Preschool Teachers’ Socialization of Emotion Knowledge: Considering Socioeconomic Risk

Abstract

Emotion knowledge supports early school success. Socializers’ emotions, contingent reactions to emotions, emotion language, and beliefs about emotions can contribute to preschoolers’ emotion knowledge, but more is known about parents’ contributions than teachers’. We expected teachers’ emotion socialization findings to parallel those in the parent literature, with potential moderation by classroom-level socioeconomic risk. Participants included 85 teachers and 327 preschoolers in high or low socioeconomic risk classrooms. Children’s emotion knowledge was assessed twice. Teachers reported on emotional expressiveness, reactions to children’s emotions, and emotion socialization beliefs. Their emotion language during book reading was codified. Hierarchical linear models predicted end-of-year emotion knowledge via emotion socialization, risk, and their interactions, controlling for child age, gender, and beginning-of-year emotion knowledge. Only one finding resembled the parent literature; most were specific to high socioeconomic risk classrooms, highlighting the importance of context in emotion socialization. Applications, limitations, and suggestions for future research are considered.

Emotion knowledge, at the preschool level, consists of accurately identifying expressions of emotions, as well as increasingly complex comprehension of situations that elicit emotions and their causes. Such knowledge is identified as important support for early school success and the growth of academic competence during elementary school (Nix, Bierman, Domitrovich, & Gill, 2013). Children who better understand emotions in preschool and at school entry are more likely to (a) develop positive and supportive relationships with peers and teachers, (b) demonstrate positive learning behaviors/attitudes and social competence, and (c) achieve at higher levels throughout early schooling; children not attaining such knowledge are at risk in these areas (Blankson et al., 2017; Denham, Bassett, Mincic et al., 2012; Denham, Bassett, Way et al., 2012; Di Maggio, Zappulla, & Pace, 2016; Garner & Waajid, 2008; Izard et al., 2001; Miller et al., 2006). In short, emotion knowledge helps to ensure children’s early school success, and its effects may be long lasting. In fact, kindergarten prosociality (including emotion knowledge) is associated with adult success in education, employment, mental health, and avoidance of crime and substance use, independent of important child, family, and contextual factors (Jones, Greenberg, & Crowley, 2015).

Fueled in part by this evidence, educators, parents, and policymakers are becoming ever more aware of the need to address emotion knowledge in early childhood educational settings and training (Buettner, Hur, Jeon, & Andrews, 2016; Dusenbury et al., 2015; O’Connor, De Feyter, Carr, Luo, & Romm, 2017). To support these initiatives, more research is warranted on the promotion of emotion knowledge in early educational contexts.

Socialization of Emotion Knowledge

Thus, given its importance, how is such emotion knowledge fostered? The components of emotion socialization theory include socializers’ expressed emotions, contingent reactions to specific emotions, teaching about emotions, and beliefs about emotions, which help young children acquire culturally appropriate emotion knowledge (Denham, 2019; Denham, Bassett, & Wyatt, 2014; Eisenberg, Cumberland, & Spinrad, 1998). Most of what we know about these mechanisms emanates from studying families.

First, children observe emotions exhibited by people with whom they interact. The modeling mechanism of emotion socialization includes specific positive and negative emotions observed by children along with socializers’ overall emotional expressiveness. Family positive expressiveness (e.g., happiness, gratitude) promotes emotion knowledge, perhaps rendering young children more open to learning about emotions; conversely, exposure to negative emotions (e.g., sadness, anger) may hamper emotion knowledge by upsetting children, making self-reflection about emotion difficult (Nixon & Watson, 2001; Raver & Spagnola, 2002).

Second, children’s emotions often elicit, even require, contingent reactions from social partners. Adults respond to young children’s emotions in ways that have been construed as either supportive (e.g., accepting, comforting) or nonsupportive (e.g., minimizing, punishing). These reactions convey important messages about emotions, bearing on preschoolers’ emotion knowledge (Perlman, Camras, & Pelfrey, 2008; Pintar Breen, Tamis-LeMonda, & Kahana-Kalman, 2018). Supportive reactions to children’s emotions may help children differentiate emotions, perhaps by making the child feel more comfortable to explore emotions and affording them practice in successfully dealing with them. In contrast, nonsupportive reactions may motivate children to avoid emotional information (Denham & Kochanoff, 2002).

