American Psychological Association (APA) - Discipline ...in this relation across ethnicity (e.g.,...

Journal of Family Psychology2000, Vol. 14, No. 3, 380-400

Copyright 2000 by the American Psychological Association, Inc.0893-3200/00/$5.00 DOI: 10.1037//0893-3200.14.3.380

Discipline Responses: Influences of Parents'Socioeconomic Status, Ethnicity, Beliefs About

Parenting, Stress, and Cognitive-Emotional Processes

Ellen E. PinderhughesVanderbilt University

John E. BatesIndiana University

Kenneth A. DodgeDuke University

Gregory S. PettitAuburn University

Arnaldo ZelliDuke University

Direct and indirect precursors to parents' harsh discipline responses to hypotheticalvignettes about child misbehavior were studied with data from 978 parents (59%mothers; 82% European American and 16% African American) of 585kindergarten-aged children. SEM analyses showed that parents' beliefs aboutspanking and child aggression and family stress mediated a negative relationbetween socioeconomic status and discipline. In turn, perception of the child andcognitive-emotional processes (hostile attributions, emotional upset, worry aboutchild's future, available alternative disciplinary strategies, and available preventivestrategies) mediated the effect of stress on discipline. Similar relations betweenethnicity and discipline were found (African Americans reported harsher disci-pline), especially among low-income parents. Societally based experiences maylead some parents to rely on accessible and coherent goals In their discipline,whereas others are more reactive.

Parent discipline practices are integral in the-ories of children's socialization. Parents' use ofphysical punishment with their child is of spe-cial interest. Numerous theories posit a role forphysical punishment in the development of an-tisocial behavior in children. According to oneset of theories, discipline responses are made inthe context of multiple influences ranging from

more distal factors such as culture, ethnicity,and socioeconomic status (SES) to more prox-imal factors such as available social supports,family structure, and family processes (Belsky,1984; Luster & Okagaki, 1993; Rubin, Stewart,& Chen, 1995). Indeed, SES and ethnic dif-ferences have been found consistently inphysical punishment (e.g., Deater-Deckard,

Ellen E. Pinderhughes, Department of Psychologyand Human Development, Peabody College, Vander-bilt University; Kenneth A. Dodge and Arnaldo Zelli,Sanford Institute of Public Policy, Duke University;John E. Bates, Department of Psychology, IndianaUniversity; Gregory S. Pettit, Department of Psy-chology, Auburn University,

This research was supported by National Instituteof Mental Health Grant MH42498, and National In-stitute of Child Health and Human DevelopmentGrant HD30572 and a William T. Grant faculty

scholar award. We are grateful for the participation ofnumerous parents, children, and teachers. Specialappreciation for assistance with this article is ex-tended to Robert D. Laird, Robert Nix, and LauraGriner Hill.

Correspondence concerning this article should beaddressed to Ellen E. Pinderhughes, Department ofPsychology and Human Development, Peabody Col-lege, Vanderbilt University, Nashville, Tennessee37203. Electronic mail may be sent to ellen.e.pinderhughes @ vanderbilt.edu.

380

SPECIAL ISSUE: SES, ETHNICITY, AND DISCIPLINE 381

Dodge, Bates, & Pettit, 1996; Dodge, Pettit, &Bates, 1994; Luster, Rhoades, & Haas, 1989;McLoyd, 1990). More proximal influences suchas stress also have been linked to punitive par-enting (e.g., McLoyd, Jayaratne, Ceballo, &Borquez, 1994). Another set of theories sug-gests that the most proximal influence on disci-pline responses can be found in parents'cognitive-emotional processes regarding situ-ationally based child misbehavior (e.g., Dix,1993). This study drew from these two sets ofliterature to examine relations among contex-tual influences, cognitive-emotional processes,and parents' use of physical or harsh punish-ment. Consistent with Luster & Okagaki(1993), who have suggested several patternsthrough which contextual influences may affectdiscipline responses, this study examined directand mediated relations among SES, ethnicity,gender, stressors, parents' perception of thechild, parenting beliefs, cognitive-emotionalprocesses, and the use of physical and harshpunishment (hereafter referred to as disciplineresponses). Two models of direct and mediatedrelations between SES and discipline responsesand between ethnicity and discipline responseswere tested. The theoretical and empirical liter-atures supporting these models are discussednext

Direct and Mediated Relations BetweenSES and Parenting

Figure 1 presents a model of direct and me-diated relations between SES and discipline re-sponses. SES has been found consistently tohave negative relations with physical and harshpunishment. Although there may be differencesin this relation across ethnicity (e.g., Deater-Deckard et al., 1996), this relation is hypothe-sized to hold within ethnicity.

Parenting Values and Beliefs

As Figure 1 depicts, one theorized pathwaylinking SES and parenting practices is parentingbeliefs (Kohn, 1963; Okagaki & Divecha,1993). With a sample of European Americanmothers, Luster et al. (1989) found that SESwas negatively related to values embracing con-formity and positively related to values endors-ing self-direction. In turn, the effect of theseparental values on parenting behaviors was me-diated by parenting beliefs about discipline,promoting exploratory behavior, and affection-ate, responsive behavior. This leads to the ques-tion of whether the relation between SES andparenting behavior could be mediated by par-enting beliefs. Several studies illustrate the link

Cognltlve-Ea»tion»lProccsMg

Figure 2. Hypothesized model of relations among socioeconomic status (SES), parenting beliefs, stress,perception of child, cognitive-emotional processes, and discipline responses.

382 PINDERHUGHES, DODGE, BATES, PETTIT, AND ZELLI

between parenting beliefs and attitudes and par-enting behavior (e.g., Dix, 1993; Luster et al.,1989). Luster and Kain (1987) found differ-ences between parents who believed that par-enting positively affects child outcomes (highefficacy) and parents who believed that parent-ing has little effect on child outcomes (lowefficacy). Parents high in efficacy endorsedlove, affection, and modeling as critical influ-ences, whereas parents low in efficacy, espe-cially fathers, tended to endorse discipline.

Stressors, Parent Perception of the Child,and Parent Cognitive-EmotionalProcesses

Another pathway through which SES is hy-pothesized to influence differences in disciplineresponses is stress: Economic hardship exposeslow-SES parents to additional stressors that un-dermine their ability to use inductive disciplinestrategies and that result in higher parental re-liance on punitive discipline (see McLoyd,1990, for a review of the literature on theserelations). Parental stress has been found to beassociated positively with punitive parentingpractices (e.g., Crnic & Greenberg, 1987;McLoyd et al., 1994; Patterson, 1986). LowSES may operate through the following specificstressors that are associated with punitive andunsupportive practices: being a single parent(e.g., Fox, Platz, & Bentley, 1995), having alarge number of children (e.g., Fox et al., 1995),having an unplanned pregnancy (e.g., Zuravin,1987), and living in an unsafe neighborhood(e.g., Abell, Clawson, Washington, Bost, &Vaughn, 1996).

Parents' Perception of the Child

Although external stressors are hypothesizedto have a direct link to parents' discipline re-sponses, their influence also may be mediatedby parents' perception of the child and parents'cognitive-emotional processes when faced withchild misbehavior. External stressors have beenfound to predict parents' negative views of thechild (Conger, McCarty, Yang, Lahey, &Kropp, 1984). And parents' perception of childfunctioning has been linked to parenting behav-iors, particularly among parents of aggressivechildren (Rubin et al., 1995). Aggressive behav-ior by children in preschool and early elemen-tary school tends to evoke negative parent emo-tions and cognitions, which lead to more

negative parenting behaviors (Rubin & Mills,1992; Rubin et al., 1995).

Parents' Cognitive-Emotional Processes

McLoyd (1990) noted that stressful life con-ditions facing low-income parents underminetheir emotional state. Dix (1991) has suggestedthat high levels of stressors negatively affectparents' cognitive-emotional processes. Sev-eral links have been found between cognitive-emotional processes and parents' discipline re-sponses. First, it has been shown (e.g., Dix,Ruble, & Zambarano, 1989; MacKinnon-Lewis,Lamb, Arbuckle, Baradaran, & Volling, 1992;Strassberg, 1995) that parents' tendency tomake hostile attributions about the child corre-lates with punitive parenting. Second, intensenegative affect about child misbehavior may berelated to use or endorsement of forceful disci-pline (Dix, 1993; Dix & Lochman, 1990).Third, parental worry about the future implica-tions of current child misbehavior may affectdiscipline responses. This process may be inten-sified when important socialization goals areinvolved (Dix, 1991, 1993). For example, aparent who values obedience is more likely tobecome upset over his or her child's defiancethan is a parent who places less value on obe-dience. These different emotional reactions todefiance might lead to different discipline re-sponses. Fourth, parental perceived control overthe misbehavior is another process that mayaffect discipline responses (Dix, 1991). Parentswho perceive that they lack control over themisbehavior in question are more likely to havenegative emotions than those who believe theycan control the misbehavior (Bugental, Blue, &Lewis, 1990). In this study, perceived controlover the misbehavior was assessed in two ways:the degree to which parents perceived that theycould prevent the problem behavior in the futureand the variety of noncoercive alternative dis-ciplinary strategies they reported as being avail-able to them.

