Journal of Family Psychology Copyright 1996 by the American Psychological Association, Inc. 1996, Vol. 10, No. 3, 243-268 O893-32O0/9OT3.OO Parental Meta-Emotion Philosophy and the Emotional Life of Families: Theoretical Models and Preliminary Data

John M. Gottman, Lynn Fainsilber Katz, and Carole Hooven University of Washington

This article introduces the concepts of parental meta-emotion, which refers to parents' emotions about their own and their children's emotions, and meta-emotion philosophy, which refers to an organized set of thoughts and metaphors, a philosophy, and an approach to one's own emotions and to one's children's emotions. In the context of a longitudinal study beginning when the children were 5 years old and ending when they were 8 years old, a theoretical model and path analytic models are presented that relate parental meta-emotion philosophy to parenting, to child regulatory physiology, to emotion regulation abilities in the child, and to child outcomes in middle childhood.

The importance of parenting practices for refers to executive functions of emotion. In this children's long-term psychological adjustment article, we discuss the evolution of the meta- has been a central tenet in developmental and emotion construct and describe its relationship family psychology. In this article, we introduce to various aspects of family and child function- a new concept of parenting that we call parental ing. We present a parsimonious theoretical meta-emotion philosophy, which refers to an model of the role of parental meta-emotions in organized set of feelings and thoughts about children's emotional development, operational- one's own emotions and one's children's emo- ize this model, and present some path analytic tions. We use the term meta-emotion broadly to tests of the model. In this model, we argue (a) encompass both feelings and thoughts about that parental meta-emotion philosophy is re- emotion, rather than in the more narrow sense lated to both the inhibition of parental negative of one's feelings about feelings (e.g., feeling affect and the facilitation of positive parenting; guilty about being angry). The notion we have (b) that it directly affects children's regulatory in mind parallels metacognition, which refers to physiology; and (c) that this, in turn, affects the executive functions of cognition (Allen & children's ability to regulate their emotions— Armour, 1993; Bvinelli, 1993; Flavell, 1979; hence, parental meta-emotion philosophy has an Fodor, 1992; Olson & Astington, 1993). In an impact on a variety of child outcomes. analogous manner, meta-emotion philosophy Background John M. Gottman, Lynn Fainsilber Katz, and Car- ole Hooven, Department of Psychology, University The Concept of Meta-Emotion of Washington. The research of this article was supported by Na- Research in developmental psychology on the tional Institute of Mental Health (NIMH) Research effects of parenting has focused on parental Grants MH42484 and MH35997, by an NIMH Merit affect and discipline, selecting variables such as Award to extend research in time, and by Research warmth, control, authoritarian or authoritative Scientist Award K2MH00257. styles, and responsiveness (see Ainsworth, Bell, This research received a great deal of support from & Stayton, 1971; Baumrind, 1967, 1971; Michael Guralnick, director of the Center for Human Becker, 1964; Cohn, Cowan, Cowan, & Pear- Developmental Disabilities (CHDD), and CHDD's son, 1992; C. P. Cowan & P. A. Cowan, 1992; core facilities, particularly the Instrument Develop- Maccoby & Martin, 1983; Patterson, 1982; and ment Laboratory at the University of Washington. Correspondence concerning this article should be Schaefer, 1959). Little attention has been placed addressed to John M. Gottman, Department of Psy- on examining the parents' feelings and cogni- chology, Box 351525, University of Washington, tions about their own affect or their feelings and Seattle, Washington 98115. cognitions about their child's affect.

