Table 1. Immediate situation: Items and factors from principal
components analysis

Factor (a)                              Mean (b)    SD

Type of Misbehavior                       4.0       1.4
  What the child did to misbehave         4.1       1.6
  Whether the child had just been         3.8       1.5
    disciplined for the same
    misbehavior
Parent-child interaction                  3.1       1.1
  The way child reacted when              3.7       1.5
    disciplined
  How much child had already              3.1       1.6
    misbehaved that day
  Whether parent had much time at the     3.0       1.5
    moment
  Whether parent was feeling supported    2.7       1.7
Location                                  2.5       1.5
  Whether parent and child were in        2.7       1.6
    public
  Whether anyone else was around          2.4       1.4
Temporary stressors                       2.3       1.1
  Whether parent was tired                2.5       1.4
  Whether parent was under a lot of       2.3       1.4
    stress that day
  Whether parent was having a bad day     2.2       1.2

Factor (a)                               N    Cronbach's [alpha]

Type of Misbehavior                     182         0.77
  What the child did to misbehave
  Whether the child had just been
    disciplined for the same
    misbehavior
Parent-child interaction                180         0.71
  The way child reacted when
    disciplined
  How much child had already
    misbehaved that day
  Whether parent had much time at the
    moment
  Whether parent was feeling supported
Location                                182         0.71
  Whether parent and child were in
    public
  Whether anyone else was around
Temporary stressors                     182         0.74
  Whether parent was tired
  Whether parent was under a lot of
    stress that day
  Whether parent was having a bad day

(a) One item -- whether the other parent was present--did not have
factor loadings sufficient to include it with any factor.
(b) Six-point Likert scale: 1 = not at all, 6 = very much.

Table 2. Demographics/description of contextual variables (a)

Parental characteristics
  Respondent's relationship to child (N = 182) (b)
    Mothers                                         92
    Fathers                                          6
    Other                                            2
  Mean maternal age, years (N = 177) (c)            26.0 (14-41)
  Mean paternal age, years (N = 164) (c)            28.3 (17-55)
  Maternal race (N = 175) (c)
    White                                           36
    Black                                           54
    Other                                           10
  Paternal race (N = 177) (c)
    White                                           33
    Black                                           53
    Other                                           14
  Maternal education (N = 177) (c)
    < High school graduate                          14
    High school graduate/GED                        33
    Some college                                    32
    [greater than or equal to] College graduate     21
  Depression (N = 167) (d)
    > Cut-off for CES-D                             22
  Knowledge of infant development (N = 167) (d,e)
    < 33% correct                                    8
    34-< 50% correct                                28
    50-< 70% correct                                62
    [greater than or equal to] 70% correct           2
  Sense of competence (N = 156) (d,f)               50.0 (39-65)
Child characteristics
  Mean child's age, months (N = 182) (b)            16.1 (12-19)
  Child's sex: (N = 182) (b)
    Boys                                            55
  Infant temperament (N = 166) (d,g)
    Irritable scale                                  6.3 (3-12)
    Active scale                                     4.9 (2-10)
Family characteristics
  Household income (N = 167) (d)
    < $5000                                         10
    $5000-9999                                      11
    $10,000-29,999                                  41
    $30,000-74,999                                  27
    [greater than or equal to] $75,000               7
    Don't know/refused                               4
  Family stress (N = 166) (d,h)
    Hassles scale                                   33.5 (19-40)
  Maternal marital status (N = 167) (d)
    Married                                         47
    Separated/divorced                               9
    Living with partner                              8
    Never married                                   36
  Number of people in household (N = 167) (d)        2
    2                                                5
    3-5                                             85
    6-8                                             10

(a) Data are given as mean (range) or percentage unless indicated
otherwise.
(b) Measured when child 12-19 months of age.
(c) Measured at birth of child.
(d) Measured when child 3-4 months of age.
(e) 98% of respondents underestimated development for
[greater than or equal to] 24% of items and 5% of respondents
overestimated development for [greater than or equal to] 24% of items.
(f) Note: higher score indicates less competence.
(g) Note: higher score indicates more of characteristic.
(h) Note: higher score indicates less hassles.
CES-D. Center for Epidemiologic Studies Depression Scale.

Table 3. Disciplinary subscales: Descriptive statistics and internal
reliability analysis for subscales

                                                        % Reporting
                             Mean Likert             use often, almost
                            score (range)     SD       always, always

Types of discipline
  Monitor                     5.10 (1-6)     1.04            87
  Verbal communication        4.73 (1-6)     1.08            76
  Model                       4.47 (1-6)     1.22            69
  Distract (a)                4.40 (1-6)     1.34            70
  Reward (a)                  4.07 (1-6)     1.64            62
  Natural consequences (a)    2.27 (1-6)     1.26            12
  Remove privilege (a)        2.25 (1-6)     1.36            17
  Corporal punishment         2.19 (1-6)     1.15            11
  Ignore                      1.97 (1-4)     0.83             5
  Time-out (a)                1.76 (1-6)     1.10             7
Modes of administration
  Positive demeanor           5.40 (2-6)     0.72            92
  Follow-through              4.98 (1-6)     0.91            87
  Consistency                 4.18 (1-6)     1.08            63
  Negative demeanor           1.99 (1-4)     0.70            10

                            No.   Cronbach's [alpha]

Types of discipline
  Monitor                   182         0.74
  Verbal communication      182         0.76
  Model                     181         0.82
  Distract (a)              182
  Reward (a)                182
  Natural consequences (a)  181
  Remove privilege (a)      182
  Corporal punishment       182         0.78
  Ignore                    182         0.64
  Time-out (a)              181
Modes of administration
  Positive demeanor         181         0.59
  Follow-through            181         0.71
  Consistency               179         0.57
  Negative demeanor         180         0.56

(a) These are single items. Responses for subscales were averaged across
the items in the subscale.
Note: Single items had response sets from never = 1 to always = 6.