Direct and Mediated Relations BetweenEthnicity and Parents' Discipline

Responses

The direct and mediated relations betweenethnicity and discipline responses were theo-rized to be the same, with one key difference.SES was expected to moderate relations be-


tween ethnicity and stress. The literature sup-porting this model is summarized next.

Ethnicity and Parenting Beliefs

African American mothers have reportedgreater use of physical discipline than do Euro-pean American mothers (Deater-Deckard et al.,1996). African American parents also have dis-played more punitive attitudes toward their chil-dren (e.g., Reis, Barbera-Stein, & Bennett,1986). How parents think about physical orsevere discipline and its purpose as a socializa-tion strategy may differ for African Americansand European Americans (e.g., Deater-Deckard& Dodge, 1997; Garcia Coll, 1990; Jackson,1997; Kelley, Power, & Wimbush, 1992; Ogbu,1981). For example, ethnic differences havebeen found in parents' acceptance of spanking(Heffer & Kelley, 1987). Thus, this study rep-resents a further step in examining ethnic dif-ferences in parenting (Deater-Deckard et al.,1996; Pettit, Bates, & Dodge, 1997) by specif-ically assessing whether parenting beliefs me-diate relations between ethnicity and discipline.

Ethnicity, Stress, Parents' Perceptions ofthe Child and Cognitive-EmotionalProcesses

Researchers have noted the moderating roleof SES in relations between ethnicity and stres-sors (McLoyd, 1990; Myers & King, 1983). Intheir study of race, SES, and distress, Kesslerand Neighbors (1986) found that low-incomeAfrican Americans were particularly vulnerableto additional race-related stressors and con-straints and reported higher levels of stress thandid low-income European Americans. In con-trast, no differences emerged between stresslevels reported by middle-income AfricanAmericans and European Americans. Thesefindings were consistent across multiple epide-miological data sets. With higher levels ofstress, low-income African American parentsmay be likely to engage in more punitive disci-pline responses (e.g., McLoyd, 1990). In thestudy reported here, SES was hypothesized tomoderate relations between ethnicity and stres-sors. Low-income African American parentswere expected to report greater stress and toendorse more physical and harsh discipline re-sponses than their European American counter-parts. Among middle-income parents, no ethnicdifferences in level of stressors were expected,

and no differences were expected in relationsbetween level of stressors and discipline.

In summary, this study tested direct and in-direct contributions of SES to discipline re-sponses and direct and indirect contributions ofethnicity to discipline responses. The study, partof a longitudinal investigation of child develop-ment, used cross-sectional data. Parents' re-sponses to multiple hypothetical vignettes in-volving child misbehavior were the measure ofdiscipline responses. Parenting beliefs, familystress, parents* perception of the child, andcognitive-emotional processes were hypothe-sized to mediate relations between SES anddiscipline responses and relations between eth-nicity and discipline responses. Parenting be-liefs included beliefs about (a) the effectivenessof spanking and (b) the appropriateness of peer-directed aggression by the child. Stressors in-cluded marital status, number of children, hav-ing an unplanned pregnancy, living in an unsafeneighborhood, and conflict in romantic relation-ships. Parent perception of the child was as-sessed as the distinctiveness of the descriptionof the child and parental affect toward the child.Cognitive-emotional processes included hostileattribution of the child, upset affect about themisbehavior, concern about the future implica-tions of the misbehavior, available alternativediscipline responses, and preventive disciplineresponses.

This study included fathers as respondents.Fathers' voices are relatively rare in the litera-ture on parenting (Phares, 1996). Althoughmothers and fathers have been found to engagein similar levels of harsh discipline (e.g.,Feldman & Wentzel, 1990), research on rela-tions among contextual influences, cognitive-emotional processes, and discipline responseshas focused primarily on mothers. Thus, inclu-sion of fathers here has the potential to extendcurrent understanding of parents' discipline re-sponses. More specifically, this design enabledexamination of parent sex as a moderator ofrelations among SES, ethnicity, parenting be-liefs, stress, perceptions of the child, cognitive-emotional processes, and discipline responses,which Phares (1996) and Rubin et al. (1995)noted are important new avenues for study.

Method

Participants

Respondents in this study were participants in theChild Development Project (described in previous


reports; e.g., Dodge et al., 1994; Pettit et al., 1997).They were selected in two cohorts from Nashville,Tennessee (a midsized mid-South urban communitywith about 25% of the participants residing in feder-ally subsidized housing); Knoxville, Tennessee (anAppalachian rural and small urban community); andBloomington, Indiana (a small midwestern city, withmany of the participants selected from working-classbackgrounds, including some rural families of Appa-lachian descent). Families were recruited at the timeof the kindergarten preregistration and (in April orAugust preceding the September of matriculation)asked to participate in a longitudinal study of childdevelopment. Two cohorts of participants were re-cruited in successive years. About 70% of familiesagreed to participate. Because about 15% of kinder-gartners do not preregister, this proportion of thesample was recruited in August by mail, letter fromschool, or telephone. Of the 585 participating fami-lies, 52% had a male child, and 48% had a femalechild; R2% were White, 16% were African American,and 2% were from other ethnic backgrounds, mostlyAsian American. When the father lived in the samehome as the child, both parents were invited to par-ticipate. Of the 978 participating parents, 581 weremothers, including adoptive mothers and stepmothers(475 White, 95 African American, and 11 other), and397 were fathers, including adoptive fathers and step-fathers (358 White, 30 African American, and 9other). Of the 585 families, both parents participatedin 393 cases, mothers only participated in 188 cases,and fathers only participated in 4 cases.

Procedure

In the summer before—or, in some cases, soonafter—the child's matriculation in kindergarten, twotrained interviewers visited the mother, falher, andchild in (heir home for a 3-hr session. In almost allcases, at least one of the two visitors was of the sameethnicity as the family. One visitor obtained writteninformed consent and then privately interviewed onerandomly selected parent while the second parentcompleted a battery of written instruments. The othervisitor interviewed the child to collect informationnot relevant to this article. After about 90 min, thetwo parents switched roles. The parent interviewconsisted of a demographic survey, a history of thechild's development, and parenting divided into pastand current (past 12 months) eras. This interview wasfollowed by the presentation of five hypothetical par-enting situations designed to solicit parent disciplineresponses and cognitive-emotional processes. Awritten battery further assessed the parent's disci-pline responses and values and the child's behavior.1

The parent interviewer was trained over a 4-weekperiod that included reading a detailed proceduremanual, observing interviews by a master inter-viewer, and conducting interviews with a supervisorpresent. Interviewers were trained to a reliability of

.80 or higher (percentage agreement across all items,with the supervisor's scores as the criterion) beforethey began to collect data on their own. Independentreliability assessments were obtained by a trainedobserver who either accompanied the interviewerfor or listened to tapes of 56 randomly selectedinterviews.

Constructs and Measures

Demographic characteristics. Parent sex (1 =mother, 2 = father) and child sex (1 = boy, 2 = girl)were coded via interviewer observation. Ethnicity(1 = White, 2 — African American, 3 = other) wascoded via direct questions to the parent. During theinterview, the parent's and partner's education andemployment status were solicited. Family SES wascomputed with Hollingshead's (1979) four-factor in-dex of social status (based on parent and partnereducation and current occupation; sample M — 39.5,SD = 14). As recommended by Hollingshead, whenfathers did not reside in the home, the mother'sscores were double weighted. Parent education levelwas coded via Hollingshead's 7-point scale {M =4.7, SD ~ 1.3). Parent past-year employment statuswas coded, on a 3-point scale, as unemployed outsidethe home, employed part time, or employed full time(M = 2.5, SD = 0.8).

General parenting beliefs were assessed with twovariables. The parent completed a newly developed,15-item written instrument, the Culture Question-naire, in which she or he read a statement and re-sponded on a 7-point scale ranging from definitedisagreement to definite agreement. Six items wereused for this study. One item (reverse scored) con-stituted the parent's belief that spanking works("When my child does something wrong, talkingabout it with her/him helps more than spanking");M = 4.57, SD = 1.6). The mean of 5 items consti-tuted a measure of the parent's aggressive values forthe child (M = 2.57, SD = 1.0; a = .52): "I wouldwant my child to defend himselfyherself even if itmeant hitting another child"; "I wouldn't try to stopmy child's fighting because my child has to show thats/he can defend herseltfhimself'; "Sometimes a phys-ical fight might help my child have a better relation-ship with another child"; "If I found out my child hitanother child, I would be disappointed no matterwhat the reason" (reverse scored); and "I wouldn't

1 Interviewers read questionnaires to parents whohad difficulty reading items to ensure that parentsunderstood the questions. Interviewers also ratedlevel of literacy (degree of difficulty reading andunderstanding questionnaires). Correlations betweenthis literacy measure and parent reports on writtenmeasures ranged from .024 (ns) to .11 (p < .01). Weconcluded that literacy had a modest impact on writ-ten responses.


mind if my child got a reputation as the toughest kidin school."

Family stress was assessed with five variables.Parent marital status was coded on a 2-point scale asmarried (which included cohabitating with a partnerfor at least the past year) or single (M = 1.26, SD =0.4). The number of children living in the home wasrecorded directly (M = 2.20, SD = 1.0). During theinterview, the child's birth history was reviewed, andthe degree of planning of the pregnancy was dis-cussed. Unplanned pregnancy was coded on a 4-pointscale as unprepared (1), not planned but accepted(2), partially planned or under discussion (3), orplanned (4; M = 2.01, SD = 1.09). Reliabilities ofthe family stress measures, computed with the kappastatistic, ranged from .63 to .87.