243 244 GOTTMAN, KATZ, AND HOOVEN

Our review of popular parenting guides also sadness was important information that some- revealed that the overwhelming majority of thing was missing in one's life. these parenting guides are based on obtaining This area of meta-emotion is probably char- and maintaining child discipline. However, one acterized by great variability even in laboratory genre of parenting guides focuses on children's experiments that elicit emotions. Researchers emotions and on how to make immediate and have reported large variability in results from everyday emotional connections with a child laboratory experiments designed to elicit emo- that are not critical and contemptuous, but ac- tion, such as the startle response. Ekman, cepting. These kinds of parenting guides can be Friesen, and Simons (1985) reported a consis- traced to the seminal influence of one child tent set of responses across participants to being psychologist: Haim Ginott (Ginott, 1956, 1971, startled, but there were huge individual differ- 1975). Although many psychological systems ences in the emotional response to having been of thought (e.g., attachment theory, psychoanal- startled, that is, in people's meta-emotions to ysis) have written about the importance of the the startle; Levenson and Sutton (personal com- child's affect, Redl (1966) and Ginott both em- munication, June 15, 1994) reported a similar phasized intervening with a child's strong neg- set of results. Meta-emotion may be a pervasive ative emotions while the child is having the and understudied dimension in emotion re- emotions. They also emphasized intervening di- search. rectly, dealing with the child's conscious thoughts and actions. This difference was so important, in our view, that it amounted to a revolution in "how one deals with children," to An Emotion-Coaching Meta-Emotion use Redl's (1966) words. Philosophy Our initial interest was in this concept of In our pilot work, we noticed that there are parents' awareness of their children's emotional some parents who are aware of the emotions in lives and their attempts to make emotional con- their lives (particularly the negative emotions), nections with their children. This interest led to who can talk about those emotions in a differ- the development of our meta-emotion interview entiated manner, who are aware of these emo- (Katz & Gottman, 1986). All the parents were tions in their children, and who assist their separately interviewed about their own experi- children with their emotions of anger and sad- ence of sadness and anger, their philosophy ness, acting like an emotion coach. This is a of emotional expression and control, and their parental meta-emotion philosophy we call an attitudes and behavior about their children's emotion-coaching philosophy. We found that an anger and sadness. In our pilot work, we dis- emotion-coaching meta-emotion philosophy covered a great variety in the emotions, experi- had five components: parents (a) said that they ences, philosophies, and attitudes that parents were aware of low intensity emotions in them- had about their own emotions and the emotions selves and in their children; (b) viewed the of their children. For example, one pair of par- child's negative emotion as an opportunity for ents said that they viewed anger as "from the intimacy or teaching; (c) validated their child's devil," and that they would not permit them- emotion; (d) assisted the child in verbally label- selves or their children to express anger. Some ing the child's emotions; and (e) problem solved parents were accepting of sadness and anger but with the child, setting behavioral limits, and did not engage in problem solving with their discussing goals and strategies for dealing with child. Other parents were not disapproving of the situation that led to the negative emotion. anger but, instead, in laissez-fairre fashion, ig- We hypothesized that these parents have a nored anger in their children. Still other parents greater ability than other parents to maneuver in encouraged the expression and exploration of the world of emotions, that they are more com- anger. There was similar variation with respect fortable with the world of emotions, and that to sadness. Some parents minimized sadness in they are better able to regulate emotions. We themselves and in their children, saying such expected them to be more affectionate with their things as, "I can't afford to be sad," or "What children and less autocratic than other parents. does a child have to be sad about?" Other par- However, it was our observation that an ents thought that emotions like sadness in their emotion-coaching meta-emotion philosophy children were opportunities for intimacy, that was different from parental warmth, and we META-EMOTION 245 tested this notion in the current study. Very emotion; for example, we might only be study- concerned, generally positive and warm parents ing how parents feel about getting angry at their can be oblivious to the world of emotion, and an children (e.g., feel guilty about getting angry). emotion-coaching meta-emotion philosophy is However, we use the term broadly to encom- something additional that these parents bring to pass feelings and thoughts about emotion. As their roles as parents. Perhaps warmth and limit the examples just provided suggest, the con- setting are correlated with these meta-emotion struct being tapped involves parents' feelings variables, but we think that they are not the and thoughts about their own and their chil- same dimensions of parenting. dren's emotions, their responses to their child's In contrast, we found that a dismissing meta- emotions, and their reasoning about these re- emotion philosophy was one in which parents sponses (i.e., what the parent is trying to teach felt that the child's sadness or anger were po- the child when responding to the child's anger). tentially harmful to the child, that it was the This broader construct indexes a fundamental parents' job to change these toxic negative emo- attitude or approach to emotion. For some peo- tions as quickly as possible, that the child ple, emotions are a welcome and enriching part needed to realize that these negative emotions of their lives; they believe, in a fundamental would not last and were not very important, and way, that it is OK to have feelings. However, that it was the parent's job to convey to the child for other people, emotions are to be avoided and a sense that he or she could ride out these minimized; the world of negative emotions is negative emotions without damage. We found seen as dangerous (see Appendix). that emotion-dismissing families could be sen- sitive to their children's feelings and wanted to be helpful, but their approach to sadness, for example, was to ignore or deny it as much as Meta-Emotion and Parenting possible. Often they perceived a child's strong From a theoretical standpoint, we think our emotion as a demand that they fix everything measures of parental meta-emotion philosophy and make it better. These parents hoped that the are embedded within a web of measures that tap dismissing strategy would make the emotion go parent-child interaction. We expect that par- away quickly. They often conveyed a sense that ents' meta-emotion philosophy is not indepen- the child's emotion was something they may dent of their parenting. Hence, in our theoretical have been forced to deal with, but it was not model, we include meta-emotion variables interesting or worthy of attention in itself. They along with parenting variables. It is our view described sadness as something to get over, ride that we need to be specific about our description out, but look beyond and not dwell on. They of parenting; for heuristic purposes (within the often used distractions when their child was sad restricted range of families in our samples), we to move the child along, and they even used discuss three dimensions of parenting behav- comfort, but within specified time limits, as if iors. they were impatient with the negative emotion itself. They preferred a happy child and often First, we started with everyday, mundane found these negative states in their child quite negativity. Inherent in the literature on parent- painful. They did not present an insightful de- ing is the idea that small things in everyday scription of their child's emotional experience parenting can be quite harmful for children (or and did not help the child with problem solving. serve as indexes of more harmful types of par- They did not see the emotion as beneficial or as enting), akin to what J. Reid has called "natter- any kind of opportunity, either for intimacy or ing" (see Patterson, 1982). Ginott (1965) wrote for teaching. Many dismissing families saw strongly about this in his discussion of the im- their child's anger (without misbehavior) as portance of (a) understanding and validating the enough cause for punishment or a Time Out. child's emotions and (b) avoiding contempt and disapproval. Thus, in our measurement of this It is important to point out that the term negativity in our laboratory-based parent-child meta-emotion is being used in its broadest interactions, we included three variables: paren- sense. Metacommunication is communication tal intrusiveness, criticism, and mockery. We about communication, and metacognition is call this type of parenting derogatory. In a cognition about cognition. Meta-emotion in the teaching task, as some of the parents in our narrow sense might refer to emotions about laboratory instructed their children, they mixed 246 GOTTMAN, KATZ, AND HOOVEN in a blend of frustration, taking over for the our view, far richer than the parenting research child as soon as the child had trouble with the literature. For example, what would we predict task (which we call intrusiveness), using criti- that an authoritative parent would do (or rec- cism and derisive humor (mockery, humiliation, ommend that he or she should do) when his or belittlement of the child). We think that this her child has just had a nightmare? Being warm dimension of parenting represents the microso- and structuring provides no real guidelines. cial processes characteristic of parental rejec- Emotion coaching does provide these guide- tion (e.g., Whitbeck, Hoyt, Simons, & Conger, lines. 1992). What is the expected relationship between Next, we also wished to measure two kinds of meta-emotion and the derogation, warmth, and positive parenting. The first is the kind of scaffolding-praising dimensions? When we be- warmth that Baumrind (1967, 1971) and others gan this study, we had two working hypotheses. described: We refer to this dimension of posi- The first hypothesis was that a coaching meta- tive parenting as warmth. Following C. P. emotion philosophy might be nested within a Cowan and P. A. Cowan (1992), we include in web of positive parenting. We proposed that an this dimension of warmth co-warmth, which emotion-coaching meta-emotion philosophy en- includes warmth between parents while inter- tails parenting that goes a step beyond the idea acting with the child. Our second dimension of of warmth; that is, we suggested that it entails positive parenting involves a positive structur- scaffolding-praising parenting. The second hy- ing, responsive, enthusiastic, engaged, and af- pothesis was that meta-emotion performs its fectionate parenting during the teaching task in major function by inhibiting parental deroga- our laboratory. This type of positive parental tion; in particular, as Ginott (1965) noted, we response goes beyond warmth. It includes the proposed that understanding and validating the responsive style that attachment theorists have child's emotions serves to avoid criticism, con- identified, but it is more complex than that. We tempt, and disapproval of the child. Most of the call it scaffolding-praising (on the general scaf- examples from Ginott's books had to do with folding concept, see Choi, 1993; Kirchner, the importance of emotion coaching in avoiding 1991; Pratt, Kerig, Cowan, & Cowan, 1988; and escalating negativity, frustration, disapproval, Vygotsky, 1987). From watching and coding and emotional distance between parents and the videotapes, we noticed that this is a dimen- children. It appears to have been first suggested sion quite different from Baumrind's authorita- foremost as a mechanism for obtaining exten- tive parenting. Parents high on the scaffolding- sive relief from spiraling negativity: Perhaps praising dimension provided structure for the validating the child's affect serves as an oppo- task, stating the goals and procedures of the nent process to derogation. Hence, it was en- game simply, in a relaxed manner, and with low information density; they then waited for their tirely reasonable to hypothesize that the major child to act and commented primarily when the effect of a coaching meta-emotion philosophy child did something right, acting like a cheering would be inhibiting parental negativity and that section at a football game, giving praise and it might have no effect on positive parenting. approval. Parents low on this scaffolding-prais- ing dimension either provided little structure for the learning situation for their children, or they Meta-Emotion and the Development of gave information rapidly, with high density, and Emotion Regulation Abilities enthusiastically, appearing to excite and con- fuse the child; such parents then waited to com- Precisely how do we think that meta- ment until their child had made a mistake. emotions affect the functioning of families and These parents were then usually critical of the act to affect child outcomes? What do we pro- child's performance. pose as the mechanism? We are particularly drawn to theories that attempt to integrate be- Are the concepts of meta-emotion and emo- havior and physiology, and our theorizing is tion coaching simply subdimensions of positive oriented toward approaches that have empha- parenting? We think not; we think that they add sized the importance of (a) the development of to current concepts in the parenting literature children's abilities in the regulation of emotion and are more general. Emotion coaching is one (Garber & Dodge, 1991) and (b) the develop- reason why the parenting advice literature is, in ment of children's abilities to self-soothe strong META-EMOTION 247 and potentially disruptive emotional states greater soothability; it is also related to greater (Dunn, 1977), focus attention, and organize ability to focus attention and greater ability to themselves for coordinated action in the service self-soothe and explore novel stimuli (DiPietro of some goal. We think that these general sets of & Porges, 1991; Fox, 1989; Hofheimer & Law- abilities underlie the development of other com- son, 1988; Huffman, Bryan, Pederson, & petencies. We are especially interested in chil- Porges, 1988; Linnemeyer & Porges, 1986; Por- dren's peer social skills, particularly because of ter, Porges, & Marshall, 1988; Richards, 1985, their predictive validity (Parker & Asher, 1987). 1987; Stifter & Fox, 1990; Stifter, Fox, & Central peer social competencies include the Porges, 1989). ability to resolve conflict, to find a sustained There is also another dimension of vagal tone common ground play activity, to compromise in that needs to be considered, namely, the ability play, and to empathize with a peer in distress to suppress vagal tone. In general, vagal tone is (e.g., see Asher & Coie, 1990; Gottman, 1983; suppressed during states that require focused or Gottman & Parker, 1986). sustained attention, mental effort, attention to We suggest that fundamental to these abilities relevant information, emotional interaction, and is the ability to soothe one's self physiologically organized responses to stress. Thus, the child's and to focus attention. These abilities underlie ability to perform a transitory suppression of being able to listen to what one's playmate is vagal tone in response to environmental (and saying, being able to take another's role and particularly emotional) demands is another in- empathize, and being able to engage in social dex that needs to be added to the child regula- problem solving. They involve the child know- tory physiology construct.2 It relates to the like- ing something about the world of emotion, both lihood of approach rather than withdrawal; his or her own and others'. We propose that this some infants with a high vagal tone who were knowledge arises only out of emotional connec- unable to suppress vagal tone in attention- tion being important in the home. In the follow- demanding tasks exhibited other regulatory dis- ing section, we briefly review our reasons for orders (e.g., sleep disorders; Huffman, Bryan, measuring child physiology as related to the Pederson, & Porges, 1992; Porges, Walter, construct of emotion regulation. Korb, & Sprague, 1975). Porges, Doussard- Roosevelt, and Portales (1992) found that 9-month-old infants who had lower baseline Regulatory Physiology1 vagal tone and less vagal tone suppression dur- ing the Bailey examination had the greatest We used Porges' (1984) suggestion that there behavioral problems at 3 years of age, as mea- may be a physiological basis for the ability to sured by the Achenbach and Edelbrock (1986) regulate emotion. To explain his notions, we Child Behavior Checklist. Measures of infant discuss two concepts related to the child's para- sympathetic nervous system (PNS) physiology. 1 The major nerve of the PNS is called the vagus We did not include the sympathetic portion of the nerve. The vagus nerve (so called because it is child's physiology in our modeling. However, we did the vagabond nerve that travels throughout the find that the child's concentration of adrenaline in the 24-hr urine sample at age 5 correlated (r = 0.39, p < body, innervating the viscera) is the X-th cranial .001) with the child's illness at age 8. nerve. The tonic firing of the vagus nerve slows 2 We hypothesize that basal vagal tone is related to down many physiological processes, such as the the child's ability to sustain attention, whereas the heart rate. Research by Porges and his col- ability to suppress vagal tone is related to the child's leagues on the PNS has indicated a strong as- ability to shift attention when that is called for. sociation between high vagal tone and good Porges and Doussard-Roosevelt (in press) pointed attentional abilities, and there is speculation that out that one must be cautious about expecting the these processes are related to emotion regula- suppression of vagal tone to always be the appropri- tion abilities. Porges (1992) reviewed evidence ate vagal response to external demands. In the neo- that suggests that a child's baseline vagal tone is natal intensive care unit, the appropriate response to related to the child's capacity to react to envi- gavage feeding turned out to be increases in vagal tone, consistent with the support of digestive pro- ronmental stimuli. There is a substantial body of cesses (DiPietro & Porges, 1991). Premature infants literature showing that basal vagal tone is re- who increased vagal tone during gavage feeding had lated to both greater behavioral reactivity and significantly shorter hospitalizations. 248 GOTTMAN, KATZ, AND HOOVEN temperament derived from maternal reports "cool," and emotionally unflappable (see Gott- (Bates, 1980) were not related to the 3-year man & Parker, 1986). It is well known that the outcome measures.3 worst thing a middle-childhood child can do when entering a group of peers is to start talking The Theoretical Challenge in Predicting about his or her own feelings as parents and children do in an emotion-coaching interaction. Peer Social Competence in Middle Thus, the basic elements and skills a child Childhood From Emotion-Coaching learns through emotion coaching (labeling, ex- Interactions pressing one's feelings, talking about one's feelings, and drawing attention to one's self) A major goal of our research was to predict become liabilities in the peer social world in peer social relations in middle childhood from middle childhood. Therefore, the basic model variables descriptive of the family's emotional life in preschool. It is now well known that the linking emotion coaching in preschool to peer ability to interact successfully with peers and to relations in middle childhood cannot be a sim- form lasting peer relationships are important ple isomorphic transfer of social skills model. developmental tasks. Children who fail at these Instead, it becomes necessary to identify a tasks, especially in the making of friends, are at mechanism that makes it possible for the child risk for a number of later problems (Parker & to learn something in the preschool period that Asher, 1987). The peer context presents new underlies the development of appropriate social opportunities and formidable challenges to chil- skills across this major developmental shift in dren. Interacting with peers provides opportuni- what constitutes social competence with peers, ties to learn about more egalitarian relation- the development in the child of what Salovey & ships, to form friendships with agemates, to Mayer (1990) and Goleman (1995) called negotiate conflicts, to engage in cooperative and "emotional intelligence," a kind of "social competitive activities, and to learn appropriate moxie." A number of researchers are addressing limits for aggressive impulses. On the other concepts related to this idea of the child's de- hand, children are typically less supportive than veloping social intelligence, such as the child's caregivers when their peers fail at these tasks. In our research, we have found that the quality of developing ability to cope with stress (Saarni, the child's peer relationships forms an impor- 1993), the child's emotional competence (Den- tant class of child outcome measures. ham, Renwick, & Hewes, 1994), the child's developing empathy (Eisenberg & Fabes, 1990; We should explain what the theoretical chal- Eisenberg & Strayer, 1987), the child's devel- lenge is in predicting peer social relations across oping social understanding (Denham, Zoller, & these two major developmental periods, from Couchoud, 1994; Dunn & Brown, 1994), the preschool to middle childhood. Major changes child's developing social and emotional compe- occur in peer relations in middle childhood. Children become aware of a much wider social tence and regulation as well as the child's de- network than the dyad. In preschool, children veloping theory of social mind (Casey & Fuller, are rarely capable of sustaining play with more 1994; Fox, 1994; Harris & Kavanaugh, 1993; than one other child (e.g., see Corsaro, 1979, Thompson, 1991), and the child's ability to 1981). However, in middle childhood, children recognize emotional expressions (Cassidy, become aware of peer norms for social accep- Parke, & Butkovsky, 1992; Walden, 1988). tance, and teasing and avoiding embarrassment suddenly emerge (see Gottman & Parker, 1986). Children become aware of clique struc- tures and of influence patterns as well as social 3 Time 1 and Time 2 down regulation scales cor- acceptance. The correlates of peer acceptance related 0.41, p = .003. The Time 1 down regulation and rejection change dramatically, particularly scale was unrelated to awareness and coaching, but it with respect to the expression of emotion. One correlated .31 (p = .020) with derogation, .44 (p = of the most interesting changes is that the so- .004) with Time 2 child negative affect, and 0.59 (p < .001) with Time 2 teacher ratings of negative cially competent response to a number of salient peer relations. It was unrelated to scaffolding- social situations, such as peer entry and teasing, praising parenting, to basal vagal tone, or to suppres- is to be a good observer who is somewhat wary, sion of vagal tone. META-EMOTION 249

The Theoretical Model: Hypotheses About 1981), particularly negativity such as aggres- Meta-Emotion, Parenting, Regulatory sion, whining, oppositional behavior, fighting Physiology, and Child Outcomes requiring parental intervention, sadness, and anxiety with peers. We also expected that an In building our theoretical model, we sought emotion-coaching meta-emotion philosophy to explain how meta-emotion philosophy might would predict the development of superior cog- be related to a variety of child outcomes. The nitive skills of the child (through superior vagal theoretical model is depicted in Figure 1. Given tone and greater ability to focus attention). We the hypothesized effects of parental meta- predicted that the relationship between the par- emotion philosophy on parenting skills, we ex- ents' meta-emotion philosophy and the child's pected that some effects would occur through achievement at age 8 years would hold over and parenting practices. To be specific, we hypoth- above preschool measures of intelligence. Thus, esized that the emotion-coaching meta-emotion we predicted that two preschool children of philosophy would be related to scaffolding- equal intelligence would differ, in part, in their praising and to the inhibition of parental dero- ultimate achievement in school as a function of gation. We also expected meta-emotion philos- the parents' meta-emotion philosophy. Finally, ophy, high scaffolding-praising, and low we examined the child's physical health as an derogation to be related to superior emotion outcome variable. Because the vagus innervates regulation abilities (as indexed by PNS func- the thymus gland (Bulloch & Moore, 1981; tioning and parental report). We also asked Bulloch & Pomerantz, 1984; Magni, Bruschi, & whether effects between child physiology and Kasti, 1987; Nance, Hopkins, & Bieger, 1987), emotion coaching were bidirectional and a central part of the immune system that is whether our effects varied with child tempera- involved in the production of T-cells, we also ment. One concern with these results was expected that basal vagal tone would be related whether parents were coaching their children to better child physical health. The path models differentially as a function of the children's also tested direct effects of meta-emotion, temperament. which are theoretically unexplained by the We also proposed that there would be a rela- model. tionship between the parents' meta-emotion coaching philosophy and a variety of child out- We recognize that our correlational studies comes. We hypothesized that a parental meta- could not provide us with a causal understand- emotion coaching philosophy would be related ing of the theoretical pathways we proposed. to the child's developing social competence However, we expected the correlational data to with other children. We expected that the yield results consistent with or disconfirming child's peer social competence would hold in these causal models, thereby suggesting some the inhibition of negative affect (Guralnick, directions for future research.