Third, teaching about emotion consists of verbally commenting upon or explaining an emotion and its relation to an observed event or expression. Adults’ discussions about emotions promote children’s emotion knowledge, perhaps by directing attention to salient emotional cues and giving children reflective distance from their feelings so that they can better interpret and evaluate feelings’ causes and consequences (Denham, Mitchell-Copeland, Strandberg, Auerbach, & Blair, 1997; Doan & Wang, 2010; Garner, Dunsmore, & Southam-Gerrow, 2008). The general trend of these findings also is found for low-income, minority families (Garner, 2006).

Finally, beliefs about how to socialize emotional competence underlie emotion socialization behaviors, and may contribute to young children’s emotion knowledge, (Halberstadt et al., 2013; Halberstadt, Thompson, Parker, & Dunsmore, 2008; Wong, Diener, & Isabella, 2008; Wong, McElwain, & Halberstadt, 2009). For example, parents who value teaching about emotions may approach children’s emotions by coaching (e.g., using children’s feelings as teachable moments and opportunities to talk about emotions’ causes and consequences; Dunsmore & Karn, 2001; Parker et al., 2012; Paterson et al., 2012). Conversely, adults who consider emotions problematic may approach children’s emotions more negatively (e.g., trying to get children to stop showing emotions by punishing them, or minimizing emotions through inattention), so that children learn only to suppress emotions (Denham & Kochanoff, 2002; Perlman et al., 2008). Alternatively, emotion socialization beliefs may lean toward a more laissez-faire attitude about emotion (e.g., not really knowing how to respond to children’s emotions), leading to restricted input about emotions (Perez-Rivera & Dunsmore, 2011).

In short, the extant literature has shown that families’ emotional lives contribute much to preschoolers’ emerging emotion knowledge. Children’s daily experiences today usually include other contexts and other important adults, however, especially their teachers in childcare or preschool. What is known, then, about how teachers’ emotion socialization contributes to developing emotion knowledge?

Teachers’ role in socialization of emotion knowledge.

Young children could also learn about emotions through their rich daily interactions with teachers (e.g., Bassett, Denham, Mohtasham, & Austin, in press; Hyson, 1994). Even when not directly involved in interaction, they may learn about emotions through observing teachers. Thus, preschool teachers may be pivotal socializers of children’s emotion knowledge (Denham, Bassett, & Zinsser, 2012). Teachers’ emotions, reactions, and teaching about emotions, undergirded by beliefs, are likely to send socialization messages to children like those encountered in the family, contributing to children’s emotional competence in similar ways. However, contextual differences also point to potentially unique contributions of teacher emotion socialization—the adult/child ratio in the classroom, for example, and concomitant need for organization, may dictate teachers’ stricter reactions to emotions (Denham & Bassett, 2019). Teachers also may feel the need to project an emotionally calm demeanor in the classroom, despite the often-stressful nature of their work (Shewark, Zinsser, & Denham, 2018). But there still is relatively little evidence on how early childhood educators socialize emotional competence (Denham, Bassett, & Zinsser, 2012).

Accordingly, in this study we seek to expand this small literature, examining teachers’ contributions to preschoolers’ emotion knowledge. First, though, we should examine the sparse evidence that corroborates the potential importance of preschool teachers’ emotion socialization. Early childhood teachers engage in a wide variety of specific emotion socialization behaviors in the classroom (Ahn & Stifter, 2006; Ersay, 2007). They show emotion, including both anger and joy (Fu, Lin, Syu, & Guo, 2010). They also encourage and discourage young children’s emotional expression via a variety of behaviors, such as comforting, distraction, problem-solving, punishment, or minimization, but infrequently validate children’s emotions (e.g., “it’s okay to feel sad”; Ahn, 2005b; Ahn & Stifter, 2006). Concerning relations with emotion knowledge, teachers’ negative reactions were negatively related to older preschoolers’ emotion knowledge (Morris, Denham, Bassett, & Curby, 2013).