During the interview, family stressors were dis-cussed in the context of influences on the child'sdevelopment. The parent was asked the following:"All families have conflicts, parents and kids. Whatkinds of conflicts, arguments, or violence was yourchild aware of during this past year (including shout-ing, physical fights, pushing, etc.)?" The interviewerprobed for descriptions of arguments and any in-volvement by outside agencies. Separate questionsasked about marital conflict, parent-child or siblingconflict, and neighborhood conflict. The parent wasencouraged to talk freely about the number, range,and severity of conflicts. On the basis of the parent'sdescription, the interviewer privately rated the degreeof partner conflict on a 6-point scale of increasingseverity (0 = no partner or no conflict, 1 = rarelyeven shouting, 2 — numerous mild verbal argumentsreported, 3 = numerous major verbal argumentsreported, 4 = mild physical fights or frequent majorverbal arguments reported, 5 = more than one majorphysical fight reported; M = 1.75, SD = 1.1). Inde-pendent rater agreement on this coding was com-puted by Pearson correlation to be .70 (p < .001).The interviewer also rated the child's exposure toviolence outside the parental dyad, including otherfamilial violence and neighborhood violence (1 =none, 2 = mild verbal arguments, 3 = major verbalarguments, 4 = mild physical fights or frequent ma-jor verbal arguments, 5 = more than one majorphysical fight; M - 1.60, SD - 0.9). Independentrater agreement (r) was .69 (p < .001).

Later during the interview, the interviewer asked anumber of questions about the parent's experience ofstress during the previous 12 months. Promptingfrom a list of 14 common stressful life events,the interviewer asked questions, including "Whatchanges or adjustments occurred in your family?""How did these changes affect your child?" "Howwere you doing during the past year?" and "Did youhave chances to get out and do the things you enjoy,with or without the family?" The interviewer thenprivately rated the parent's experience of stressors(stress rating) on a 5-point scale ranging from mini-mal challenges (1) to severe and/or frequent chal-

lenges (5; M = 2.62, SD = 1.1). Independent rateragreement was .75 (p < .001).

Perception of child. Two scores were derived tomeasure the parent's affect about the child and eval-uation of the child. At the beginning of the interview,the parent was asked to describe the child freely forup to 5 min without interruption. The interviewerthen rated the parent's positive affect toward thechild on a 3-point scale ranging from mostly negativeto mostly positive (M = 2.60, SD = 0.6). The Pearsoncorrelation coefficient for agreement was .56 (p <.01). The interviewer also rated the level of distinc-tive description of the child—specifically, the par-ent's ability to describe the child uniquely, with greatdetail and understanding—on a 3-point scale rangingfrom vague and indistinct to distinct and insightful(Af = 2.13, SD = 0.7), Reliability, computed with thePearson correlation coefficient, was .67 {p < .001).

Cognitive-emotional processes in parenting wereassessed with five variables. In the last part of theinterview, parents were presented with each of fivehypothetical vignettes designed to elicit cognitive-emotional processes in decision making in responseto child misbehavior. For each vignette, the parentwas asked to imagine that her or his child had en-gaged in a particular misbehavior in an ambiguouscontext For example, in one vignette, the parent wasasked to imagine the following: "You go to yourchild's school to pick him or her up. You see all thekindergarten children running to get into line. One ofthe other children runs hard and bumps into yourchild. The other kids laugh. Your child gets upset andpushes the other kid to the ground" (Story 2). Theother vignettes described the child calling the winnerof a race a bad name (Story 1), the child threateningpeers who had excluded the child from a game (Story3), the child excluding a cousin from a birthday partyinvitation list (Story 4), and the child teasing a peer(Story 5).

After each vignette, the parent first was asked whythe child might have acted in the manner described.As a means of measuring the parent's hostile attribu-tion to the child, responses were scored as 1 if theparent indicated that the child had acted with hostileintent or in a negative-trait-like manner (or, in 3% ofthe cases, if the parent could not generate an expla-nation) and 0 if the parent attributed the blame to anyother cause (including the demands of the situation,the child's temporary state, the child's misinterpreta-tion of peer behavior, or the peer's behavior) orreported that the behavior was not indicative of aproblem (across five vignettes, M = .24, SD ~ .21).

Second, the parent was asked to rate how she or hewould feel if the child had behaved this way (upsetaffect); responses ranged from very good and notupset (1) to very bad and upset (5; M — 3.99, SD =0.5). Next, the parent was asked what she or he woulddo if the child behaved this way. Free responses weresolicited that could include multiple behaviors. Thefull response was coded as 1 if it included physical


punishment and 0 if not (M = .06, SD = .13). Theresponse was separately coded as 1 if it includedreasoning, explanations, discussion, inductive rea-soning, or proactive guidance and 0 if not, as ameasure of an acceptable alternative to physical pun-ishment (M = .77, SD = .24). A parent's responsecould be coded as including both of these alternativesor neither alternative.

Third, as a measure of severity of discipline re-sponse, the parent was asked to rate whether she or hewould punish the child and, if so, how much. Re-sponse options were not at all (1), a little (2), mod-erately (3), somewhat (4), and very sternly (5; M =2.50, SD = 0.9).

Next, the parent was told "Children sometimes dothings that make their parents worry that they willhave some problems. Some things parents don'tworry about because they know their children willgrow out of it. If your child acted this way most of thetime, how worried would you be that your childwould grow up to have problems later on?" Parentsrated their worry about the child's future on a 5-pointscale ranging from not at all (1) to very worried (5;M - 3.43, SD - 0.9).

Finally, the parent was asked "What could you asa parent do to prevent your child from acting this wayin the first place?" The free response was coded as ameasure of available preventive strategies in the par-ent's repertoire, with 1 coded if the response includeda before-the-fact preventive strategy and 0 coded ifthe parent reported that the behavior could not beprevented or the parent could generate only an after-the-fact response (M = .09, SD = .16). Reliabilitywas assessed on this measure with a subset of fami-lies (« = 24; Pettit et al., 1997). The correlationbetween independent raters for the number of timesmothers identified a preventive strategy was .56. Be-cause interviewers were trained to a high level ofreliability (.80) before beginning to collect data, re-liability of coding on variables for which code op-tions (i.e., hostile attribution, physical punishment,and alternative punishments) were clear and unam-biguous was not specifically assessed.

Harsh discipline response styles were assessedwith four variables. In addition to the two scores forharsh discipline assessed in response to each of thefive hypothetical vignettes (physical punishment andseverity of discipline response), two other scoreswere computed for harsh discipline used by the par-ent. The first score was a rating made by the inter-viewer after asking the following questions: "Whatkinds of misbehavior over the past year did yourchild do that you had to deal with?" "What kinds ofthings did you have to do to deal with his/her mis-behavior?" "How often did you have to physicallypunish your child, such as spank, grab, or shake?"and "What was the most severe thing you had to doduring this period?" The interviewer then privatelyrated the parent's harshness of discipline on a 5-pointscale ranging from nonrestrictive, mostly positive

guidance (1) to severe, strict, often physical punish-ment (5; M = 2.60, SD = 0.9). Independent rateragreement on this rating was .73 (p < .001). (Be-cause this rating was made early in the interview, itwas not affected by parents* responses to the fivevignettes.)

The fourth measure of harsh discipline was ob-tained from parent self-reports on a revision of theConflict Tactics Scales (CTS; Straus, 1979). Cohort 1parents were asked to report the frequency withwhich they displayed aggressive behaviors towardtheir child for each of two eras of the child's life(birth to 4 years of age and the previous 12 months).Response options ranged from never (0) to more thanonce a month (5). A physical aggression toward childscale was derived as the mean of 10 items (a = .87;items were, for each of two eras, "threatened to hit orthrow something at your child"; "threw something atyour child"; "pushed, grabbed, or shoved your child";"hit or tried to hit your child"; and "hit or tried to hitwith something"). Straus and Gelles (1990) reported"strong evidence of construct validity" (p. 72) for thisscale. For parents in Cohort 2, the "threatened to hitor throw something at your child" item was dividedinto 2 items, and the six response options wererescaled to range from never to almost every day.Thus, 14 items were included in this measure. Tocontrol for these cohort differences, we standardizedscores within cohorts (M = 0.0, SD = 1).

Results

Preliminary Analyses: Cognitive-Emotional Processes and DisciplineResponses

Because scholars differ about whether rela-tions between cognitive-emotional processesand parenting behavior are best assessed withinsituation (e.g., Sigel, 1986) or across-situation(e.g., Goodnow, 1984) analyses were initiallyconducted separately for each of the five hypo-thetical situations. Regression analyses revealedcross-situation consistency in relations betweencognitive-emotional processes and physicalpunishment and severity of punishment. In ad-dition, analyses testing parent sex differences inthese relations showed that the processes con-tributing to discipline responses did not differappreciably for mothers and fathers. Giventhese findings, and to describe the process ofparenting decision making most efficiently, weaveraged scores across the five situations.