Direct Pathway

Child Outcomes A

Child Emotion Regulation Abilities

Figure I. Summary model for how parental meta-emotion might influence child outcomes. 250 GOTTMAN, KATZ, AND HOOVEN

Method was only mildly interesting for the children and hard to recall. The second task involved teaching the child An abbreviated set of procedures is presented in how to play an Atari game that the parents had this article in the interests of conserving space. See learned to play while the child was hearing the story. Gottman and Katz (1989) and Katz and Gottman The interaction lasted 10 min. (1993) for more detail. Children's film viewing. During a second visit to the laboratory, children were shown segments of emotion-eliciting films. Our main interest was to Participants obtain indexes of physiological activity during emo- tional and nonemotional events. Each film clip was Fifty-six normal families were recruited from the preceded by a neutral story and an emotion induction Champaign-Urbana, Illinois, community for this film clip of an actress who acted out the emotions of study; 24 of the families had a male and 32 a female the protagonist in the upcoming story. The function 4- to 5-year-old child. We used a telephone version of of the emotion induction was to direct the child to the Locke-Wallace marital satisfaction scale identify with the protagonist and experience the spe- (Krokoff, 1984) to ensure that our study included cific emotion in question. Although each film clip couples with wide range of marital satisfaction levels. was designed to elicit a specific emotion, the emotion The mean marital satisfaction score was 111.1 (SD = elicitation was not very successful; instead, we ob- 29.6). Three years later, we recontacted 53 of the 56 tained a range of facial expressions of emotion during families (94.6% of the original sample). in each film. Hence, we refer to the films by their titles instead of by the emotion they were intended to induce. The child viewed clips from six films: (a) Fly Procedure fishing; (b) Wizard ofOz (Leroy & Flemming, 1939), flying monkey scene; (c) Charlotte's Web (Barbara, The procedure involved laboratory sessions and Nicolas, & Takamoto, 1973), Charlotte dies; (d) home interviews for both parents and children. We Meaning of Life (Goldstone & Gilliam, 1983), res- used a combination of naturalistic interaction, highly taurant scene; (e) Wizard of Oz, taking Toto away; structured tasks, and semistructured interviews. and (f) Daisy (Alcroft & Mitton, 1984). Home and laboratory visits consisted of two home Child's physiological functioning. We assessed visits-one with the marital couple and one with the the following physiological variables from the child child—and three laboratory visits—one with the cou- under baseline conditions, during parent-child inter- ple only, one with the couple and their 4 to 5-year-old action, and during film-viewing: child, and one with the child alone. Families were seen at Time 1, when children were 4-5 years old, 1. Cardiac interbeat interval (IBI). This measure and again three years later at Time 2, when children was determined by measuring the time interval be- were on average 8 years old. tween successive spikes (R-waves) of the electrocar- diogram (EKG). 2. Pulse transmission time to the finger (PTT-F). Time 1 This was a measure of the elapsed time between the R-wave of the EKG and the arrival of the pulse wave at the finger. Assessment 3. Finger pulse amplitude (FPA). This was an Meta-emotion interview. All parents were sepa- estimate of the relative volume of blood reaching the rately interviewed about their own experience of sad- finger on each heart beat. ness and anger; their philosophy of emotional expres- 4. Skin conductance level (SCL). This measure sion and control; and their feelings, attitudes, and was sensitive to changes in levels of sweat in the behavior about their children's anger and sadness eccrine sweat glands located in the hand. (Katz & Gottman, 1986). Their behavior during this 5. General somatic activity (ACT). To measure interview was audiotaped. A script for this semistruc- somatic activity, the participant's chair was mounted tured interview is available from John M. Gottman or on a platform that was coupled to a rigid base in such Lynn Fainsilber Katz. a way as to allow an imperceptible amount of flexing. Parent- child interaction. The parent- child inter- Child intelligence. The Wechsler Preschool action session consisted of a modification of two Scales of Intelligence (WPPSI; Wechsler, 1974) procedures used by P. A. Cowan and C. P. Cowan Block Design, Picture Completion, and Information (1987). In the first task, parents were asked to obtain subscales were administered to each child. information from their child. The parents were in- formed that the child had heard a story, and they were Measure and Coding asked to find out what the story was. The story that the children heard did not follow normal story gram- Meta-emotion coding system. The audiotapes of mar and was read in a monotone voice; thus, the story the meta-emotion interview were coded using a spe- META-EMOTION 251 cific checklist rating system that codes for parents' Intrusiveness, which involved physical interference awareness of their own anger and sadness, their own with the child's actions (e.g., grabbing the joy stick). regulation of anger and sadness, and their acceptance The KACS also consists of seven parental affect and assistance (coaching) with their child's anger and codes. The three positive affect codes are as follows: sadness (Hooven, 1994). For each dimension, the (a) Affection, which consisted of praise and physical coding manual was quite detailed and specific. The affection; (b) Enthusiasm, which was coded as prais- Awareness score was a sum of 12 subscales: experi- ing and excitement at the child's performance; and encing the emotion; being able to distinguish the (c) Humor, which involved parental laughter or jok- emotion from others; having various experiences ing. The three negative affect codes are as follows: with the emotion; being descriptive of the experience (a) Criticism, which involved direct disparaging of the emotion; being descriptive of the physical comments or put-downs of the child's behavior or sensations connected with this emotion; being de- performance; (b) Anger, in which parents were visi- scriptive of the cognitive processes connected with bly frustrated by the child's actions or demonstrated this emotion; providing a descriptive anecdote; disappointment, annoyance, or irritation toward the knowing the causes of the emotion; being aware of child; (c) Derisive Humor, in which parents used remediation processes; answering questions about the humor at the child's expense (e.g., through sarcasm emotion easily, without hesitation or confusion; talk- or by making fun of the child). ing at length about this emotion; and showing interest Parent-child interaction was coded continuously in or excitement about this emotion. Coaching was a real time with coding synchronized to the original sum of 11 scales: showing respect for the child's parent-child interaction. The total number of times experience of the emotion, talking about the situation each variable occurred in the 10-min parent-child and the emotion when the child is upset, intervening interaction session was recorded and totals across in situations that give rise to the emotion, comforting time were calculated for each of the 14 parent-child the child, teaching the child rules for appropriate interaction variables. This index was therefore an expression of the emotion, educating the child about estimate of the frequency of the parenting behavior the nature of this emotion, teaching the child strate- within a 10-min period. Independent observers coded gies to soothe the child's own emotion, being in- mothers and fathers. Engagement and affect dimen- volved in the child's experience of the emotion, feel- sions were also coded by independent observers. ing confident about how to deal with this emotion, Reliability was calculated across coders using a cor- having given thought and energy to the emotion and relation coefficient. Because total number of seconds what one wants the child to know about this emotion within each parent code was the variable computed (goals), and using strategies that are age and situation and used in all data analyses, the appropriate reliabil- appropriate. Because coders used rating scales, the ity statistic was a correlation coefficient, rather than appropriate statistics to use for computing interrater Cohen's kappa or percentage agreement. For the reliability were correlations between independent ob- KECS, the mean correlation was .96, with a range of servers for each scale, rather than Cohen's kappa, .86 to .99; for the KACS, the mean correlation was which is appropriate for categorical data. The range .93, with a range of .84 to .97. We computed the sum of interobserver reliabilities for the awareness and of derisive humor, intrusiveness, and criticism for coaching scales was 0.73 to 0.86. both parents to form our derogation variable. The Observational coding of parent-child interaction. Kahen systems were also used to measure the Parenting was coded using the Cowans' Observa- scaffolding-praising dimension, which consisted of tional System, the Kahen Engagement Coding Sys- parental affection, engagement, positive structuring, tem (KECS), and the Kahen Affect Coding Systems responsiveness, and enthusiasm; we computed the (KACS; Gottman, in press). The KECS consists of sum of these variables across parents. Although it is seven parental engagement codes, including three certainly reasonable to expect to find differential positive, three negative, and one neutral code. The effects of mothers and fathers on children and we three Kahen positive engagement codes are as fol- have evidence that maternal and paternal parenting lows: (a) Engaged, which consisted of parental atten- were uncorrelated (the correlation between mother tion toward the child; (b) Positive Directiveness, in and father derogation was .21 and was .12 between which parents issued a directive statement that began mother and father scaffolding-praising), we summed in a positive way (e.g., "move to your right"); and (c) across parents* scores for the sake of economy. The Responds to Child's Needs, in which parents re- scaffolding-praising dimension differs from Baum- sponded to a child's question or complaint. The three rind's authoritative parenting in that it includes a negative engagement codes are as follows: (a) Dis- responsive and enthusiastic parenting style; in the engaged, in which parents were not attending to the teaching task, this was reflected in parents' (a) effec- child; (b) Negative Directiveness, in which parents tively structuring and scaffolding the child's learning issued a directive statement that began in a negative and (b) generally waiting until the child did some- way (e.g., "Don't move around so much"); and (c) thing right and men praising enthusiastically, rather 252 GOTTMAN, KATZ, AND HOOVEN