Regarding teaching about emotions, early childhood teachers also use emotion language in the classroom, although relatively infrequently (Ahn, 2005a; Yelinek & Grady, 2017). They explain and question during teacher-led activities and use socializing and guiding language during free play (e.g., “we smile when we say hello”, “you can pound these blocks if you’re mad”). In addition, their book reading is a noteworthy outlet for emotion socialization that can promote children’s emotion knowledge. For example, when teachers explain causes and consequences of story characters’ emotions (e.g., “Do you think she is sad because the ball fell in the river?”), children’s emotion knowledge increases across a school year (Bassett et al., in press). Moreover, when teachers read books including an enriched emotional lexicon, and then converse about them, children show growth in emotion knowledge (Grazzani, Ornaghi, Agliati, & Brazzelli, 2016; see also Bergman Deitcher, Aram, Khalaily-Shahadi, & Dwairy, 2020, for concordant cross-cultural results). Finally, regarding early childhood teachers’ beliefs about emotions, Ahn (2005a) noted that her observations of Korean early childhood educators showed congruence between their emotion-related beliefs and emotion socialization behaviors.

We seek to extend these initial theoretical and empirical efforts. Hence, our overarching goal for the current study was to further understand teachers’ emotion socialization in preschool classrooms, by examining teachers’ self-reported emotion socialization (i.e., emotions, reactions to children’s emotions, beliefs) and directly observing their emotion language during picturebook reading. Knowing how preschool teachers’ emotion socialization is related to children’s developing emotion knowledge could inform practice recommendations and professional development.

Socioeconomic Risk

Thus, understanding teacher emotion socialization is an important goal, but it may be an incomplete one. Over and above elucidating teachers’ emotion socialization attitudes and behaviors, attention must be paid to contextual issues that can be critical in the development of preschoolers’ emotion knowledge. Socioeconomic status is one such contextual issue. For example, poverty’s stressors prominently include parents’ struggle to pay bills and “make ends meet.” These stressors may elicit overwhelming, negative emotions within the family, rendering children’s acquisition of emotion knowledge more difficult (Denham, Bassett, Mincic, et al., 2012; Denham, Bassett, Zinsser, & Wyatt, 2014; National Scientific Council on the Developing Child, 2005/2014; Raver, Blair, Garrett-Peters, & Family Life Project Key Investigators, 2015). Specifically, children’s ability to detect and appraise emotional stimuli is greater in the absence of environmental adversities such as poverty, household chaos, and interparental aggression (Raver et al., 2015; see also Erhart, Dmitrieva, Blair, & Kim, 2019).

Socioeconomic risk also can be related to differences in emotion socialization in the family, via the emotional climate that poverty can induce. For example, the stress of living in poverty understandably can dampen maternal positive expressiveness, even controlling for depression (Davis, Suveg, & Shaffer, 2015). At the same time, such stress may exacerbate mothers’ punitive or dismissing reactions to children’s emotions (Shaffer, Suveg, Thomassin, & Bradbury, 2012). If living in poverty makes it difficult for parents to demonstrate positive emotion socialization, could early experiences with teachers buffer such effects? Indeed, low-income children are especially sensitive to quality social-emotional programming, with advantageous promotion of their emotional competence (Fishbein et al., 2016; Nix, Bierman, Domitrovich & Gill, 2013). Thus, we argue that teacher emotion socialization may be especially important for children living in poverty.

In considering socioeconomic risk, it is important to recognize that teachers’ emotion socialization is conveyed to their entire classrooms. Even when not the target of a teachers’ expressiveness, reactions, emotion language, and beliefs (or when there is no specific target for these components of emotion socialization), each child is exposed to and arguably learns about emotions. Thus, in this initial examination of teachers’ emotion socialization, socioeconomic risk, as it moderates such classroom-level emotion socialization, is operationalized as attendance in a classroom serving low-income families. We classify each classroom, then, as either at high or low socioeconomic risk depending on program target population.

The Current Study

Building from these considerations, the overarching goal of the present study focused on the contribution of teachers’ emotion socialization behaviors to children’s emotion knowledge. In accordance with findings from extant parent emotion socialization literature, we expect that self-reported positive and negative emotional expressiveness, along with supportive and unsupportive reactions to children’s emotions, language used to teach about emotion, and attitudes about emotion socialization will contribute to children’s year-end emotion knowledge, over and above earlier emotion knowledge. Thus, we expect that teachers’ positive expressiveness would be positively related to children’s emotion knowledge, helping them to become receptive to learning about emotions. In contrast, teacher negativity would create an atmosphere where addressing emotions would be difficult. Further, it is expected that teachers’ supportive reactions to children’s emotions - encouraging, comforting, helping - would be positively related to children’s emotion knowledge, with the converse true for their punishing or minimizing reactions. Supportive responses from teachers would help children “stay in the moment” to learn more about emotions. It is also expected that teachers who discuss emotions when reading picturebooks convey content to children about emotions. Finally, teachers’ beliefs about coaching emotions will contribute to children’s emotion knowledge.