Multivariate analyses of covariance (withSES as the covariate) were conducted on thefive cognitive-emotional processing cross-situational variables (hostile attributions, upset


affect, worry about the future, alternative disci-pline strategies, and preventive discipline strat-egies) and two hypothetical discipline cross-situational variables (severity of punishmentand physical punishment); parent ethnicity, gen-der of parent, and gender of child were inde-pendent variables. The adjusted group meansare listed in Table 1. Parent ethnicity yielded asignificant main effect, F(7, 926) = 3.13, p <.001. Univariate analyses revealed a main effectdue to parent ethnicity on three variables. Rel-ative to African Americans, European Ameri-can parents made fewer hostile attributionsabout their child, F(l, 932) = 4.95, p < .03;were less worried about their child's future, F(l,932) = 8.45,/> < .001; and endorsed less severepunishment, F(l, 932) = 7.46, p < .001.

Analyses of covariance were conducted onthe two actual discipline variables (from theglobal interviewer rating and from the CTS),with parent ethnicity, parent sex, and child sexas independent variables. Table 1 also containsthese adjusted group means. Parent ethnicityyielded a main effect on both measures of actualdiscipline. Relative to African American par-ents, European American parents were rated byinterviewers as using less harsh punishment,F(l, 908) = 4.68, p < .05, and they reportedless harsh punishment on the CTS, F(l, 892) -5.72, p < .10. Child sex yielded a significanteffect for interviewer rating: Parents of boyswere rated as more harsh, F(l, 908) = 5.21, p<.05. In regard to CTS scores, a parent sex effect,F(l, 892) = 11.26, p < .001), indicated moreharsh punishment among mothers than fathers,and a Parent Sex X Ethnicity effect reflectedgreater differences between African Americanmothers and fathers, F(l, 892) = 5.72, p < .05.

Structural Equation Models

The hypotheses drawn from the theoreticalmodel were tested through structural equationmodeling (SEM) analyses performed withAMOS software (Arbuckle & Wothke, 1999).Broadly speaking, we evaluated the fit of twomodels positing that the effects of SES, ethnic-ity, and stress on discipline responses are me-diated by parenting beliefs, perception of thechild, and cognitive-emotional processing ofhypothetical social situations. More specifi-cally, a first model, diagramed in Figure 2,hypothesized that SES influenced discipline re-sponses both directly and indirectly. The indi-rect effect operated though parenting beliefs and

23

8

• • 4

« s

ffl

'-<M©©

•o r» «? Q -g- s r* * cis3!

OO©'©'©'©"©'©©'

©©•©©©©•©"©©

OO^HOOO-^©--

pppoppppp©d©'©'©b©d©'

©©©©©©©©©

1 I-

388 PINDERHUGHES, DODGE, BATES, PETTIT, AND ZELU

Cognitive-EmotiontlPlDCeSHB

Figure 2. Mediation model: Relations of socioeconomic status (SES) with discipline responses mediatedthrough parenting beliefs, stress, perception of the child, and cognitive-emotional processes. Latent pathcoefficients are presented in standardized units. The coefficients in parentheses are the latent "total effect"coefficients. Adjusted goodness-of-fit index = .95, ^(153, N = 723) = 416.15, p < .01. *p =£ .05.**p ^ .01.

stress. This model also hypothesized that stressinfluenced discipline responses both directlyand indirectly. The indirect effect operatedthrough parents' perception of the child andparents' cognitive-emotional processes. Here-after, we refer to this six-construct model as theSES-discipline responses model. The secondmodel, diagramed in Figure 3, was identical tothe SES-discipline responses model with theexception that an ethnicity dichotomy score(i.e., a score that compared African Ameri-can families with the remaining sample) re-placed the SES variable in the model. Hereafter,we refer to this six-construct model as theethnicity-discipline responses model.

Table 2 presents the intercorrelations amongall variables used in the SEM analyses. Inter-correlations among indicators of the same con-struct are italicized. In general, indicators ofconstructs were modestly but significantly cor-related. The measurement model did not varyacross the two models or SEM analyses. Thefactor loadings and error terms for the SES-discipline responses model are presented in theAppendix. Parent conflict, preventive strategies,and CTS aggression score were not used in theSEM analyses because they represented poormeasurement indicators of the three latent con-

structs of stress, cognitive-emotional pro-cesses, and discipline responses, respectively.

For each model, SEM analyses proceeded infour separate steps. In Step 1, we tested the fullmediation model (i.e., a model that included allhypothesized direct and indirect paths of influ-ence linking the six latent constructs).

In Step 2, we modified the full mediationmodel to estimate the total unmediated effect ofeither SES (in the SES-discipline responsesmodel) or ethnicity (in the ethnicity-disciplineresponses model) on discipline responses. Bynot estimating (i.e., constraining to zero) theindirect paths linking each potential mediator—parenting beliefs, stress, perception of the child,and cognitive-emotional processes—to disci-pline responses, we forced the total effect ofSES or ethnicity on discipline responses to op-erate through the direct path only. Because thislatter "total effect" model was nested within theprevious full mediation model, chi-square dif-ference tests could be conducted to test thebenefit (or advantage or value) of including theindirect, mediated relations. A significant in-crease in model chi-square and a substantialincrease in the estimate of the direct effect ofSES on discipline responses would provide ev-


Cognttive-EmotionatProcesses

Figure 3. Mediation model: Relations between ethnicity and discipline responses mediated through parent-ing beliefs, stress, perception of the child, and cognitive-emotional processes. Latent path coefficients arepresented in standardized units. The coefficients in parentheses are the latent "total effect" coefficients.Adjusted goodness-of-fit index = .95, )?(153t N = 723) = 402.89, /> < .01. *p £ .05. **p s .01.

idence of indirect relations as presented in thefull mediation model.

In Step 3, we followed a similar procedure toestimate the total effect of stress on disciplineresponses and to evaluate the merit of hypoth-esizing that effects of stress on discipline aremediated by perception of the child andcognitive-emotional processes. In particular,we constrained to zero the two paths Unking,perception of the child and cognitive-emotionalprocesses to discipline responses, estimated theresulting model chi-square, and compared it sta-tistically with the full mediation model chi-square from Step 1.

In Step 4, the full mediation model was testedsimultaneously among mothers and fathers (i.e.,an SEM multiple-group analysis) to evaluatewhether the relations posited in the model var-ied with—or were moderated by—parent sex.After estimating the paths among latent vari-ables in the two groups simultaneously, wetested the model again with the general con-straint that the estimates of these paths wereidentical across the two groups. If parent sexmoderated the relations hypothesized in the fullmediation model, the chi-square value of thelatter model with equal constraints would be

significantly larger than the previous modelwithout the constraints.

For the SES-discipline responses model,SEM analyses were performed with and withoutcontrol for the demographic variable of ethnic-ity. For the ethnicity-discipline responsesmodel, SEM analyses were performed with andwithout control for the demographic variable ofSES. On inspection, findings were very similaracross these analyses. For each model, we thuschose to report findings obtained when themodel included the demographic control. Miss-ing data were handled with listwise deletion.Because the parents (and not the child) were theunit of analysis, all of the analyses reportedhereafter were conducted with data from 723parents (423 mothers and 300 fathers).

SES-Discipline Responses Model

Figure 2 shows the latent path estimates ob-tained for the full mediation model of Step 1.Lower SES predicted harsher parental disci-pline, greater stress, and more negative parent-ing beliefs, and greater stress predicted lesspositive perception of the child and more in-tense cognitive-emotional processes. Also,

I

i

is1c

* * # # #* • • * #a1

**ON

1**<—<

**

1#&

1**H I

**o

1**001

#*•—1

**

C4

I

O1

*#—-*

o

1#

1

*1—1

**

©

1 1 1 1

I I

3

o

r**o

r r* #* *o *o

r*#

ss

SSS o

F5 _ ^H -H

00 JO I

I

I f 8I

I§

;VI


more intense cognitive-emotional processespredicted harsher parental discipline. This fullmediation model fit the data satisfactorily,X*(153) = 416.15, p < .01, ratio = 2.7),goodness-of-fit index (GFI) = .95.

More important, the SES effect on disciplineresponses was mediated by the intervening ef-fects of parenting beliefs, stress, perception ofthe child, and cognitive-emotional processes.When the mediating paths from these variableswere constrained in Step 2, there was a statisti-cally significant increase in the model chi-square, A^(4) = 176.01,/? < .01. The value ofthe path linking SES and discipline responseswas - .22 (p < .05) in the full mediation modelwhen indirect effects were estimated. This valueincreased to —.51 ip < .01), however, in thetotal effect model when all indirect effects wereconstrained to zero. Together, these findingsfrom Step 2 indicate that not allowing parentingbeliefs, stress, perception of the child, andcognitive-emotional processes to mediate therelation between SES and discipline responsesresulted in a significantly worse model chi-square. These variables accounted for nearly60% (i.e., 1 - |.22A51| = .57) of the total effectof SES on discipline responses.

Step 3 of the analyses indicated that the ef-fects of family stress on discipline responsesalso were mediated by perception of the childand cognitive-emotional processes. In particu-lar, when the two mediating paths leading todiscipline responses were constrained, we againobtained a statistically significant increase inthe model chi-square, A^(2) = 9.13, p < .05,and a substantial increase in the effect of stresson discipline responses (/3 = .23 vs. /3 = —.08;see Figure 2).