than waiting until the child made a mistake and then interval at baseline, interbeat interval variability (a being critical. measure of vagal tone; Izard, Porges, Simons, & The parents' behavior during the parent-child in- Haynes, 1991), reactivity of heart rate variability teraction was also coded using P. A. Cowan and C. P. from baseline to the mean of the parent-child inter- Cowan's (1987) coding system. This coding system action (used as an index of the child's ability to codes parents behavior on dimensions of warmth- modulate vagal tone by DiPietro, Porges, & Uhly, coldness, presence or lack of structure and limit set- 1992), and mean skin conductance level during base- ting, whether or not parents back down when their line (first visit to the lab) were also computed as child is noncompliant, anger and displeasure, unre- indexes of the amount of the child's chronic physio- sponsiveness or responsiveness, and whether or not logical arousal and PNS functioning. parents make maturity demands of their child. The behavior of parents toward each other during their interactions with their child (their coparenting) was also coded on dimensions of warmth, cooperation or Time 2 competition, anger, disagreement, responsiveness, pleasure in coparenting, clarity of communication, Assessment and amount of interaction. For the purposes of this study, only the warmth dimension (parenting and Overview. Time 2 assessment consisted of coparenting) was of interest. For the parenting di- teacher ratings of child outcomes and couple's re- mension, coders rated the overall degree of warmth ports of considerations of marital dissolution. Fami- and the highest level of warmth and coldness exhib- lies were recontacted 3 years later for follow-up ited by each parent. For the coparenting dimension, assessments of child and marital outcomes. Children coders rated the overall degree of warmth and the were, on average, 8 years old (M = 96.9 months; highest level of warmth and coldness exhibited by the range = 82-110 months). Ninety-five percent (53 out couple toward each other. Warmth was defined as the of 56) of the families in the initial sample and 86% sum of all the warmth variables minus the sum of all (48 out of 56) of the children's teachers at follow-up the coldness variables. Interreliability for the warmth agreed to participate in the Time 2 assessments. variable was .64. Ratings of children's behavior problems. Teachers completed the Child Adaptive Behavior Child regulatory physiology. An estimate of the Inventory (CABI; P. A. Cowan & C. P. Cowan, child's baseline vagal tone was taken when the child 1990). We used the CABI as a measure of child was listening to the introduction to an interesting outcomes for two reasons. First, the CABI was de- story taken from an animated cartoon film (clip from veloped on a normal sample and contains subscales Charlotte's Web), a variable we called basal vagal. that are less pathological in nature than the Child The child's ability to withdraw vagal tone was esti- Behavior Checklist (Achenbach & Edelbrock, 1986). mated as a difference between this estimate of basal Second, the CABI also controls for teacher rating vagal tone and the child's vagal tone during an ex- bias by having teachers complete the scale on all citing film clip designed to elicit strong emotion (the same-sexed children in the classroom and deriving z scene from The Wizard of Oz when the flying mon- scores for the target child. The CABI has good inter- keys kidnap Dorothy). We expected vagal tone to be nal consistency (average alpha = .81; range = .66 to withdrawn and heart rate to increase when the child .90) and predictive validity (P. A. Cowan & Cowan, was emotionally engaged with the fearful stimuli in 1990). The factors and subscales that comprise the this second film clip. We called this second variable CABI include the following: (a) the Antisocial Fac- delta vagal. This second variable indexed the child's tor, which consists of the Hyperactivity, Antisocial ability to suppress vagal tone when engaging with a Behavior, Negative Engagement with Peers, Hostil- strong emotional stimulus that included an environ- ity, Fairness-Responsibility (keyed negatively), mental demand for changing attentional focus or reg- Calm Response to Challenge (keyed negatively), and ulating emotion. In this case, the engagement with Kindness-Empathy (keyed negatively) subscales; the environment involved the demands for an emo- and (b) the Internalizing Factor, which consists of the tional response being elicited by the emotional film Introversion, Depression, Victim-Rejected, Tension, as well as the demands to focus attention on the Atari and Extroversion (keyed negatively) subscales. videogame the child played immediately after each Differential Emotions Scale. We used the Differ- film clip. The index of vagal tone was computed as ential Emotions Scale (Izard, 1982) as a measure of the amount of variance in the cardiac interbeat inter- temperament, given Goldsmith and Campos' (1982) val spectrum that was within the child's respiratory definition of temperament in terms of affect expres- range using spectral time-series analysis. This index sion (for a review and critique, see Bates, 1987). The of vagal tone measures respiratory sinus arrhythmia, Differential Emotions Scale is a 36-item question- a measure of PNS tonus, which has been found to naire that mothers complete with regard to the fre- index attentional processes and emotion regulation quency with which they have observed their children abilities (Porges, 1984). Mean levels of interbeat display specific emotions in the past week. Each item META-EMOTION 253 is rated on a 5-point scale ranging from 1 (rarely or ment at Time 2, even controlling for Time 1 never) to 5 (very often). We computed the total child intelligence. number of emotions and the total number of positive and negative emotions for the week. Teacher ratings of peer aggression. Teachers Selection and Validity of Key also completed the Dodge Peer Aggression Scale Theoretical Variables (Dodge, 1986). The Dodge Peer Aggression scale consists of items that measure the degree to which the Reduced Set of Meta-Emotion Variables child uses overt aggression with peers. Child academic achievement. Children were in- In the interest of parsimony, we needed to cut dividually administered the Peabody Individual down the choice of variables for the modeling, Achievement Test—Revised (PIAT-R) as a measure and it was thus necessary to limit the number of of academic achievement (e.g., see Costenbader & meta-emotion variables. We started with 12 Adams, 1991). They were administered the mathe- matics, reading recognition, reading comprehension, variables (awareness of own emotion, aware- and general information tests. ness of the child's emotion, and coaching, for Emotion regulation questionnaire. Mothers filled father and mother and for sadness and anger). out a newly developed 45-item questionnaire about Two variables from this set of 12 were con- the degree to which their child requires external structed, one of which was the sum of parental regulation of emotion (Katz & Gottman, 1986). This awareness of the parents' own emotions and the questionnaire includes items that reflect instances sum of the parents' awareness of the child's when the parent needs to down regulate the child to emotions and the other of which was the sum of reduce the child's level of activity, inappropriate the parental coaching of the child's emotions. behavior, and misconduct. The alpha coefficient for Table 1 gives a summary of the correlations of the scale was .74. these two summary meta-emotion variables Child physical health. Child illness was assessed (i.e., awareness and coaching) with the original by parental report using a version of the Rand Cor- 12 meta-emotion variables that were obtained. poration Health Insurance Study measures (available from John M. Gottman or Lynn Fainsilber Katz; see Therefore, it is a presentation of item-total cor- Gottman & Katz, 1989). The following Likert or relations that demonstrate the internal construct true-false items were summed: "In general, would validity of the summary codes. These correla- you say that this child's health is excellent, good, fair tions show that the two variables we constructed or poor?," "The child's health is excellent," "The are related to all the individual awareness and child seems to resist illness very well," "When some- coaching variables. thing is going around this child usually catches it," "This child has had a nosebleed in the past 30 days" (alpha = 0.82). Table 1 Results Correlations of the Two Summary Meta- Emotion Variables With Original 12 Variables We begin by discussing the selection and Original meta- validity of variables used for building the theo- emotion variables Awareness Coaching retical model. The reduced set of variables used to index meta-emotion philosophy are pre- Father sadness Awareness own .80*** .55*** sented, and we address the construct validity of Awareness child .68*** .37** the derogatory parenting and scaffolding- Coaching .26* .63*** praising variables by indicating that derogatory Father anger parenting is not anger and that scaffolding- Awareness own .75*** .33** praising parenting is not warmth. We then Awareness child .69*** 49*** Coaching 44*** j4*** present the results of our path-analytic model- Mother sadness ing, looking separately at models related to pa- Awareness own .56*** .32** rental derogation and to parental scaffolding- Awareness child .66*** .63*** praising. We also test the temperament Coaching .48*** .72*** hypothesis to see whether parents select their Mother anger Awareness own .57*** .29* parenting style to be consistent with the child's Awareness child .64*** .36** behavior. Finally, we review the result that Coaching .37** .66*** meta-emotion predicts child academic achieve- *p<.05. **p<.0\. ***p< .001. 254 GOTTMAN, KATZ, AND HOOVEN

Validity of Child Regulatory Physiology anger is not predictive of marital dissolution, whereas contempt is (Gottman, 1993, 1994; To examine the validity of the vagal tone Gottman & Krokoff, 1989; Gottman & Leven- variables, we correlated baseline vagal tone and son, 1992). Ginott (1965) distinguished anger suppression of vagal tone with other indexes of versus emotional interactions with the child that autonomic arousal relating to parasympathetic communicated contempt for the child, global activation and general nervous system arousal. versus specific criticism, and criticism versus As Table 2 shows, baseline vagal tone and blaming or communications suggesting that the suppression of vagal tone were related to the child was incompetent. He suggested that the child's mean heart rate during the parent-child latter communications are harmful to a child, interaction, to the child's lower resting heart rate whereas anger is not and may even be healthy. during the first visit to the laboratory when it was To deal with a tendency to overgeneralize and a mildly stressful and novel situation, to lower conclude that anger is harmful for children, we baseline skin conductance during the second selected the parental derogation codes from the visit to the laboratory after initial adaptation, to Kahen Coding Systems, which were designed to higher heart rate variability, and to lower heart rate measure Ginott's cluster of negative parenting; reactivity in the parent-child interaction. derogation is not the same as anger parents express toward children. We tested this specif- ically by looking at the relationships between Validity of Parenting Variables parental anger and derogation. Parental anger was uncorrelated with our measure of parental Derogatory parenting is not anger. In de- derogation (for father anger, r = —.11; for velopmental psychology today, there is a gen- mother anger, r = — .06), uncorrelated with the eral equation of anger and negativity (see Cum- meta-emotion codes, and uncorrelated with neg- mings & Davies, 1994), which we consider ative child outcomes. unfortunate. When the construct of negative af- fect is broken into specific negative emotions, Scaffolding-praising is not warmth. It is there is evidence that specific negative emotions important to emphasize that scaffolding- serve different functions. In marital interaction, praising is not merely a dimension of global positivity, such as is tapped by the Cowan pa- rental warmth variable. Correlations computed between the warmth codes from the Cowan Table 2 system and the scaffolding-praising code indi- Validity of the Two Child Physiology cated that only maternal warmth was related to Variables Selected for Model Building our scaffolding-praising variable (r = .32, p < Variable Base vagal Delta vagal .05). Paternal warmth, maternal and paternal Mean heart rate coldness, and co-warmth all showed no relation during parent- to scaffolding-praising. child interaction -.26* -.40** Mean heart rate during first Summary of the Seven Dimensions of the time in the Theoretical Model laboratory -.28* -.33** Heart rate Aside from the outcome variables, there are reactivity in seven dimensions of the theoretical model. The parent-child interaction -.24* -.30* two meta-emotion variables were derived from Heart rate the meta-emotion coding system used to code variability .25* .39** the meta-emotion interview: coaching (a sum of Skin conductance 11 scales) and awareness (a sum of 12 sub- level after adaptation to scales). The two parenting dimensions were ex- laboratory -.37** -.22 tracted from the Kahen observational scales, Note. Basal vagal = child's baseline vagal tone; scaffolding-praising (sum of three scales of the delta vagal = child's vagal tone when emotionally KECS—engagement, positive directiveness, engaged. and responsiveness to the child's needs—and * p < .05. ** p < .01. two scales of positive affect of the KACS— META-EMOTION 255 affection-praise and enthusiasm) and deroga- fect these physiological variables. That is, we tion (the intrusiveness code of negative engage- fundamentally believed that these physiological ment on the KECS and criticism and derisive variables are not entirely engraved in stone, humor of the KACS negative affect codes). The even if they are biological; rather, we believed regulatory physiology dimensions were basal that these variables are malleable and shaped in vagal tone (when the child was listening to the part by parents through their emotional interac- introduction to an interesting story taken from tions with the child. These conceptual pathways an animated cartoon film, Charlotte's Web) and enable us to assess whether meta-emotion phi- the child's ability to suppress vagal tone, which losophy and parenting in some way are related was estimated as a difference between this es- to the child's regulatory physiology, or, con- timate of basal vagal tone and the child's vagal versely, if the child's regulatory physiology is tone during an exciting film clip designed to related to meta-emotion philosophy or parent- elicit strong emotion (the scene from the Wizard ing. We also predicted that there would be sta- of Oz when the flying monkeys kidnap Dor- tistically significant path coefficients connect- othy). The emotion regulation dimension was ing the child physiology variables to the child taken from a questionnaire mothers completed outcome variables (Path 4), suggesting that the (when the child was 8 years old) about the child physiology at age 5 years predicts child degree to which their child required external outcome at age 8 years (Path 7). We also pre- down regulation of emotion. dicted the physiological variables would predict the emotion regulation variable, and that the Building the Theoretical Model regulation and the parenting variables would relate to child outcome (Paths 5 and 8). We The basic template for the theoretical model predicted that parenting would have a direct is presented in Figure 2. There are eight con- effect on child regulation (Path 6). There may ceptual pathways consistent with our theoretical be direct effects between the meta-emotion vari- formulation and a direct pathway from meta- ables and the child outcome variables, but, to the emotion to child outcome. First, we predicted extent that this is true, we have not completely that there would be statistically significant path succeeded in our theory, because we will not have coefficients from the meta-emotion variables to had a mechanism to explain these effects. the parenting variable (Path 1). This was a va- In the interest of data reduction, we created lidity test to see if the variables derived from the following four child outcome variables: our interview related to actual parenting behav- 1. Child achievement was the sum of the ior. Second, we tested the significance of the mathematics and reading comprehension paths from meta-emotion and parenting to the scores. physiological variables (Paths 2 and 3). We 2. Child's emotion regulation abilities was hypothesized that meta-emotion, operating (in assessed with the Down-Regulation scale of the part) through parenting, would significantly af- Katz-Gottman Emotional Regulation Question-