However, given little extant research and important contextual differences in classrooms versus families (e.g., dealing with multiple rather than individual children), we cannot rule out unique teacher contributions to children’s emotion knowledge. Teachers’ roles differ from parents’; they may see themselves as readying preschoolers to function in a school setting, using a more “no nonsense” approach, perhaps including pressure for children to minimize emotional outbursts, so that classroom order can be maintained. In the classroom context, such emotion socialization might support development of emotion knowledge via focusing children’s attention.

Further, contributions of classroom-level teacher emotion socialization may be especially important to developing emotion knowledge for children living in poverty. In particular, given these children’s greater risk of exposure to parents’ frustration, anger, and proneness to losing emotional control (Raver et al., 2015), children living in poverty may especially benefit from learning about emotions within an emotionally positive, coaching environment (National Scientific Council on the Developing Child, 2005/2014). At the same time, they also may benefit from a stricter, less “coddling”, approach to their emotions, reflective of the increased need to regulate emotions in their environment (Dunbar, Leerkes, Coard, Supple, & Calkins, 2017).

Thus, our first research question is: How does teachers’ self-reported emotion socialization and discussion about emotions contribute to young children’s emotion knowledge over the preschool year? Our second research question is: How do contributions of teachers’ emotion socialization vary by the socioeconomic risk status of their classrooms?

Method

Participants

Participants included 85 teachers and 327 children in three- and four-year-old classroom (m_age = 48.10, sd_age = 7.22; 54% boys). Children attended privatized (i.e., private or university-based) child care (n_centers = 20 n_teachers = 66; n_children = 240; termed “low socioeconomic risk”. On average less than 15% of students in each center received subsidized care) or government- or church-sponsored centers (n_centers = 3; n_teachers = 19; n_children = 87; termed “high socioeconomic risk”. To be enrolled, families were documented to be experiencing some level of poverty). Sixty-five percent of teachers had attained a BA degree or better, 51% had taught for less than 10 years, and 54% were less than 35 years old. In terms of ethnicity and race, 63.5% of teachers were European American, 18.8% African American, and 5.9% Asian, with 11.9% identifying as Latinx.

Given our second problem question, we examined correlates of classroom economic risk status. Teachers of classrooms at higher socioeconomic risk tended to be better remunerated (χ ² (4) = 37.81, p < .001), and more likely to be African American, not European American (χ ² (2) = 5.72, p < .06). The total group of children for whom demographic data were available was 73.3% European American, 13.8% African American, 6.2% Asian, and 6.7% other, as well as 16.6% Latinx. Children in classrooms at higher socioeconomic risk were more likely to be African American (and less likely to be European American or Asian, χ ² (3) = 57.43, p < .001); they were also more likely to be Latinx (χ ² (1) = 57.42, p < .001).

Procedure

Before the beginning of the study, it was approved by the local Institutional Reviewer Board. Subsequently, participants were recruited at the beginning of the school year; after meeting with each center’s director, we obtained consent from participating teachers. Then, children and families in these teachers’ classrooms were invited to participate at recruitment events, information sessions held at the facilities, and/or through the help of facility personnel.

Direct assessments of children’s emotion knowledge were performed in mid- to late fall, after children had become acclimated to the classroom (October to early December, T1), and in late spring, near the end of the school year (mid-March to early May, T2), with an average of five months between assessments. Teachers completed self-report measures and read two emotion-laden, wordless, picturebooks (as a direct observation of their emotion coaching behavior) between T1 and T2.

Measures

Emotion socialization: Modeling.