Step 4 of the analyses revealed that the rela-tions among SES, discipline responses, stress,parenting beliefs, perception of the child, andcognitive-emotional processes were not mod-erated by parent sex. That is, there was not astatistically significant difference in the chi-square values of the structural equation modelswith and without the equality constraints on allpaths between latent variables, A^(12) = 8.19,ns. Thus, there was no evidence that any rela-tion between two latent constructs was differentfor mothers or fathers.

Ethnicity-Discipline Responses Model

Figure 3 shows the latent path estimatesobtained for the full mediation ethnicity-

discipline responses model. Even though themultivariate analysis of covariance showed eth-nic differences for two of the discipline re-sponses indicators (severity of punishment andrating of harsh punishment), in this model theeffect was small. Ethnicity predicted harsherdiscipline responses and greater stress. In thisand subsequent analyses in which ethnicity hada significant effect, being African Americanwas linked to harsher discipline responses andgreater stress. Greater stress also predicted lesspositive perception of the child, and more in-tense cognitive-emotional processes predictedharsher discipline responses. This model pro-vided a satisfactory fit to the data, ^(153) =402.89, p < .01, ratio = 2.6, GFI = .95.

The ethnicity effect on discipline responseswas mediated by the intervening effects of par-enting beliefs, stress, perception of the child,and cognitive-emotional processes. When themediating paths from these variables were con-strained in Step 2, there was a statistically sig-nificant increase in the model chi-square,A^(4) = 173.5, p < .01. The value of the pathlinking ethnicity and discipline responses was.06 (ns) in the mil mediation model when indi-rect effects were estimated. This value in-creased to .12 (p < .05), however, in the totaleffect model when all indirect paths were con-strained to zero (see Figure 3). Together, thesefindings from Step 2 indicate that not allowingparenting beliefs, stress, perception of the child,and cognitive-emotional processes to mediatethe relation between ethnicity and disciplineresponses resulted in a significantly worsemodel chi-square. These variables accountedfor nearly 50% (i.e., 1 - |.06/.12| = .50) of thetotal effect of ethnicity on discipline responses.

Step 3 of the analyses indicated that the ef-fects of family stress on discipline responsesalso were mediated by perception of the childand cognitive-emotional processes. In particu-lar, when the two mediating paths leading todiscipline responses were constrained, we againobtained a statistically significant increase inthe model chi-square, A^(2) = 8.35, p < .05,and a substantial increase in the effect of stresson discipline responses (/3 = .11 vs. /3 = —.01;see Figure 3).

Step 4 of the analyses revealed that the rela-tions among ethnicity, discipline responses, par-enting beliefs, stress, perception of the child,and cognitive-emotional processes were mod-erated by parent sex. When the full model was


tested with and without the equality constrainton all paths between latent variables, the anal-ysis yielded a statistically significant difference,A^(12) = 26.35, p < .01. Two paths differed,those between ethnicity and stress (.45 formothers vs. .10 for fathers) and between stressand cognitive-emotional processes (.40 formothers vs. .08 for fathers). These two differ-ences suggested that, especially for mothers, (a)being African American was linked to higherlevels of stressors than was being EuropeanAmerican, and (b) having higher levels of stresswas linked to more intense cognitive-emotionalprocesses.

SES Moderation of the Relations AmongEthnicity, Stress, and DisciplineResponses

Finally, the hypothesis that SES moderatedthe relation between ethnicity and stress wasexplored. This analysis entailed splitting thesample into two groups using a median split onSES (Af = 42.5). Because the middle-incomeAfrican American sample was so small (n —21), an SEM analysis could not be used to

assess differences in this relation amongmiddle-income parents. Figure 4 shows the la-tent path estimates obtained for the full media-tion model among low-income parents (n =355). Ethnicity predicted harsher discipline re-sponses and greater stress, and greater stresspredicted a less positive perception of the child.Relations were greater for African Americanparents than for European American parents.Also, more intense cognitive-emotional pro-cesses predicted harsher discipline responses.This full mediation model fit the data satisfac-torily, ^(108) = 253.80, p < .01, ratio - 2.18,GFI = .93.

The ethnicity effect on discipline responseswas mediated by the intervening effects of par-enting beliefs, stress, perception of the child,and cognitive-emotional processes. When themediating paths from these variables were con-strained in Step 2, there was a statistically sig-nificant increase in the model's chi-square,A^(4) = 133.8, p < .01. The value of the pathlinking ethnicity and discipline responses was.13 (ns) in the full mediation model when indi-rect effects were estimated. This value in-creased to .23 (p < .01), however, in the total

Cognitlve-EmotiooalProcesses

Figure 4. Mediation model: Relations between ethnicity and discipline responses among low-income parentsmediated through parenting beliefs, stress, perception of the child, and cognitive-emotional processes. Latentpath coefficients are presented in standardized units. The coefficients in parentheses are the latent "total effect"coefficients. Adjusted goodness-of-fit index = .93, ^(108, N = 355) = 253.80, p < .01. *p ^ .05.**p < .01.


effect model when all indirect paths were con-strained to zero. Together, these findings fromStep 2 indicate that not allowing parenting be-liefs, stress, perception of the child, andcognitive-emotional processes to mediate therelation between ethnicity and discipline re-sponses resulted in a significantly worse modelchi-square. These variables accounted for morethan 40% of the total effect of ethnicity ondiscipline responses (i.e., 1 - |.13/.23| = .43).

Step 3 of the analyses indicated that the ef-fects of family stress on discipline responsesalso were mediated by perception of the childand cognitive-emotional processes. In particu-lar, when the two mediating paths leading toharsh discipline responses were constrained, weagain obtained a significant increase in the mod-el1 s chi-square, A;p(2) = 9.10, p < .05, and asubstantial increase in the effect of stress ondiscipline responses (|3 = .17 vs. |3 = —.12; seeFigure 4). Given sample limitations (n = 131fathers), the Step 4 analyses to test for gendermoderation were not run.

Because the SEM could not be run on themiddle-income sample, separate bivariate cor-relations between ethnicity and stress—the crit-ical path hypothesized to differ—were con-ducted for middle- and low-income samples.Comparison of differences between low-incomeand middle-income parents in these correlationsrevealed a statistically higher relation amonglow-income parents (r = .237) than amongmiddle-income parents (r = .096), z = 1.96.

Discussion

In this study, two models hypothesizing di-rect and mediated relations between SES anddiscipline responses and between ethnicity anddiscipline responses were tested. Findings ofSEM analyses provided some support for thesemodels. After discussing implications of thesefindings, we focus attention on specific con-structs within the models.

SES—Discipline Responses Model

Significant direct and mediated relationswere found between SES and discipline re-sponses. As hypothesized, low-income parentstended to endorse more harsh discipline re-sponses in part because they held stronger be-liefs about the value of spanking, and they ex-perienced higher levels of stress. In turn, theirhigher levels of stress were associated with

more negative perceptions of the child and moreintense cognitive-emotional processes. Thesefindings suggest that SES differences in disci-pline responses are due to differences in parent-ing beliefs and to more intense cognitive-emotional processes that are linked to higherlevels of stress. What remains unclear iswhether the two paths from these characteristicsare mutually distinct or co-occur within individ-uals. That is, do some low-income parents useharsh discipline responses primarily because oftheir parenting beliefs, whereas others use harshdiscipline responses primarily because their lev-els of stress and cognitive-emotional reactivityare high? Or do low-income parents use harshdiscipline responses because they both believein their effectiveness and have high levels ofstress and cognitive-emotional reactivity?Studies examining within-group differencesamong low-SES parents (e.g., Kelly et al.,1992) will shed further light on this question.

Ethnicity-Discipline Responses Model

In contrast, the ethnicity- discipline re-sponses model yielded significant but modestdirect and mediated effects. For example, thetotal effect of ethnicity (.12) accounted for lessthan 2% of the variance associated with disci-pline responses, whereas the total effect of SES(-.51) accounted for 26% of the variance asso-ciated with discipline responses. At first glance,these findings suggest that relations betweenethnicity and the specific discipline responsesassessed are minimal. However, Kessler andNeighbors (1986) noted that SES and ethnicityinteract to minimize the true effect of ethnicity,particularly when stress is involved. They sug-gested that separate analyses of mediated rela-tions between ethnicity and discipline wouldreveal a stronger effect among lower incomeparents than among middle-income parents.

Unfortunately, power limitations precludeduse of SEM to examine relations among theconstructs with middle-income parents. How-ever, the SEM analysis with low-income par-ents revealed significant relations between eth-nicity and discipline responses. Consistent withthe findings of Kessler and Neighbors (1986),African American parents reported higher lev-els of stress and harsher discipline.

McLoyd (1990) noted a double impact ofstressors associated with the combination oflow SES and minority group status: AfricanAmericans are subjected to heightened soci-


etally based stressors associated with being amember of a minority group. As a result, Afri-can American parents may experience more in-tense cognitive-emotional processes and re-spond more harshly. Moreover, relative tomajority-status youth, African American youthhistorically have been at risk for more harshinstitutional and societal treatment, despite sim-ilar levels of problem behavior (Snyder, 1996;Stehno, 1982). Given this history, AfricanAmerican parents must incorporate into theirsocialization practices strategies that will pre-pare their children to function in settings inwhich others may respond harshly (e.g., Peters,1985; Stevenson, 1994; Tatum, 1987). Such afuture orientation may predispose low-SES Af-rican American parents to experience more in-tense cognitive-emotional processes. Indeed,review of the adjusted means in Table 1 sug-gests a pattern among African American parentsto report greater worry about the future impli-cations of current misbehavior than EuropeanAmerican parents. This pattern is especiallystriking among mothers of girls. Future researchin this area should examine more closely whichcognitive-emotional processes are salient forAfrican American parents.