Child Outcomes

Regulation Delta Vagal

Physiology

Basal Vagal

Figure 2. Outline of the general expected path structure of the theoretical model. 256 GOTTMAN, KATZ, AND HOOVEN naire, assessed when the children were 8 years ings, R = 0.62; for child physical illness, R = old (called down regulation). Because this is 0.60. The figures present the path coefficients, technically an outcome variable and not a pro- with z scores for each coefficient in parentheses. cess variable (because it was measured when As can be seen from these figures, the model the children were 8 years old), the emotion building using our theory was generally suc- regulation variable appears in every child out- cessful. We were able to find linkages for the come model. major pathways we proposed between meta- 3. Child's peer relations was the sum of three emotion and parenting, between meta-emotion teacher rating scales: the CABI Negative Peer and the physiological variables, between parent- scale, the CABI Antisocial Scale, and the ing and the physiological variables for child Dodge Peer Aggression scale. peer relations, between physiology and emotion 4. Child illness was our health measure at regulation, between emotion regulation and age 8. child outcome, and between parenting and child We built two sets of theoretical models. outcome. For child illness, we suggested that Given the hypotheses that meta-emotion philos- basal vagal tone should relate to less physical ophy might either be related to the inhibition of illness, and it turned out that the significant parental derogation or the facilitation of paths to child illness were parental emotion scaffolding-praising parenting, we constructed coaching and basal vagal tone. separate models to examine hypotheses related In all models, either awareness or coaching of to derogation and scaffolding-praising. All the child's emotions was negatively related to modeling computations were performed using the negative parenting variable, supporting the Bentler's (1992) computer program, EQS. hypothesis that awareness or coaching is an inhibitor of parental derogation of the child. It Modeling With Derogatory Parenting was interesting that in all models the child's ability to suppress vagal tone at age 5 was a The results of our model building are pre- significant predictor of the child's emotion reg- sented in Figures 3, 4, and 5. The following are ulation at age 8. The greater the child's ability our goodness-of-fit statistics: for the academic to suppress vagal tone at age 5, the less the achievement outcome variable, ^(14, N = parents had to down regulate the child's nega- 56) = 13.68, p = .474, Bentler-Bonett normed tive affects, inappropriate behavior, and overex- fit index (BBN) = .981; for the peer relations citement at age 8. outcome variable, ^(13, N = 56) = 17.95,/? = .159, BBN = .986; for child illness, ^(12, N = Modeling With Scaffolding-Praising 56) = 13.04, p = .366, BBN = .987. Hence, all Parenting models fit the data. Multiple Rs for the outcome variables varied: for child academic achieve- We tested whether the same equations we had ment, R = 0.41; for negative peer teacher rat- used for negative parenting would fit when the

, Awareness

.73 (6.89)

Figure 3. Path model for child academic achievement with parental derogation. Numbers in parentheses are z scores. META-EMOTION 257

.48 (4.20)

(Awareness) .45 (3.65) Child Peer .73 (6.68) ( Derogation Relations -.47 (-3.72) A .31 (2.48) -.47 (-3.65) >( Down Reguate )

v .61 (5.53)

Figure 4. Path model for child peer relations (teacher ratings) with parental derogation. Numbers in parentheses are z scores. parental derogation variable was replaced with the coaching. For teacher ratings of peer interac- scaffolding-praising parenting variable. We ex- tion, the scaffolding-praising variable was not pected that some of the path coefficients would directly related to the outcome, but it was sig- change, but we first tested whether or not a similar nificantly related to coaching. For child illness, model would fit the data. For the academic the scaffolding-praising variable was unrelated achievement outcome variable, a very similar to the outcome, but it was significantly related model fit the data, ^(14, AT = 56) = 13.12, to awareness. p = .517, BBN = .986; for the peer relations outcome variable, ^(15, N = 56) = 24.14, p = .063, BBN = .978; for child illness, ^(15, Direction of Effects Between Emotion N = 56) = 18.82, p = .222, BBN = .981. These Coaching and Child Regulatory results are presented as Figures 6, 7, and 8. Physiology: Testing the Temperament Was scaffolding-praising related to the vari- Hypothesis ables of the model? For child academic achieve- ment, the scaffolding-praising variable was sig- Although our path models present data sup- nificantly related to the outcome and to porting the notion that emotion coaching can

-.43 (-3.17)

Figure 5. Path model for child illness with parental derogation. Numbers in parentheses are z scores. 258 GOTTMAN, KATZ, AND HOOVEN

Figure 6. Path model for child academic achievement with positive parenting in the model. Numbers in parentheses are z scores. SCAF-PR = scaffolding-praising. affect a child's vagal tone, it is possible that the reversed. First, we considered the models with direction of effects may be reversed. That is, it derogatory parenting. For the child outcome of is quite possible that either (a) child physiology negative peer relations, ;^(13, N = 56) = is a part of child temperament, and parents may 19.32, p = .113, BBN = .985, and the path select parenting style to be consistent with in- coefficient from basal vagal tone to emotion dividual differences in child behavior or (b) coaching was .37, z = 3.62. For the child out- emotion coaching changes a child's vagal tone. come of child achievement, )^(14, N = 56) = Perhaps parents are more likely to do emotion 11.28, p = .664, BBN = .989, and the path coaching with a child higher in vagal tone, or coefficient from basal vagal tone to emotion perhaps emotion coaching can affect a child's coaching was .35, z = 3.74. For the child out- basal vagal tone. Although we could not answer come of child illness, ^(12, N = 56) = 13.36, causal questions in our path modeling, we did p = .344, BBN = .986, and the path coefficient conduct additional analyses to see if results from basal vagal tone to emotion coaching was either were consistent with or disconfirmed the .33, z = 3.05. Next, we considered the models temperament hypothesis. with scaffolding-praising. For the child out- We tested the hypothesis that child vagal tone come of academic achievement, ;^(14, N = might affect emotion coaching by reversing the 56) = 12.77, p = .545, BBN = .986, and the arrow between these two variables. The models path coefficient from basal vagal tone to emo- fit just as well with the direction of effects tion coaching was .34, z = 3.35. For the child

42 (-3.35) Delta Vagal ) >( Down Reguate

Figure 7. Path model for child peer relations (teacher ratings) with positive parenting in the model. Numbers in parentheses are z scores. SCAF-PR = scaffolding-praising. META-EMOTION 259

A -.53 H.02)

.02 (.13) Child Illness "A

.43(3.18)

Figure 8. Path model for child illness with positive parenting in the model. Numbers in paren- theses are z scores. SCAF-PR = scaffolding-praising. outcome of negative peer relations, ^(15, N = .001). Hence, it appears that the direct benefits 56) = 19.82, /> = .179, BBN = .982, and the of coaching are unaffected by the child's basal path coefficient from basal vagal tone to emo- vagal tone. tion coaching was .40, z = 3.69. For the child outcome of child illness, ^(14, N = 56) = 19.77, p = .138, BBN = .982, and the path Meta-Emotion Predicts Child Academic coefficient from basal vagal tone to emotion Achievement at Time 2, Controlling for coaching was .37, z = 3.64. Thus, our modeling Time 1 Child Intelligence can rule out neither a direction for effects, nor the possibility that the effects are bidirectional. If, as we hypothesized, the meta-emotion Given recent theorizing that vagal tone is a variables affect school achievement through temperament dimension (Porges, Doussard- emotion regulation, we should expect that the Roosevelt, Portales, & Suers, 1994), one con- relationships between the meta-emotion vari- cern with these results concerns the direction of ables and the child's achievement at age 8 will effects; the question is whether parents are hold, even controlling the child's Time 1 IQ (at coaching their children differentially as a func- age 5). To test this hypothesis, we performed a tion of the children's temperament. To further regression analysis forcing in the three IQ scales test this hypothesis, we correlated coaching (WPPSI Block Design, Picture Completion, and with our temperament measures from the Dif- Information Scale scores) before entering the ferential Emotions Scale. Coaching was uncor- mother's awareness of her own sadness in pre- related with the amount of child negative affect dicting the child's math scores and, in a second (r = .02, ns), the amount of child positive affect analysis, before entering the father's coaching (r = .20, ns), and the amount of child total of the child's anger. The F tests for change were affect (r = .16, ns). We were also concerned computed as well as the partial correlations. For that the direct effects of coaching on child out- the prediction of the child's mathematics scores comes might be qualified as a function of the from the mother's awareness of her own sad- child's basal vagal tone. Coaching was signifi- ness, F(4, 48) = 6.12, p < .05 (partial correla- cantly directly correlated with only one of the tion = 0.34). For the prediction of the child's three outcomes, the child's Time 2 physical reading comprehension scores from the father's illness (r = -.55, p < .001). The child's basal coaching of the child's anger, F(4, 44) = 9.41, vagal tone was also significantly correlated with p < .01 (partial correlation = 0.37). For the this outcome (r = .30, p < .05); however, the prediction of the child's total achievement score partial correlation between coaching and the from both the mother's awareness of her own child's physical illness, controlling basal vagal sadness and the father's coaching of the child's tone, remained significant (r = -.56, p < anger, the two variables were summed for the 260 GOTTMAN, KATZ, AND HOOVEN

analysis, F(4, 45) = 4.13, p < .05 (partial child's negative child peer relations. There were correlation = 0.29). direct links between child physiology and child outcome only for child illness. We note, in Summary passing, that in path analysis, a path coefficient is a partial correlation; thus, for example, the Let us summarize the results of our modeling pathway between delta vagal and down regulate in Figure 9 with respect to the eight conceptual partials out basal vagal tone. For this reason, we paths numbered in Figure 2. We recognize that speak of these pathways in terms of the regula- our models are probably not independent of one tory physiology variables instead of in terms of another. single variables. Path 3: Parenting and child physiology. Conceptual Paths This hypothesized pathway was not supported by any of the models. Path 1: Meta-emotion and parenting. Meta- Path 4: Parenting and emotional down reg- emotion was related to both the inhibition of ulation. This pathway was not supported by parental derogation of the child and to the fa- any of the models. cilitation of scaffolding-praising parenting. Path 5: Parenting and child outcome. This Path 2: Meta-emotion and child physiology. pathway was supported in three of the six mod- For models including the derogation parenting els: for derogatory parenting, there was a direct variable, coaching was directly and signifi- link with child academic achievement and child cantly related in all the models (achievement, peer relations, whereas for scaffolding-praising ratings of child peer relations, and child health) parenting, there was a direct link only with to the child's physiology. This linkage was academic achievement. also evident in the models that included the Path 6: Child physiology and child outcome. scaffolding-praising parenting variable for the This linkage was supported only for child ill- achievement and peer relations models (margin- ness and only in the derogation model. ally for the child health model, z = 1.80). This Path 7: Child physiology and later emotional suggests the intriguing hypothesis that perhaps down regulation. This pathway, which was parents can influence a child's physiology by from the 5-year-old suppression of vagal tone emotion coaching; however, we cannot differ- variable to the 8-year-old emotional down reg- entiate directionality of effects without an ex- ulation variable, was supported in all of the periment. As we posited, the child's physiology models. was significantly related to parental emotion Path 8: Emotional down regulation and child regulation ratings, which, in turn, were related outcome. This linkage was supported in two of to child outcomes in the models for derogatory the six models, for models predicting peer rela- parenting and scaffolding-praising and the tions.