To investigate emotions modeled in the classroom, we used the Classroom Expressiveness Questionnaire (CEQ; Halberstadt & Wilson, 2010). Adapted for teachers from the validated Self-Expressiveness in the Family Questionnaire (SEFQ; Halberstadt, Cassidy, Stifter, Parke, & Fox, 1995), the CEQ examines teachers’ emotional expressiveness in the classroom. The questionnaire’s 30 hypothetical scenarios represent a range of positive and negative emotions, and teachers indicate the frequency with which they express them in their classroom (from “1, not at all frequently” to “9, very frequently”). For example, one negative emotion scenario is “expressing anger at a student’s carelessness”; a positive emotion scenario is “expressing excitement over future classroom plans”.

This measure was adapted for use in classrooms by replacing family-oriented language with classroom references (Halberstadt & Wilson, 2010). Cronbach’s αs for this sample were .73 for Positive Expressiveness (12 items) and .72 for Negative Expressiveness (17 items, with one item, “Crying at the end of the year when students leave”, removed). In this study, items for Positive and Negative Expressiveness scales were averaged for use in subsequent analyses.

Emotion socialization: Reactions to children’s emotions.

To investigate teachers’ reactions to children’s emotions, we used the Coping with Children’s Negative Emotions Scale, Teacher Version (TCCNES). Designed for research with parents but adapted for use with teachers under guidance of the authors (Fabes, Eisenberg, & Bernzweig, 1990), the TCCNES contains 10 items describing hypothetical situations in which a child in their classroom is expressing a negative emotion. Teachers are asked to indicate their likelihood of response to six choices within each item on a 7-point response scale, from “1, very unlikely” to “7, very likely”. Their responses are aggregated to create six subscales as follows: Distress Reactions (e.g., “feel uncomfortable, embarrassed myself”), Punitive Reactions (e.g., “tell child to straighten up or s/he’ll have to sit out for a while”), Minimization Reactions (e.g., “tell child that s/he is overreacting”), Expressive Encouragement (e.g., “encourage child to talk about his/her nervous feelings”), Emotion-Focused Reactions (e.g., “suggest child think about something relaxing…”), and Problem-Focused Reactions (e.g., “help child think of places s/he hasn’t looked yet”).

Fabes, Poulin, Eisenberg, & Madden-Derdich (2002) found good-to-excellent internal reliability for the original scales, as well as significant test-retest reliability across four months. They also reported, as evidence of construct validity, interrelations with other self-report indices of parental reactions to children’s emotions. In this study, Cronbach’s αs for this sample, for each 10-item scale, were .46 for Distress, .73 for Punitive, .61 for Minimizing, .89 for Emotion Encouragement, .84 for Emotion-Focused, and .89 for Problem-Focused. Items for each scale were each averaged for use in subsequent analyses; because of Distress scale’s low α, however, it was not included hereafter. Average scores for “nonsupportive” (i.e., Punitive and Minimizing) and all “supportive” (i.e., Emotion Encouragement, Emotion-Focused, and Problem-Focused) scales were also aggregated to examine the overall contribution of these techniques (αs = .68 and .94).

Emotion socialization: Teaching about emotions.

Teaching about emotions was gleaned from teachers’ reading of two emotionally evocative wordless picturebooks. These books included pictures only, and portrayed situations involving happiness, surprise, sadness, anger, and fear. Instructions to the teachers were deliberately vague to allow maximum flexibility; they were simply asked to talk with the children about these books in whatever manner they wished. Reading sessions were transcribed for ease of coding. See Pentimonti et al. (2012) and Ziv, Smadja, and Aram (2015) for similar procedures in the literature.

The coding system for the picturebook task was derived from the Parent-Child Affect Communication Task, which is well-established in the parent emotion socialization literature (PACT; Denham & Kochanoff, 2002), and has been used to track teachers’ emotion language as well (Morris et al., 2013). Words referring to discrete emotions and their behavioral expression were counted, along with their positive or negative valence (repetitions of the children’s emotion labels were not included in the tally). Functions of utterances containing emotion words were also tallied for commenting (e.g., “she looks sad”), explaining (e.g., “He’s happy because his friend is there”), and/or clarifying (e.g., “his face actually looks sad instead of angry, doesn’t it?”). Total number of positive and negative words, and percentage of total emotion words for each function code, were used in subsequent analyses.