Exploration of SES differences in bivariatecorrelations between ethnicity and stress—thekey path hypothesized to differ—did indicate astatistically weaker effect among middle-SESparents than among low-SES parents. Amongmiddle-SES families, the stressors studied heredo not exert a differential impact due to ethnic-ity. Perhaps economic advantage, in limiting theimpact of stressors, attenuates ethnic differ-ences in discipline responses. If this is the case,any ethnic differences in parenting amongmiddle-SES parents may reflect the impact ofdistinct parenting beliefs. Alternatively, perhapsother stressors exert a stronger differential im-pact on ethnic differences in parenting amongmiddle-SES parents. McLoyd (1999) has notedthat, even among middle-SES individuals,meaningful economic disparities (e.g., job sta-bility and net financial assets on which to rely inthe event of job loss) exist between AfricanAmericans and European Americans. Furtherstudy of SES and ethnicity interactions withlarger samples of middle-SES African Ameri-can parents and more comprehensive conceptu-alization of stressors may clarify whether eco-nomic advantage serves as a buffer against the

impact of stress on parenting among AfricanAmericans.

Specificity of Models

Parent sex differences among relations inboth models were tested. The findings werediscrepant: Parent sex moderated only the eth-nicity model. Among mothers especially, Afri-can Americans reported higher levels of stres-sors than did European Americans, and higherlevels of stress were linked to more intensecognitive-emotional processes. One would ex-pect that, irrespective of race or SES, womenwould be more vulnerable than men to the stres-sors examined in this study. Several stressorsare likely to differentially affect women: un-planned pregnancy, number of children in thehome, and single-parent status in particular. Inaddition, these stressors occur with greaterprevalence among African American womenand are more likely to lead to an escalatingchain of subsequent stressors than for EuropeanAmerican women (National Center for Educa-tion Statistics, 1993; Ventura, Taffel, Mosher,Wilson, & Henshaw, 1995). Thus, these find-ings suggest that European American mothersmay have had more buffers to counteract expe-riencing the stressors or the impact of stressors.

Another relevant question is the extent towhich the models are age specific. Several is-sues impinge on the generalizability of thesemodels. Age-related differences in the norma-tive challenges associated with parenting couldaffect parents' cognitive-emotional processesregarding misbehavior. And parents' views ofphysical discipline and what constitutes harshdiscipline could vary with the child's age. Forexample, parents who endorse physical disci-pline may not view it as viable or acceptablewith adolescents. Thus, tests of model general-izability across age groups would require use ofdevelopmentally appropriate measurements ofparenting constructs. However, extant researchand theory provide some support for the pathlinking SES or ethnicity, stress, and parentingfor children of various ages (Conger, Ge, Elder,Lorenz, & Simons, 1994; Patterson, 1982,1986).

Parenting Beliefs

Parenting beliefs had a significant relationonly with SES. Consistent with the work ofKohn (1963, 1977) and Luster and colleagues


(Luster & Okagaki, 1993; Luster et al., 1989),low-income parents were more likely thanmiddle-income parents to endorse the physicaluse of power to resolve conflict, whether be-tween parent and child or between children. Thefinding of no relation between ethnicity andparenting beliefs contrasts with other findingssuggesting that African American parents andprofessionals are more likely to endorse spank-ing (e.g., Deater-Deckard & Dodge, 1997; Hef-fer & Kelley, 1987). This inconsistency may bedue to differences in the specific constructs as-sessed. In this study parents reported their be-liefs in the effectiveness of spanking their child,whereas in the Deater-Deckard and Dodge andHeffer and Kelly studies respondents reportedthe appropriateness or acceptability of spankingas a strategy. Do these findings suggest that theeffectiveness of physical discipline and accept-ability of physical discipline represent distinctbeliefs about physical discipline? If so, how arethese beliefs related? For example, can one be-lieve that a discipline strategy is effective, yetunacceptable? Further research is needed toclarify these subtle yet possibly important dis-tinctions among beliefs about spanking.

It also is possible that no ethnic differenceswere found because conservative religiousbeliefs—which are prevalent in the areas fromwhich the sample was drawn—are linked toendorsement of physical discipline (Ellison,Bartkowski, & Segal, 1996; Hynn, 1994) forboth European American parents and AfricanAmerican parents. Unfortunately, we could notexamine this possibility; its clarification awaitsfurther research with other samples.

Inclusion of beliefs about the merit of otherdiscipline responses (e.g., withdrawal of privi-leges and time-out) and parents' general prefer-ence for certain discipline responses wouldbroaden understanding of relations among con-structs in this model. Because African Ameri-can parents tend not to equate severity of disci-pline with physical discipline (Jackson, 1997),beliefs about the use of severe discipline re-sponses warrant examination as well. Parentingbeliefs also could be more comprehensively as-sessed with the inclusion of beliefs about posi-tive interactions with children such as praise,nurturance, supportive parenting (e.g., Pettit etal., 1997), and management of negative affect,which are important elements of parenting thatmay relate to discipline responses.

Inclusion of racial socialization goals such as

preparation for bias (e.g., Boykin & Toms,1985) in the model would enhance understand-ing of relations between parenting beliefs anddiscipline responses and could clarify furtherthe field's developing understanding of varia-tions in African American parenting (e.g.,Kelley et al., 1992; Stevenson, 1994; Thornton,Chatters, Taylor, & Allen, 1990). Relatedly,inclusion of parent's racial identity (e.g.,Helms, 1990) in the model might clarify within-group differences in relations between cognitive-emotional processes and discipline responses.Among African American parents, those whoseidentity is grounded in one perspective of theirgroup's history and current challenges maythink about and react differently to their chil-dren's misbehavior than parents who have adifferent orientation of the group's history andfuture (e.g., Spencer, 1983; Thornton et al.,1990).

Cognitive-Emotional Processes

This study is one of few that establishescognitive-emotional processes as a mediator ofthe relations between stress and discipline re-sponses and, indirectly, between SES and dis-cipline responses and ethnicity and disciplineresponses. In both models, findings suggestedthat parents who experience more stress endorseharsher discipline responses, in part becausethey experienced more intense cognitive-emotional processes.

As proximal predictors, the four cognitive-emotional processes accounted for at least 48%of the variance associated with discipline re-sponses. When confronted with child misbehav-ior, parents who attributed hostile intent to theirchildren, who were highly upset by and worriedabout the future implications of the misbehav-ior, and who had few alternative discipline strat-egies were more likely to choose physical pun-ishment and more severe punishment. Thesefindings are consistent with extant research(Dix, Reinhold, & Zambarano, 1990; Mills &Rubin, 1990; Mize, Pettit, & Brown, 1995) inthis area, as well as models of social informa-tion processing (Dodge, 1993). Dix (1991,1993) suggested that these processes becomeintensified when important socialization goalsare involved. Bacon and Ashmore (1986) fur-ther suggested that long-term socializationgoals may be related to discipline responses. Anext research step might be to examine moreexplicitly the effect of parents' long-term so-

396 PINDERHUGHES, DODGE, BATES, PETTCT, AND ZELLI

cialization goals on cognitive-emotional pro-cesses and discipline responses.

Discipline Responses

The potency of cognitive-emotional pro-cesses as a proximal influence on disciplineresponses suggests that the specific disciplineresponses conceptualized and assessed in thisstudy reflect more reactive discipline responses.That is, more intense cognitive-emotional pro-cesses were associated with more harsh disci-pline responses. In its explicit focus on physicaland severe punishment, this study examinedonly one aspect of discipline: certain power-assertive strategies. Parents routinely useother disciplinary strategies—power assertive(e.g., time-out) and non-power assertive (e.g.,reasoning)—in the service of preparing chil-dren to function effectively as adults (e.g.,Baumrind, 1997; Jackson, 1997; Lytton, 1997).As a group, discipline strategies may vary in thedegree to which they draw on reactive or delib-erative processes. These findings suggest thatboth cognitive-emotional processes and under-lying beliefs are related to use of harsh disci-pline responses. Future research that includesbroader assessments of different disciplinestrategies (i.e., including reasoning, withdrawalof privileges, and time-out) and assesses thedegree of deliberation and reactivity leading touse of discipline responses would clarify furtherthe relations found in this study.

Methodological Issues

In this study, parents' cognitive-emotionalprocesses about situationally specific child be-haviors were aggregated, as recommended byHolden (1995) and Goodnow (1995). However,perhaps aggregation of attitudes and disciplineresponses obscures within-individual variationthat may exist across situations (Sigel, 1986).Because the current study included only fivesituations that were not classified and aggre-gated into more general types, Sigel's hypothe-sis could not be tested thoroughly. Future re-search should examine situational sources ofwithin-individual variation in relations betweenparent cognitions and emotions and disciplineresponses. Nevertheless, note that parentsshowed some degree of similarity across thedifferent situations they encountered.