Delta Vagal "— - Physiology ,- Basal Vagal

Figure 9. Revision of the path model derived from our theory. META-EMOTION 261

Unexpected Direct Pathway child's regulatory physiology and behavior are related to child outcomes. In general, aside from In two models for derogatory parenting two direct connections (from awareness to peer (teacher ratings of peer relations and child ill- relations and from coaching to child illness), ness) and in one model for scaffolding-praising there are two pathways through the models that (child illness), there was a direct pathway from predict child outcomes: one from the meta- either awareness or coaching of the child's emo- emotion variables through child physiology and tions and the child outcome. For one of these emotion regulation and one from the meta- models (derogatory parenting: child peer rela- emotion variables through parenting. tions), more awareness predicted more negative The pathway from meta-emotion to the ratings, opposite to what we might have pre- child's regulatory physiology suggests the hy- dicted, whereas for child illness, more coaching pothesis that coaching the child's emotions has of the child's emotions was related to less child a soothing effect on the child that may change illness, which we predicted. One possible ex- some key aspects of the child's PNS. If this planation for the unexpected linkage is that at were true, then the child's ability to self-soothe, times parental awareness of emotion is not such to regulate negative emotion, and to focus at- a good thing for children. Perhaps being attuned tention may be a result of temperament and may to a child's negative emotion fosters its expres- also be affected by the parents' meta-emotion sion, and even teachers pick up on this when the philosophy. However, the data also support an child is 8 years old and rate it negatively. As interpretation that suggests that parents may be part of this explanation, perhaps the arrow of causation should be reversed, meaning that par- 4 ents have heightened awareness of their child's Because the arrow for awareness has coaching negative emotions because the child is highly coming into it and, as a variable, a line from aware- ness represents a partial correlation of awareness, expressive of them. With this possibility in controlling for coaching, we wondered if the problem mind, we attempted to fit path models with the arose because some families are high in awareness arrow reversed for the three models for which but do not coach. We split the families who were awareness was positively related to a negative above the median in awareness into two groups: child outcome. The results of these post hoc Group 2 included those above the median (N = 18), analyses were as follows. For the derogatory and Group 1 included those below the median (N = parenting models, with the child peer relations 8) in coaching. Thus, Group 1 represented 8 families outcome, the new model did not fit the data, high in awareness but low in coaching, whereas ^(13, N = 56) = 48.60, p < .001. Hence, this Group 2 represented 18 families high in both aware- hypothesis is not consistent with our data.4 The ness and coaching. Group 1 had children whose play quality at age 5 was significantly lower than Group 2, revised figure (Figure 9) places the Time 1 F(l, 24) = 4.29, p < .05, Group 1 M = 14.13, Group meta-emotion variables as the exogenous, or 2 M = 16.00. There were also marginal effects at age driving, variables in the model, having their 5 in the observed negative peer interaction with a best purported causal effects on parenting, which friend; in noncompliance, F(l, 24) = 3.05, p = .093, predicts 8-year-old child outcome, and on child Group 1 = 2.75, Group 2 = .83; and in crying, F(l, physiology, which has its major effects on out- 24) = 3.74, p = .065, Group 1 = .63, Group 2 = .00. come through the child's emotion regulation Also, when the children were 8 years old, teachers abilities at age 8. These results are consistent used the Dodge Rating Scale and rated the children with the theory we proposed. on three scales of peer relations. The teachers rated the children in Group 1 significantly lower in the Dodge Social Skills scale, F(l, 25) = 7.51, p < .05, Group 1 = -3.80, Group 2 = -0.21; higher in the Dodge Aggression Scale, F(l, 25) = 4.56, p < .05, Discussion Group 1 = -3.29, Group 2 = -0.40; and lower in the Dodge scale of Overall Peer Relations, F(l, The preliminary results of our investigation 21) = 5.83, p < .05, Group 1 = -2.21, Group 2 = of meta-emotion are encouraging. The data sug- 0.21 than Group 2. These post hoc results support the gest that meta-emotion is related to parenting hypothesis that the negative linkage between aware- behavior and to child regulatory physiology; ness and negative child outcomes may be attribut- that child regulatory physiology at age 5 pre- able, in part, to those families who are aware of their dicts emotion down regulation ability at age 8; child's emotion and perhaps are even accepting of it, and that parenting, meta-emotion, and the but who do not emotion coach their children. 262 GOTTMAN, KATZ, AND HOOVEN selecting emotion coaching as a good strategy Fox (Davidson, 1984; Davidson, Ekman, Saron, of parenting with children who have higher basal Senulis, & Friesen, 1990; Davidson & Fox, vagal tone. However, parents are not differentially 1982; Davidson, Schaffer, & Saron, 1985; Dav- coaching their children as a function of our mea- idson & Tomarken, 1990; Fox & Davidson, sures of their temperaments. The truth probably 1987, 1988), we can reason that talking about lies in a combination of these two directions of negative emotions while having the emotions influence, or perhaps the influence is bidirectional; might entail changing a right frontal-limbic- an experimental test of the effects of training par- autonomic experience of emotion into a left ents in emotion coaching on basal vagal tone frontal dominant experience in which the left would help sort out these alternatives. We have frontal lobe assumes control over the right fron- now developed and pilot tested a parent training tal lobe and its limbic and autonomic connec- intervention in emotion coaching. tions. The positive affects, language, and anger The social skills that are related to peer social are lateralized in the left frontal lobe, and Dav- competence in middle childhood are not the idson called these left-lateralized emotions ap- same as the skills that emotion-coaching parents proach emotions. Anger is considered an ap- are building in their children. It has been well proach emotion because it generally tends to established that calling attention to oneself and engage people with the world rather than lead- one's feelings are precisely the opposite of what ing to disengagement and withdrawal. This may socially competent children do in a variety of be, in part, the reason that many clinical inter- critical peer situations in middle childhood (for ventions suggest a healing process in which data on peer entry, see Putallaz & Gottman, depression changes into anger. The other nega- 1981, and Asher & Gottman, 1981; for data on tive affects, such as fear, disgust, and sadness, teasing, see Gottman & Parker, 1986, and Asher are lateralized in the left frontal lobe; Davidson & Coie, 1990). They do not express their emo- called these withdrawal affects. tions. On the contrary, they act cool and unruf- We posit that talking about negative with- fled, as if they have had an "emotion-ectomy." drawal emotions while having them entails a The point here is that children whose parents greater sense of approach rather than with- were emotion coaches at age 5 are acting in a drawal, a greater sense of control of the nega- way that teachers describe as socially competent tive emotions, and greater parasympathetic con- at age 8. Hence, they are skillful enough to trol of autonomic reactions. These hypotheses know what to do to be competent with peers at suggest to us a series of studies using concom- age 8, and these social skills are not at all the itant electroencephalographic (EEG) and auto- same as what they learned from their parents at nomic measurement while children are either age 5. Apparently, emotion-coached children experiencing a negative emotion, being dis- have learned to be savvy about peer social sit- tracted from an emotional response they are uations, and they can do what is called for. They having (similar to what dismissing parents do), may have a heightened sense of awareness of or talking about the emotions while having their own emotions, a better ability to self- them (similar to what coaching parents do). The regulate their own upset, and a greater ability to conditions in these experiments would involve a attend to the salient aspects of any challenging systematic dismantling of the emotion-coaching peer situation. This sense of heightened aware- parents' behaviors; this group would be com- ness may lead them to be more likely to know pared to appropriate controls for rival hypothe- what is called for and to do it. This interpreta- ses and to a dismantling of the counterpart of tion of our results suggests that what children the coaching parent, reflecting a meta-emotion learn from emotion coaching is not at all philosophy we call dismissing. These studies modeling-specific social skills. It is far more are currently underway in our laboratory, but likely that what they have acquired are the tools we do not hesitate to point out that these hy- to learn how to learn in emotionally challenging potheses are highly speculative. situations, even if that calls for inhibiting emo- tional responding. The results of this first study of meta-emotion are encouraging, but replication and extension From a theoretic standpoint, it is interesting to to a more diverse sample of families is clearly consider why and how parental coaching would needed. We are currently conducting a replica- affect the child's regulatory physiology. On the tion and extension study. Our initial sample was basis of the seminal work of Davidson and of not racially or ethnically diverse, and it was META-EMOTION 263 limited to families with two married parents. validation of metacognition. Journal of Psychol- Subsequent investigation needs to assess ogy, 127, 203-211. whether these results generalize to various kinds Asher, S. R., & Coie, J. D. (Eds.). (1990). Peer of single-parent families and to blended fami- rejection in childhood. New York: Cambridge Uni- lies. Replication is also important because of the versity Press. relatively small sample size of the present study Asher, S. R., & Gottman, J. M. (1981). The develop- and its possible effects on structural equations ment of children's friendships. New York: Cam- bridge University Press. modeling (Loehlin, 1992). Naturalistic samples Barbara, H. (Producer), Nicolas, C. A., & Takamoto, of parents interacting with their children during I. (Directors). (1973). Charlotte's web [film]. emotional moments are also needed to see how Paramount. emotion-coaching families actually talk to their Bates, J. E. (1980). The concept of difficult temper- children during times of heightened emotion. ament. Merrill-Palmer Quarterly, 26, 299-319. These samples should be obtained across vari- Bates, J. E. (1987). Temperament in infancy. In J. D. ous developmental levels of both the children Osofsky (Ed.), Handbook of infant development and the parents. We also need to assess the (pp. 1101-1149). New York: Wiley. stability of parental meta-emotion philosophy Baumrind, D. (1967). Child care practices anteceding over time. Changes to the meta-emotion inter- 3 patterns of preschool behavior. Genetic Psychol- view are also needed, including an expansion to ogy Monographs, 75, 43-88. include other emotions (particularly fear, guilt, Baumrind, D. (1971). Current patterns of parental and shame) and the development of a measure authority. Developmental Psychology Monograph, of child meta-emotion philosophy. 4(1, Pt. 2), 1-103. Becker, W. C. (1964). Consequences of different The second general pathway through our kinds of parental discipline. In M. L. Hoffman & L. models was from the meta-emotion variables W. Hoffman (Eds.), Review of child development through parenting. What this adds to the parent- research (Vol. 1, pp. 169-208). New York: Rus- ing literature is that perhaps parents' own basic sell Sage Foundation. feelings and thoughts about their emotions are Bentler, P. M. (1992). EQS: Structural equations strongly related to or underlie the way they program manual. Los Angeles: BMDP Statistical parent. The results of our modeling with longi- Software. tudinal correlational data suggest a series of Bulloch, K., & Moore, R. Y. (1981). Innervation of experiments that would further test the model. the thymus gland by brain stem and spinal cord in mouse and rat. American Journal of Anatomy, 162, Each arrow in the model suggests an interven- 157-66. tion for altering the variable at the head of the Bulloch K., & Pomerantz, W. (1984). Autonomic arrow by manipulating the variable at the foot of nervous system innervation of thymic-related lym- the arrow. Thus, parent training in meta- phoid tissue in wildtype and nude mice. Journal of emotion should affect two pathways: It should Comparative Neurology, 228, 57-68. alter both parenting and the child's regulatory Bvinelli, D. J. (1993). Reconstructing the evolution physiology, and it may also have a direct effect of mind. American Psychologist, 48, 493-509. on child outcome. We are now planning to Casey, R. J., & Fuller, L. J. (1994). Maternal regu- conduct this research. lation of children's emotions. Journal of Nonver- bal Behavior, 18, 57-89. Cassidy, J., Parke, R. D., & Butkovsky, C. (1992). Family correlates of childhood peer relations. References Child Development, 63, 603-618. Choi, S. (1993). Mothers' and older siblings' scaf- Achenbach, T. M, & Edelbrock, C. (1986). Manual folding of preschool children's pretend play. Dis- for the Teacher's Report Form and Teacher Ver- sertation Abstracts International, 5i(7-A), 2234- sion of the Child Behavior Profile. Burlington, VT: 2235. University of Vermont, Department of Psychiatry. Cohn, D. A., Cowan, P. A., Cowan, C. P., & Pearson, Ainsworth, M. D. S., Bell, S. M., & Stayton, D. J. J. (1992). Mothers' and fathers' working models of (1971). Individual differences in the strange situa- childhood attachment relationships, parenting tion behavior of one-year-olds. In H. R. Schaffer styles, and child behavior. Development & Psycho- (Ed.), The origins of human social relations (pp. pathology, 4, 417-431. 17-57). London: Academic Press. Corsaro, W. (1979). We're friends, right?: Children's Alcroft, B. (Producer), & Mitton, D. (Director). use of access rituals in a nursery school. Language (1984). Daisy [film]. Strand Home Video. in Society, 8, 315-336. Allen, B. A., & Armour, T. E. (1993). Construct Corsaro, W. (1981). Friendship in the nursery school: 264 GOTTMAN, KATZ, AND HOOVEN