All picturebook reading was audiotaped via automatic recorder and transcribed professionally. Coder training consisted of introduction to the manual, discussion of examples, and coding of practice and reliability transcripts. Seven coders were exposed to three practice transcripts; if these were coded within a satisfactory margin with the master coder, they then coded 18 reliability transcripts, both using data from an earlier project. Inter-rater reliability coefficients for coding positive and negative emotion terms (expressed as intraclass correlations) were .95 and .93, respectively. Intraclass correlations for coding the functions commenting, explaining, and clarifying ranged from .85 to .98, with an average of .93 (all ps < .001). Scores used indicate the number of tallies for each category across the two book readings.

Emotion socialization: Beliefs about emotions.

To measure teachers’ beliefs about teaching about emotions and contingent responding to children’s emotions, we created the Teacher Emotion Socialization Self-Test (TESST) based on a measure for parents created by Hakim-Larson, Parker, Lee, Goodwin, and Voelker (2006) and shortened by Paterson and colleagues (2012). Paterson et al.’s items were lightly edited for use in a classroom setting, as with the CEQ and TCCNES. The TESST consists of 20 statements to which teachers respond on a seven-point scale of endorsement, from “1, strongly disagree” to “7, strongly agree”. Four five-item subscales are generated: Dismissing (e.g., “children have a right to feel angry”), Punishing (e.g., “when a child gets angry, my goal is to get him/her to stop”), Laissez-Faire/Ineffective (e.g., “When my child is angry, I am not quite sure what he or she wants me to do”), and Emotion Coaching (e.g., “when a child is sad, I try to help him/her explore what is making him/her sad”) styles. These factors were identified by Paterson et al. (2012), who also offer information on their very similar measure’s good-to-excellent reliability. There is also supportive evidence for the measure’s concurrent validity, in terms of its relations with aspects of emotion discussions and with self-reports of emotions and reactions to children’s emotions (Baker, Fenning, & Crnic, 2011; Halberstadt et al., 2013; Paterson et al., 2012; Swartz & McElwain, 2012). In this study, Cronbach’s α for five-item scales were .71 for Emotion Coaching, .65 for Punishing, .81 for Dismissing, and .76 for Laissez-Faire/Ineffective. Items for each scale were averaged for use in subsequent analyses.

Child outcome: Emotion knowledge.

The Affect Knowledge Test-Shortened (AKT-S; Denham, Bassett, Brown, Way, & Steed, 2015) was used to assess preschoolers’ emotion knowledge. AKT-S assessed preschoolers’ understanding of emotion using puppets with detachable faces that depict happy, sad, angry, and afraid expressions. For labeling (six items), children were asked to identify sad, angry and afraid facial expressions by verbally naming them (expressive knowledge), and then by non-verbally pointing to the corresponding face when presented with the emotion word (receptive knowledge). For situation knowledge, nine vignettes were enacted using puppets, accompanied by vocal and visual affective cues emitted by the puppet/experimenter. For three stereotypical emotion knowledge vignettes, the puppet depicted the emotion most people would feel (e.g., fear during a nightmare). In the remaining six nonstereotypical vignettes, the puppet depicted emotions different from each mother’s reports of her child’s likely feelings. Among nonstereotypical situations, three vignettes pitted positive versus negative emotion (e.g., happy or sad to come to preschool); the rest pitted negative versus negative emotion (e.g., angry at or afraid of a sibling’s aggression).

For each scenario, children affixed a flannel face to report the puppet’s emotion, and received two points for correct identification of emotion, or one point for identifying correct valence but not correct emotion (e.g., sad for afraid). The score used in subsequent analyses was the mean of standard scores for all subtests’ items. Internal consistency reliability α for this study was .84 at T1 and .72 at T2. The AKT-S has demonstrated reliability and validity (Denham et al., 2015).

Analytic Plan

After calculating descriptive statistics for all study variables and inter-scale correlations for teacher and child measures separately, we conducted the principal analyses to address the specific research questions. 2-Level Hierarchical Linear Modeling (HLM; Raudenbush & Bryk, 2002) analyses. The unconditional model for emotion knowledge was first examined before the full multi-level models. The associated ICC was calculated, estimating the amount of variance at the classroom level and thus appropriateness of HLM.