This study's findings must be viewed in thecontext of its limitations. Aside from those pre-

viously discussed (i.e., selection of measuresand model conceptualization), another limita-tion of this study is the use of single reporters,which runs the risk of respondent bias. How-ever, one discipline measure involved an inter-viewer rating of parent response to misbehaviorduring the previous year. The rating includedthe interviewer's knowledge of the misbehaviorand parent discipline responses, thus minimiz-ing the risk of respondent bias. Future studies inthis area should incorporate observation of par-ent discipline behaviors; yet, this method facesthe challenge of restricted ranges in behaviors(e.g., Deater-Deckard et al., 1996).

Social support has been found to moderatethe impact of stress on parenting (McLoyd,1990), Inclusion of measures of quality of sup-port may clarify the modest relations betweenethnicity and stress as well.

Inclusion of parents' own history of beingparented also could have enhanced this study'sdesign. Grusec, Hastings, and Mammone (1994)found relations between parents representationsof being parented and their parental cognitions.Perhaps these former representations influenceparenting: Parents reared within an explicit sys-tem of values and beliefs may be more likely toincorporate their own system of cognitions intotheir socialization. In contrast, parents lacking amodel of a coherent system of beliefs and val-ues may rely on reactive processes. Research onthese individual differences would help clarifythe relations between cognitive-emotional pro-cesses and discipline responses.

Finally, the cross-sectional design precludesconclusions about the causality of these rela-tions. Although we suggest that contextual fac-tors and cognitive-emotional processes tempo-rally precede parents' discipline responses,perhaps long-standing parental cognitions—such as attribution biases—temporally precede,thus influencing cognitive- emotional pro-cesses. Or perhaps parents who clearly informtheir child of their beliefs about certain types ofdiscipline experience certain cognitive-emotional processes in the face of child misbe-havior. Thus, longitudinal inquiry and experi-mentally designed studies may clarify furtherhow these processes temporally interrelate.

Conclusion and Implications forApplication and Policy

Culture and context exert their influences ondiscipline responses. This study is one of few to


show that mechanisms of these influences ap-pear to have at least two paths. One importantpath, via cognitive-emotional processes andperceptions of the child, may reactively guidediscipline responses in the mediation of culturaland contextual influences. Individuals experi-encing higher levels of stress are more likely toreport more reactive processes and, in turn, touse more harsh discipline responses. Becauseharsh discipline often is linked with inconsis-tency (e.g., Patterson, 1982), it may negativelyaffect children's functioning. Thus, any parentexperiencing these intense cognitive-emotionalprocesses may benefit from interventions tar-geted to reduce their intensity. Promotion ofself-monitoring of cognitions and affect, as wellas problem solving about alternative disciplinestrategies, may help make discipline decisionsless reactive and harsh. However, focus only onthese proximal processes probably will be in-sufficient for low-income parents, who facestressors related to economic hardship. Policiesand programs that effectively reduce these stres-sors may, in turn, reduce the intensity of themore proximal and mediating cognitive-emotional processes. Because low-income Af-rican American parents are especially vulnera-ble to the combined impact of these stressorsand other societally based constraints, interven-tions that are culturally sensitive and that ad-dress the specific barriers that African Amer-ican parents perceive they face may be moreeffective.

The second path influencing discipline re-sponses, via parenting beliefs, represents endur-ing beliefs that may proactively guide disciplineresponses. Group differences in this relationreflect the impact of culture—for example, SESor ethnicity—on how parents may perceive theneed to socialize their children for future suc-cess. Because these beliefs and discipline re-sponses may be linked to goals that parents havefor their children's success as adults (Kohn,1963), they may pose unique challenges forinterventions that are grounded in a different setof values (e.g., promotion of autonomy vs. pro-motion of adherence to authority).

These findings suggest that the degree towhich the two paths operate jointly or sepa-rately may differ as a result of SES or ethnicity.That is, both parenting beliefs and cognitive-emotional processes mediated SES-disciplinerelations, whereas only cognitive-emotionalprocesses mediated ethnicity-discipline rela-

tions. Future research on both paths using moresensitive measures of relevant constructs willextend knowledge of influences on parentingand identify further directions for parentinginterventions.

References

Abell, E., Clawson, M., Washington, W. N., Bost,K. K., & Vaughn, B. E. (1996). Parenting values,attitudes, behaviors, and goals of African Americanmothers from a low-income population in relationto social and societal contexts. Journal of FamilyIssues, 17, 593-613.

Arbuckle, J. L., & Wothke, W. (1999). AMOS 4.0user's guide. Chicago, IL: Small Waters Corpora-tion.

Bacon, M. K., & Ashmore, R. D. (1986). A consid-eration of the cognitive activities of parents andtheir role in the socialization process. In R. D.Ashmore & D. M. Brodzinsky (Eds.), Thinkingabout the family: Views of parents and children (pp.3-33). Hillsdale, NJ: Erlbaum.

Baumrind, D. (1997). Necessary distinctions. Psy-chological Inquiry, 8, 176-182.

Belsky, J. (1984). The determinants of parenting: Aprocess model. Child Development, 55, 83-96.

Boykin, A. W., & Toms, F. D. (1985). Black childsocialization: A conceptual framework. In H. P.McAdoo & J. L. McAdoo (Eds.), Black children:Social, educational, and parental environments(pp. 33-51). Beverly Hills, CA: Sage.

Bugental, D. B., Blue, J., & Lewis, J. (1990). Care-giver beliefs and dysphoric affect directed to diffi-cult children. Developmental Psychology, 26, 631-638.

Conger, R. D., Ge, X., Elder, G. H., Lorenz, F. O., &Simons, R. L. (1994). Economic stress, coercivefamily process, and developmental problems of ad-olescents. Child Development, 65, 541-561.

Conger, R. D., McCarty, J. A., Yang, R. K., Lahey,B. B., & Kropp, J. P. (1984). Perception of child,child-rearing values, and emotional distress as me-diating links between environmental stressors andobserved maternal behavior. Child Development,55, 2234-2247.

Crnic, K., & Greenberg, M. (1987). Maternal stress,social support, and coping: Influences on earlymother-child relationship. In C. Boukydis (Ed.),Research on support for parents and infants in thepostnatal period (pp. 25-40). Norwood, NJ: Ablex.

Deater-Deckard, K., & Dodge, K. A. (1997). Exter-nalizing behavior problems and discipline revisited:Nonlinear effects and variation by culture, context,and gender. Psychological Inquiry, 8, 161-175.

Deater-Deckard, K., Dodge, K. A., Bates, J. E., &Pettit, G. S. (1996). Physical discipline among Af-rican American and European American mothers:Links to children's externalizing behaviors. Devel-opmental Psychology, 32, 1065-1072.


Dix, T. (1991). The affective organization of parent-ing: Adaptive and maladaptive processes. Psycho-logical Bulletin, 120, 3-25.

Dix, T. (1993). Attributing dispositions to children:An interactional analysis of attribution in socializa-tion. Personality and Social Psychology Bulletin,19, 633-643.

Dix, T., & Lochman, J. E. (1990). Social cognitionsand negative reactions to children: A comparison ofmothers of aggressive and nonaggressive boys.Journal of Social and Clinical Psychology, 9, 418-438.

Dix, T., Reinhold, D. P., & Zambarano, R. J. (1990).Mothers' judgment in moments of anger. Merrill-Palmer Quarterly, 36, 465-486.

Dix, T., Ruble, D. N., & Zambarano, R. J. (1989).Mothers' implicit theories of discipline: Child ef-fects, parent effects, and the attribution process.Child Development, 60, 1373-1391.

Dodge, K. A. (1993). Social-cognitive mechanismsin the development of conduct disorder and depres-sion. Annual Review of Psychology, 44, 559-584.

Dodge, K. A., Pettit, G. S., & Bates, J. E. (1994).Socialization mediators of the relation between so-cioeconomic status and child conduct problems.Child Development, 65, 649-665.

Ellison, C. G., Bartkowski, J. P., & Segal, M. L.(1996). Do conservative Protestant parents spankmore often?: Further evidence from the nationalsurvey of families and households. Social ScienceQuarterly, 77, 663-673.

Feldman, S. S., & Wentzel, K. R. (1990). The rela-tionship between parenting styles, sons' self-restraint, and peer relations in early adolescence.Journal of Early Adolescence, 10, 439-454.

Flynn, C. P. (1994). Regional differences in attitudestoward corporal punishment. Journal of Marriageand the Family, 56, 314-324.

Fox, R. A., Platz, D. L., & Bentley, K. S. (1995).Maternal factors related to parenting practices, de-velopmental expectations, and perceptions of childbehavior problems. Journal of Genetic Psychology,156, 431-441.

Garcia Coll, C. T. (1990). Developmental outcome ofminority infants: A process-oriented look into ourbeginnings. Child Development, 61, 270-289.

Goodnow, J. (1984). Parents' ideas about parentingand development: A review of issues and recentwork. In M. E. Lamb, A. L. Brown, & B. Rogoff(Eds.), Advances in developmental psychology(Vol. 3, pp. 193-242). Hillsdale, NJ: Erlbaum.

Goodnow, J. (1995). Parents' knowledge and expec-tations. In M. Bomstein (Ed.), Handbook of parent-ing: Vol. 3. Status and social conditions of parent-ing (pp. 305-332). Mahwah, NJ: Erlbaum.