Social organization in the peer environment. In S. DiPietro, J. A., Porges, S. W., & Uhly, B. (1992). Asher & J. M. Gottman (Eds.), The development of Reactivity and developmental competence in pre- children's friendships. New York: Cambridge term and full-term infants. Developmental Psy- University. chology, 28, 831-841. Costenbader, V. K., & Adams, J. W. (1991). A Dodge, K. A. (1986). A social information process- review of the psychometric and administrative ing model of social competence in children. In M. features of the PIAT-R: Implications for the Perlmutter (Ed.), Minnesota Symposium on Child practitioner. Journal of School Psychology, 29, Psychology (Vol. 18, pp. 77-125). Hillsdale, NJ: 219-228. Erlbaum. Cowan, C. P., & Cowan, P. A. (1992). When parents Dunn, J. (1977). Distress and comfort. Cambridge, become partners. New York: Basic Books. MA: Harvard University Press. Cowan, P. A., & Cowan, C. P. (1987, April). Cou- Dunn, J., & Brown, J. (1994). Affect expression in ple's relationships, parenting styles and the child's the family, children's understanding of emotions, development at three. Paper presented at the Soci- and their interactions with others. Merrill Palmer ety for Research in Child Development, Baltimore, Quarterly, 40, 120-137. MD. Eisenberg, N. A., & Fabes, R. A. (1990). Empathy: Cowan, P. A., & Cowan, C. P. (1990). Becoming a Conceptualization, measurement, and relation to family: Research and intervention. In I. Sigel & A. prosocial behavior. Motivation and Emotion, 14, Brody (Eds.), Methods of family research (pp. 131-149. 1-51). Hillsdale, NJ: Erlbaum. Eisenberg, N. A., & Strayer, J. (Eds.). (1987). Em- Cummings, E. M., & Davies, P. (1994). Children and pathy and its development. New York: Cambridge marital conflict: The impact of family dispute res- University Press. olution. New York: Guilford Press. Ekman, P., Friesen, W. V., & Simons, R. C. (1985). Davidson, R. J. (1994). Temperament, affective style, Is the startle reaction an emotion? Journal of Per- and frontal lobe asymmetry. In G. Dawson & K. sonality and Social Psychology, 49, 1416-1426. W. Fischer (Eds.), Human behavior and the devel- Flavell, J. H. (1979). Metacognition and cognitive oping brain (pp. 518-536). New York: Guilford monitoring: A new area of cognitive-developmen- Press. tal inquiry. American Psychologist, 34, 906-911. Davidson, R. J., Ekman, P., Saron, C. D., Senulis, J. Fodor, J. A. (1992). A theory of the child's theory of A., & Friesen, W. (1990). Approach-withdrawal mind. Cognition, 44, 283-296. and cerebral asymmetry: Emotional expression and Fox, N. A. (1989). The psychophysiological corre- brain physiology. Journal of Personality and So- lates of emotional reactivity during the first year of cial Psychology, 58, 330-341. life. Developmental Psychology, 25, 364-372. Davidson, R. J., & Fox, N. A. (1982). Asymmetrical Fox, N. A. (Ed.). (1994). The development of emo- brain activity discriminates between positive and tion regulation. Monographs of the Society for Re- negative affective stimuli in human infants. Sci- search in Child Development, 59(2-3, Serial No. ence, 218, 1235-1237. 240). Davidson, R. J., Schaffer, C. E., & Saron, C. (1985). Fox, N. A., & Davidson, R. J. (1987). Electroenceph- Effects of lateralized presentations of faces on self- reports of emotion and EEG asymmetry in de- alogram asymmetry in response to the approach of pressed and non-depressed subjects. Psychophysi- a stranger and maternal separation in 10-month- ology, 22, 353-364. old infants. Developmental Psychology, 23, 233- Davidson, R. J., & Tomarken, A. J. (1990). Laterality 240. and emotion: An electrophysiological approach. In Fox, N. A., & Davidson, R. J. (1988). Patterns of F. Boiler & J. Grafman (Eds.), Handbook of neu- brain electrical activity during facial signs of emo- ropsychology (Vol 3, pp. 419-441). Amsterdam: tion in 10-month-old infants. Developmental Psy- Elsevier. chology, 24, 230-236. Denham, S. A., Renwick, D. S., & Hewes, S. (1994). Garber, J., & Dodge, K. A. (Eds.). (1991). The de- Emotional communication between mothers and velopment of emotion regulation and dysregula- preschoolers: Relations with emotional compe- tion. New York: Cambridge University Press. tence. Merrill Palmer Quarterly, 40, 488-508. Ginott, H. G. (1965). Between parent and child. New Denham, S. A., Zoller, D., & Couchoud, E. A. York: Avon Books. (1994). Socialization of preschoolers' emotional Ginott, H. G. (1971). Between parent and teenager. and social competence. Developmental Psychol- New York: Avon Books. ogy, 30, 928-936. Ginott, H. G. (1975). Teacher and child. New York: DiPietro, J. A., & Porges, S. W. (1991). Relations Avon Books. between neonatal states and 8-month developmen- Goldsmith, H. H., & Campos, J. J. (1982). Toward a tal outcome in preterm infants. Infant Behavior & theory of infant temperament. In R. N. Emde & R. Development, 14, 441-450. Harmon (Eds.), The development of attachment META-EMOTION 265

and affiliative systems (pp. 161-193). New York: Izard, C. (1982). Measuring emotions in human de- Plenum. velopment. In C. Izard (Ed.), Measuring emotions Goldstone, J. (Producer), & Gillian, T. (Director). in infants and children (pp. 3-20). New York: (1983). Meaning of life [film]. Universal. Cambridge University Press. Goleman, D. (1995). Emotional intelligence. New Izard, C. E., Porges, S. W., Simons, R. F., & Haynes, York: Bantam Books. O. M. (1991). Infant cardiac activity: Developmen- Gottman, J. M. (1983). How children become friends. tal changes and relations with attachment. Devel- Monographs of the Society for Research in Child opmental P)tychology, 27, 432-439. Development, 48(2, Serial No. 201). Katz, L. F., &\Gottman, J. M. (1986). The meta- Gottman, J. M. (1993). The roles of conflict engage- emotion interview. Unpublished manual, Univer- ment, escalation, or avoidance in marital interac- sity of Washington, Department of Psychology. tion: A longitudinal view of five types of couples. Katz, L. F., & Gottman, J. M. (1993). Patterns of Journal of Consulting and Clinical Psychology, 61, marital conflict predict children's internalizing and 6-15. externalizing behaviors. Developmental Psychol- Gottman, J. M. (1994). What predicts divorce? ogy, 29, 940-950. Hillsdale, NJ: Erlbaum. Kirchner, D. M. (1991). Using verbal scaffolding to Gottman, J. M. (in press). Meta-emotion: The mea- facilitate conversational participation and language sures. Hillsdale, NJ: Erlbaum. acquisition in children with pervasive developmen- Gottman, J. M, & Katz, L. F. (1989). The effects of tal disorders. Journal of Childhood Communica- marital discord on young children's peer interac- tion Disorders, 14, 81-98. tion and health. Developmental Psychology, 25, Krokoff, L. (1984). Anatomy of negative affect in 373-381. working class marriages. Dissertation Abstracts Gottman, J. M., & Krokoff, L. (1989). Marital inter- International, 45, 7A. (University Microfilms No. action and satisfaction: A longitudinal view. Jour- 84-22109) nal of Consulting and Clinical Psychology, 57, Leroy, M. (Producer), & Flemming, V. (Director). 47-52. (1939). Wizard of Oz [film]. MGM. Gottman, J. M., & Levenson, R. W. (1992). Marital Linnemeyer, S. A., & Porges, S. W. (1986). Recog- processes predictive of later dissolution: Behavior, nition memory and cardiac vagal tone in 6-month- physiology and health. Journal of Personality and old infants. Infant Behavior and Development, 9, Social Psychology, 63, 221-233. 43-56. Gottman, J. M., & Parker, J. (Eds.). (1986). Conver- Loehlin, J. C. (1992). Latent variable models: An sations of friends. New York: Cambridge Univer- introduction to factor, path, and structural analy- sity Press. sis. Hillsdale, NJ: Erlbaum. Guralnick, M. J. (1981). Peer influences on the de- Maccoby, E. E., & Martin, J. A. (1983). Socialization velopment of communciative competence. In P. in the context of the family: Parent-child interac- Strain (Ed.), The utilization of classroom peers as tion. In E. M. Hetherington (Ed.), Handbook of behavior change agents (pp. 31-68). New York: child psychology, Vol. 4, Socialization, personality Plenum. (pp. 1-102). New York: Wiley. Harris, P. L., & Kavanaugh, R. D. (1993). Young Magni, F., Bruschi, F., & Kasti, M. (1987). The children's understanding of pretense. Monographs afferent innervation of the thymus gland in the rat. of the Society for Research in Child Development, Brain Research, 424, 379-385. 58(1, Serial No. 231). Nance, D. M., Hopkins, D. A., & Bieger, D. (1987). Hofheimer, J. A., & Lawson, (1988). Neurophysio- Re-investigation of the innervation of the thymus logical correlates of interactive behavior in preterm gland in mice and rats. Brain, Behavior, and Im- newborns. Infant Behavior and Development, 11, munity, 1, \JiA-41. 143. Olson, D. R., & Astington, J. W. (1993). Thinking Hooven, C. (1994). The meta-emotion coding system. about thinking: Learning how to take statements Unpublished manuscript, University of Washing- and hold beliefs. Educational Psychologist, 28, ton, Seattle. 7-23. Huffman, L. C, Bryan, Y. E., Pederson, F. A., & Parker, J. G., & Asher, S. R. (1987). Peer relations Porges, S. W. (1988). Infant temperament: Rela- and later personal adjustment: Are low-accepted tionships with heart rate variablility. Unpublished children at risk? Psychological Bulletin, 102, 357- manuscript, National Institute of Mental Health, 389. Rockville, MD. Patterson, G. R. (1982). Coercive family process. Huffman, L. C, Bryan, Y. E., Pederson, F. A., & Eugene, OR: Castalia. Porges, S. W. (1992). Autonomic correlates of re- Porges, S. W. (1984). Heart rate oscillation: An index activity and self-regulation at twelve weeks of age. of neural mediation. In M. G. H. Coles, J. R. Unpublished manuscript, National Institute of Jennings, & J. A. Stern (Eds.), Psychophysiologi- Mental Health, Rockville, MD. cal perspectives: Festschrift for Beatrice and John 266 GOTTMAN, KATZ, AND HOOVEN