Table 4. Hierarchical regression analysis with multiple [R.sup.2]'s of
factors affecting disciplinary practices (a)

                                          Hierarchical
                                    Model 1          Model 2
                                    [R.sub.1.sup.2]  [R.sub.2.sup.2]
Disciplinary practices              Site             Parent

Corporal punishment (b) (N = 152)   0.14*            0.19
Consistency (b) (N = 145)           0.07             0.22***
Negative (b) (N = 152)              0.00             0.06
Model (b) (N = 147)                 0.00             0.18***
Ignore (b) (N = 151)                0.05             0.13*
Verbal communication (b) (N = 145)  0.00             0.12***
Distract (c) (N = 147)              0.02             0.10*
Follow through (b) (N = 156)        0.00             0.10*
Natural consequences (N = 154)      0.01             0.06
Monitor (b) (N = 147)               0.00             0.09
Reward (c) (N = 161)                0.02             0.11*
Time-out (c) (N = 144)              0.02             0.07
Remove privileges (c) (N = 156)     0.01             0.09*
Positive (b) (N = 145)              0.00             0.07

                                    Model 3          Model 4
                                    [R.sub.3.sup.2]  [R.sub.4.sup.2]

Disciplinary practices              Child            Family

Corporal punishment (b) (N = 152)   0.21             0.27*
Consistency (b) (N = 145)           0.25             0.26
Negative (b) (N = 152)              0.13***          0.17
Model (b) (N = 147)                 0.23**           0.26
Ignore (b) (N = 151)                0.14             0.18
Verbal communication (b) (N = 145)  0.16             0.19
Distract (c) (N = 147)              0.15             0.19
Follow through (b) (N = 156)        0.12             0.15
Natural consequences (N = 154)      0.09             0.11
Monitor (b) (N = 147)               0.12             0.14
Reward (c) (N = 161)                0.13             0.15
Time-out (c) (N = 144)              0.08             0.12
Remove privileges (c) (N = 156)     0.12             0.15
Positive (b) (N = 145)              0.10             0.11

                                    Final model
                                    Total [R.sup.2]
Disciplinary practices              Situation

Corporal punishment (b) (N = 152)   0.33*
Consistency (b) (N = 145)           0.32**
Negative (b) (N = 152)              0.29***
Model (b) (N = 147)                 0.27
Ignore (b) (N = 151)                0.21
Verbal communication (b) (N = 145)  0.20
Distract (c) (N = 147)              0.20
Follow through (b) (N = 156)        0.19*
Natural consequences (N = 154)      0.16*
Monitor (b) (N = 147)               0.19*
Reward (c) (N = 161)                0.19*
Time-out (c) (N = 144)              0.18*
Remove privileges (c) (N = 156)     0.17
Positive (b) (N = 145)              0.14

[R.sub.2.sup.2]; for model i where i = 1[right arrow]4.
(a) Model 1 = site; Model 2 = Model 1 + parent variables; Model 3 =
Model 2 + child variables; Model 4 = Model 3 + family variables; Final
Model = Nodel 4 + situation variables.
(b) A scale.
(c) A single item.
Significance testing for F tests of the added group of predictors:
* < 0.05, ** < 0.01, *** < 0.005.

Table 5. Non-hierarchical regression analysis of individual variables
affecting disciplinary practices (a)

                        Parent                 Child
                     parameter est.        parameter est.
Model                (Std. [beta])         (Std. [beta])

Consistency          KIDI underestimation  Smiley baby
                     -8.4 (-0.17)*         0.30 (0.17)*
Corporal punishment
Negative                                   Child's age
                                           0.21 (0.18)*
Model                Maternal age          Smiley baby
                     -0.18 (-0.30)***
                     Race (c)              0.39 (0.20)**
                     -1.9 (-0.27)*

                        Family         Situation           Site
                     parameter est.  parameter est.    parameter est.
Model                (Std. [beta])   (Std. [beta])     (Std. [beta])

Consistency                          Type misbehavior
                                     0.24 (0.21)*
                                     Parent-child
                                     interaction
                                     -0.12 (-0.22)*
Corporal punishment  Hassles (b)     Type misbehavior  Site
                     -0.08 (-0.18)*  0.13 (0.16)*      0.90 (0.19)*
                                     Parent-child
                                     interaction
                                     0.07 (0.17)*
                                     Temp stressors
                                     -0.15 (-0.23)**
Negative                             Temp stressors
                                     0.13 (0.22)*
Model

Only variables significant at 0.05 or less shown.
Significance testing probability > |T| * < 0.05, ** < 0.01, *** < 0.005.
(a) est., established; Std., standard; KIDI, knowledge of infant
development.
(b) Note: higher score indicates less hassles.
(c) White = 1; Black = 2.