Grusec, J. E., Hastings, P., & Mammone, N. (1994).Parenting cognitions and relationship schemas.New Directions for Child Development, 66, 5-19.

Heffer, R. W., & Kelley, M. L. (1987). Mothers'acceptance of behavioral interventions for children:

The influence of parent race and income. BehaviorTherapy, 2, 153-163.

Helms, J. E. (Ed.). (1990). Black and White racialidentity—Theory, research and practice. NewYork: Greenwood Press.

Holden, G. W. (1995). Parental attitudes toward chil-drearing. In M. Borstein (Ed.), Handbook of par-enting: Vol. 3. Status and social conditions of par-enting (pp. 359-392). Mahwah, NJ: Erlbaum.

Hollingshead, A. A. (1979). Four-factor index ofsocial status. Unpublished manuscript, Yale Uni-versity, New Haven, CT.

Jackson, J. F. (1997). Issues in need of initial visita-tion: Race and nation specificity in the study ofexternalizing behavior problems and discipline.Psychological Inquiry, 8, 204-211.

Kelley, M. L., Power, T. G., & Wimbush, D. D.(1992). Determinants of disciplinary practices inlow-income Black mothers. Child Development,63, 573-582.

Kessler, R. C , & Neighbors, H. W. (1986). A newperspective on the relationships among race, socialclass, and psychological distress. Journal of Healthand Social Behavior, 27, 107-115.

Kohn, M. L. (1963). Social class and parent-childrelationships: An interpretation. American Journalof Sociology, 68, 471-480.

Kohn, M. L. (1977). Class and conformity: A study ofvalues. (2nd ed.). Chicago: University of ChicagoPress.

Luster, T., & Kain, R L. (1987). The relation be-tween family context and perceptions of parentalefficacy. Early Child Development and Care, 29,301-311.

Luster, T., & Okagaki, L. (1993). Multiple influenceson parenting: Ecological and life-course perspec-tives. In T. Luster & L. Okagaki (Eds.), Parenting:An ecological perspective (pp. 227-250). Hillsdale,NJ: Erlbaum.

Luster, T., Rhoades, K., & Haas, B. (1989). Therelation between parental values and parenting be-havior: A test of the Kohn hypothesis. Journal ofMarriage and the Family, 51, 139-147.

Lytton, H. (1997). Physical punishment is a problem,whether conduct disorder is endogenous or not.Psychological Inquiry, 8, 211-214.

MacKinnon-Lewis, C , Lamb, M. E., Arbuckle, B.,Baradaran, L. P., & Volling, B. L. (1992). Therelationship between biased maternal and filial at-tributions and the aggressiveness of their interac-tions. Development and Psychopathology, 4, 403—415.

McLoyd, V. C. (1990). The impact of economichardship on Black families and children: Psycho-logical distress, parenting, and socioeconomic de-velopment. Child Development, 61, 311-346.

McLoyd, V. C. (1999, June). Ethnicity, race, andsocial class as defining features of families. Paperpresented at the Family Research Consortium Sum-mer Institute, Bretton Woods, NH.


McLoyd, V. C , Jayaratne, T. E., Ceballo, R., &Borquez, J. (1994). Unemployment and work inter-ruption among African American single mothers:Effects on parenting and adolescent socioemotionalfunction. Child Development, 65, 562-589.

Mills, R. S. L., & Rubin, K. H. (1990). Parentalbeliefs about problematic social behaviors in earlychildhood. Child Development, 61, 138-151.

Mize, J., Pettit, G. S., & Brown, E. G. (1995). Moth-ers' supervision of their children's peer play: Rela-tions with beliefs, perceptions, and knowledge. De-velopmental Psychology, 31, 311-321.

Myers, H. R, & King, L. (1983). Mental health issuesin the development of the Black American child. InG. Powell, J. Yamamoto, A. Romero, & A. Morales(Eds.), The psychosocial development of minoritygroup children (pp. 275-306). New York: Brunner/Mazel.

National Center for Education Statistics. (1993). Trendsin the well-being of American youth [On-line].Available: http://www.ed.gov/pubs/YoulhmdicatQrs/Demographics .html

Ogbu, J. U. (1981), Origins of human competence: Acultural-ecological perspective. Child Develop-ment, 52, 413-429.

Okagaki, L., & Divecha, D. J. (1993). Developmentof parental beliefs. In T. Luster & L. Okagaki(Eds.), Parenting: An ecological perspective (pp.35-68). Hillsdale, NJ: Erlbaum,

Patterson, G. R. (1982). Coercive family process.Eugene, OR: Castalia.

Patterson, G. (1986). Performance models for anti-social boys. American Psychologist, 41, $$2-44$.

Peters, M. (1985). Racial socialization of youngBlack children. In H. McAdoo & J. McAdoo (Eds.),Black children (pp. 159-173). Beverly Hills, CA:Sage.

Pettit, G. S., Bates, J. E., & Dodge, K. A. (1997).Supportive parenting, ecological context, and chil-dren's adjustment: A seven-year longitudinal study.Child Development, 68, 908-923.

Phares, V. (1996). Fathers and developmental psy-chopathology. New York: Wiley.

Reis, J., Barbera-Stein, L., & Bennett, S. (1986).Determinants of parenting. Family Relations, 36,547-564.

Rubin, K. H., & Mills, R. S. L. (1992). Parents'thoughts about children's socially adaptive andmaladaptive behaviors: Stability, change, and indi-vidual differences. In I. E. Siegel, A. V.McGillicuddy-DeLisi, & J. J. Goodnow (Eds.), Pa-rental belief systems: The psychological conse-quences for children (pp. 41-68). Hillsdale, NJ:Erlbaum.

Rubin, K. R , Stewart, S. L., & Chen, X. (1995).Parents of aggressive and withdrawn children. InM. Borstein (Ed.), Handbook of parenting: Vol. 3.Status and social conditions of parenting (pp. 255-277). Mahwah, NJ: Erlbaum.

Sigel, I. E. (1986). Reflections on the belief-behaviorconnection: Lessons learned from a research pro-gram on parental belief systems and teaching strat-egies. In R. D. Ashmore & D. M. Brodzinsky(Eds.), Thinking about the family: Views of parentsand children (pp. 35-66). Hillsdale, NJ: Erlbaum.

Snyder, H. N. (1996). The juvenile court and delin-quency cases. The Future of Children: The JuvenileCourt, 6(3), 53-63.

Spencer, M. B. (1983). Children's cultural values andparental child rearing strategies. Developmental Re-view, 3, 351-370.

Stehno, S. M. (1982). Differential treatment of mi-nority children in service systems. Social Work, 27,39-46.

Stevenson, H. C , Jr. (1994). Racial socialization inAfrican American families: The art of balancingintolerance and survival. The Family Journal:Counseling and Therapy for Couples and Families,2, 190-198.

Strassberg, Z. (1995). Social information process-ing in compliance situations by mothers ofbehavior-problem boys. Child Development, 66,376-389.

Straus, M. A. (1979). Measuring intrafamily conflictand violence: The Conflict Tactics (CT) Scales.Journal of Marriage and the Family, 41, 75-88.

Straus, M. A., & Gelles, R. J. (1990). Physical vio-lence in American families. New Brunswick, NJ:Transaction.

Tatum, B. D. (1987). Assimilation blues: Black fam-ilies in a White community. Northampton, MA:Hazel-Maxwell.

Thornton, M. C , Chatters, L. M., Taylor, R. J., &Allen, W. R. (1990). Sociodemographic and envi-ronmental correlates of racial socialization byBlack parents. Child Development, 61, 401-409.

Ventura, S. J., Taffel, S. M., Mosher, W. D., Wilson,J. B,, & Henshaw, S. (1995). Trends in pregnanciesand pregnancy rates: Estimates for the UnitedStates, 1980-92. Monthly Vital Statistics Report,45,(11, Suppl), 1-24. Retrieved from World WideWeb: http://cdc.gov/search.htm.

Zuravin, S. J. (1987). Unplanned pregnancies, familyplanning problems, and child maltreatment. FamilyRelations: Journal of Applied Family and ChildStudies, 36, 135-139.


Appendix

Standardized Factor Loadings and Error Variances for the SES-Discipline Response Measurement Model

Measure Loading

SESFather educationMother educationHollingshead score

StressInterviewer ratingsUnplanned pregnancyViolence exposureSingle parenthoodNumber of children

Discipline responsesInterviewer ratingsSeverityPhysical punishment

Parenting beliefsAggression valuesSpanking

Child perceptionPositive affectUnique description

Cognitive-emotional processesAttributionsBeing worriedUpset affectAlternative strategies

.86

.76

.90

.40

.28

.40

.43

.18

.39

.65

.52

.19

.53

.52

.60

.21

.58

.63

.34

Variance

.51

.65

.44

.92

.97

.92

.90

.98

.92

.76

.85

.98

.84

.85

.80

.97

.81

.78

.93

Note. SES = socioeconomic status.

Received October 22, 1998Revision received March 15, 2000

Accepted April 19, 2000

American Psychological Association (APA) - Discipline ...in this relation across ethnicity (e.g.,...

Documents

Transcript of American Psychological Association (APA) - Discipline ...in this relation across ethnicity (e.g.,...