Lacey (pp. 229-241). New York: Van Nostrand Richards, J. E. (1985). The development of sustained Reinhold. visual attention in infants from 14 to 26 weeks of Porges, S. W. (1992). Autonomic regulation and at- age. Psychophysiology, 22, 409-416. tention. In B. A. Campbell, H. Hayne, & R. Rich- Richards, J. E. (1987). Infant visual sustained atten- ardson (Eds.), Attention and information process- tion and respiratory sinus arrhythmia. Child Devel- ing in infants and adults (pp. 201-223). Hillsdale, opment, 58, 488-496. NJ: Erlbaum. Saarni, C. (1993). The socialization of emotion. In M. Porges, S. W., & Doussard-Roosevelt, J. A. (in Lewis & J. M. Haviland (Eds.), Handbook of emo- press). The psychophysiology of temperament. In tions (pp. 435-446). New York: Guilford Press. J. D. Noshpitz (Ed.), Handbook of child and ado- Salovey, P., & Mayer, J. D. (1990). Emotional intel- lescent psychology. New York: Wiley Press. ligence. Imagination, Cognition and Personality, Porges, S. W., Doussard-Roosevelt, J. A., & Portales, 9, 185-211. L. A. (1992). Difficultness and cardiac vagal tone: Schaefer, E. S. (1959). A circumplex model for ma- Stable indices of temperament in infants and tod- ternal behavior. Journal of Abnormal and Social dlers. Unpublished manuscript. Psychology, 59, 226-235. Porges, S. W., Doussard-Roosevelt, J. A., Portales, L. Stifter, C. A., & Fox, N. A. (1990). Infant reactivity: A., & Suers, P. E. (1994). Cardiac vagal tone: Physiological correlates of newborn and 5-month Stability and relation to difficulties in infants and 3 temperament. Developmental Psychology, 26, year olds. Developmental Psychobiology, 27, 289- 582-588. 300. Stifter, C. A., Fox, N. A., & Porges, S. W. (1989). Porges, S. W., Walter, G. F., Korb, R. J., & Sprague, Facial expressivity and vagal tone in 5- and 10- R. L. (1975). The influence of methylphenidate on month-old infants. Infant Behavior and Develop- heart rate and behavioral measures of attention in ment, 12, 127-137. hyperactive children. Child Development, 46, 727- Thompson, R. A. (1991). Emotional regulation and 733. emotional behavior. Educational Psychology Re- Porter, F. L., Porges, S. W., & Marshall, R. E. (1988). view, 3, 269-307. Newborn pain cries and vagal tone: Parallel Vygotsky, L. S. (1987). The collected works of L. S. changes in response to circumcision. Child Devel- Vygotsky (R. W. Rieber & A. S. Corton, Eds.). opment, 59, 495-505. New York: Plenum. Pratt M. W., Kerig, P., Cowan, P. A., & Cowan, C. P. Walden, T. A. (1988). A developmental approach to (1988). Mothers and fathers teaching 3-year-olds: social behavior. Journal of Social Behavior and Authoritative parenting and adult scaffolding of Personality, 3, 67-84. young children's learning. Developmental Psy- Wechsler, D. (1974). Selected papers of David chology, 24, 832-839. Wechsler. New York: Academic Press. Putallaz, M., & Gottman, J. M. (1981). An interac- Whitbeck, L. B., Hoyt, D. R., Simons, R. L., & tional model of children's entry into peer groups. Conger, R. D. (1992). Intergenerational continuity Child Development, 52, 986-994. of parental rejection and depressed affect. Journal Redl, F. (1966). When we deal with children. New of Personality and Social Psychology, 63, 1036- York: Free Press. 1045. META-EMOTION 267

Appendix Coaching and Awareness Variables

We created two variables from our coding of the awareness and coaching saw these emotions as toxic meta-emotion interview: coaching and awareness. To and dangerous. give readers a qualitative sense of the emotional People high in awareness tended to believe that nature of these two variables, we present a few ex- one should not stifle one's emotions; that it was good, amples of quotes from parents who were high and healthy, and positive to pay attention to emotion; and low on these two variables. Parents high on coaching that emotions are always there, a part of life, and it is said things like the following: "I feel close to my best to be aware of them. They said things like it is child when he is sad," "When Jason is sad it makes best to get anger or sadness "out of your system" by me feel like a real Dad, now my heart just goes out to becoming aware of it and then being able to deal with him," "When my child is sad I let her know that I it. They believed that it was important to recognize understand," "When Markie is sad I want to know smaller and less intense expressions of emotion to what he is thinking," "It makes me want to hold him prevent them from escalating. They could speak in a close," "I feel affectionate," "I feel love even when differentiated manner about each emotion and the he's upset and angry at me." bodily sensations of each; for example, some parents Some examples of emotion metaphors and con- talked about the "delicious" aspects of sadness in cepts parents expressed were as follows: "Anger some romantic movies but the awful grief that ac- gives me energy and drive," " I think that sadness can companies an important loss. These parents often be good and even productive," "Sadness tells you to described the physical sensations that accompanied slow down," "When I am sad I know something is an emotion, for example, "Sometimes I get so mad missing from my life," "Anger is just like clearing that my stomach is in knots." your throat, just clear it and go on," "In my view, People low in awareness said that these negative attending to sadness is cleansing," "I want her to be effects of anger and sadness were often so aversive sad like in movies. It means she can feel and empa- for them that they tended to prefer to minimize their thize," "I often have the good cry, and I think she importance or not to notice them at all so they does that too," "Getting angry can be a relief, like a wouldn't have to deal with them. That is what we storm that finally happens." tapped with this variable and is probably the essence Parents low on coaching included parents who of being low on the awareness variable. For example, were either disapproving of the child's emotion or one parent said, "When he gets on my nerves like dismissing of the child's emotion. Examples of com- that, I just tune him out," and, "He's not sad much. It ments they made include the following: "Seeing my hurts me to see him sad though. I have to go out for child sad makes me uncomfortable," "I think that a run." Many parents who expressed discomfort with sadness is OK as long as it's under control," "A their child's negative affect tended to view it as toxic child's anger deserves a time out," "I get annoyed and believed that it was their role to get the child out when my child acts sad," "Children often act sad just of the negative mood as quickly as possible. Many of to get their own way," "She looks cute and silly when these parents also said that they and their children she's angry, like a little midget," "I warn him about rarely showed the emotion, or (to prove that they not developing a bad character," "Her shouting scares could survive the negative mood) that they can "roll me," "Molly gets into these black moods." with the punches." They seemed to be at a loss when One example of an emotion metaphor expressed asked to describe how they could tell when they or by a parent low on coaching was as follows: "When their children had the emotion, and they seemed people get angry they are just relieving themselves on unaware of what might make a child feel sad or angry others." Other disapproving examples had to do with and what might be done about it. They often ex- loss of control; humiliation in public; metaphors of pressed the philosophy that the way to cope with fire, pressure, heat, and other explosion and violence negativity is to emphasize the positive aspects of life metaphors for anger; and generally defeat, hopeless- and to substitute a positive emotion for the negative ness, and pathology concerns about sadness. emotion. They said that negative emotion must sim- For the awareness variable, we found that only ply be endured; that the passage of time alone will people who are aware of emotion and can differen- solve emotional problems; that to get over a negative tially talk about nuances of emotion and emotion emotion one should just get on with life's routines, intensity find emotional expression to be acceptable. ignore the emotion and just go on; that anger or If the low awareness reaction to negative emotions sadness meant loss of control; that feelings are pri- were to be characterized as a specific emotion paired vate, not public; and that it is embarrassing to be sad to anger and sadness, it would be fear. People low in or angry and better not to be aware of it. These

(Appendix continues on next page) 268 GOTTMAN, KATZ, AND HOOVEN

Appendix (continued) parents were often most aware of the demand com- and sadness, even if their names, awareness and ponent of a child's emotion (that they fix the world coaching, do not sound very emotional, and make it perfect so that their child will not have this awful emotion). Received January 22, 1995 To summarize, these two variables represent Revision received December 14, 1995 highly emotional reactions and metaphors to anger Accepted December 15, 1995 •

New Editors Appointed, 1998-2003

The Publications and Communications Board of the American Psychological Association announces the appointment of five new editors for 6-year terms beginning in 1998.

As of January 1,1997, manuscripts should be directed as follows:

• For the Journal of Experimental Psychology: Animal Behavior Processes, submit manuscripts to Mark E. Bouton, PhD, Department of Psychology, University of Vermont, Burlington, VT 05405-0134.

• For the Journal of Family Psychology, submit manuscripts to Ross D. Parke, PhD, Department of Psychology and Center for Family Studies-075,1419 Life Sciences, University of California, Riverside, CA 92521-0426.

• For the Personality Processes and Individual Differences section of the Journal of Personality and Social Psychology, submit manuscripts to Ed Diener, PhD, Department of Psychology, University of Illinois, 603 East Daniel, Champaign, IL 61820.

• For Psychological Assessment, submit manuscripts to Stephen N. Haynes, PhD, Department of Psychology, University of Hawaii, 2430 Campus Road, Honolulu, HI 96822.

• For Psychology and Aging, submit manuscripts to Leah L. Light, PhD, Pitzer College, 1050 North Mills Avenue, Claremont, CA 91711-6110.

Manuscript submission patterns make the precise date of completion of the 1997 volumes uncertain. Current editors, Stewart H. Hulse, PhD; Ronald F. Levant, EdD; Russell G. Geen, PhD; James N. Butcher, PhD; and Timothy A. Salthouse, PhD, respectively, will receive and consider manuscripts until December 31, 1996. Should 1997 volumes be completed before that date, manuscripts will be redirected to the new editors for consider- ation in 1998 volumes.