Next, separate full models were created. Given that only seven participants’ emotion knowledge data were missing at T2, data were handled with listwise deletion, resulting in 305 children with T2 scores. We controlled for age, gender, and the T1 premeasure of the outcome variable via their inclusion at Level 1. Thus, on Level 1, the T2 emotion knowledge outcome (Ŷ) was predicted by the classroom intercept, β₀, plus the effect of child age, β₁, plus the effect child gender (1=boy; 2=girl), β₂, plus the effect of T1 emotion knowledge, β₃. All predictors were centered around the grand mean except for child gender, which was uncentered.

$$\underset{\_}{\text{Level-}1\text{ Model:}}\text{ Y}=ß0+ß1*\left(\text{AGE\_MONTHS}\right)+ß2*\left(\text{C\_GENDER}\right)+ß3*\left(\text{T}1\_\text{EMOTION KNOWLEDGE}\right)+\text{R}$$

Continuing to Level 2, teacher emotion socialization predictors, classroom socioeconomic risk, and interactions between teacher emotion socialization predictors and risk were included. Thus, for child i in classroom j, B0 is equal to the classroom average for T2 emotion knowledge, plus effects for levels of risk and each teacher predictor γ_0x, interactions of risk and each teacher predictor, plus error, μ₀. The Level-2 equation models between-classroom variance using risk uncentered (1=high-risk classroom; 0=low-risk classroom), and each teacher behavior predictor and its interaction with risk grand mean centered. Using the Classroom Expressiveness Questionnaire as an exemplary predictor:

$$\text{Level-}2\text{ Model: }ß0=\text{G}00+\text{G}01*(\text{RISK})+\text{G}02*\left(\text{CEQ\_POSITIVE}\right)+\text{G}03*\left(\text{CEQ\_NEGATIVE}\right)+\text{G}04*\left(\text{RISK}*\text{ CEQ\_POSITIVE}\right)+\text{G}05*\left(\text{RISK}*\text{ CEQ\_NEGATIVE}\right)+\text{U}0$$

Results

Descriptive statistics and correlations.

Tables 1 through 3 show descriptive statistics and intercorrelations. As can be seen in Table 1, teachers self-reported relatively positive emotions, positive rather than negative reactions to children’s emotions, and prevalent attitudes endorsing both emotion coaching and dismissing children’s emotions (i.e., merely accepting their negative emotions). In terms of the picturebook reading task, across two books, there were moderate numbers of emotion words used, with the predominant function of such language as comments and explanations; there was a great deal of variability in the teachers’ use of emotion language (i.e., relative standard deviations of 40 – 96 percent). Children’s emotion knowledge was good, as evidenced by the average item scores for the subtests that were used to create emotion knowledge aggregates at T1 and T2.

Table 1.

Descriptive Statistics for Teacher Self-Report and Book-Reading and Child Emotion Knowledge

Teacher Emotion Socialization Measure	Mean (SD)	Possible Range/Actual Range
Positive Expressiveness a	7.72 (0.82)	5.53 – 9.00 / 1 – 9
Negative Expressiveness a	2.99 (0.79)	1.61 – 5.59 / 1 – 9
Punitive Reactions a	1.61 (0.65)	1.00 – 4.10/ 1 – 7
Minimizing Reactions a	1.75 (0.49)	1.00 – 3.30/1 – 7
Emotion Encouragement a	5.75 (0.97)	2.80 – 7.00/ 1 – 7
Emotion-Focused Reactions a	5.26 (1.04)	2.00 – 7.00/ 1 – 7
Problem-Focused Reactions a	6.04 (0.89)	2.90 – 7.00/ 1 – 7
Total Positive Words	9.33 (6.25)	0.00 – 34.50 / NA
Total Negative Words	15.43 (8.79)	0.00 – 38.00 / NA
Percent Comments	38.48 (15.66)	0.00 – 76.92/ NA
Percent Explanations	29.05 (19.03)	0.00 – 86.36/ NA
Percent Clarifying	8.52 (8.18)	0.00 – 29.03/ NA
Coaching a	6.18 (0.68)	4.60 – 7.00/ 1 – 7
Dismissing a	5.11 (1.09)	2.00 – 7.00/ 1 – 7
Punishing a	2.59 (0.89)	1.00 – 4.40/ 1 – 7
Laissez-Fairea	2.04 (0.81)	1.00 – 4.00/1 – 7

Child Emotion Knowledge Subscales	T1	T2	T1 Range	T2 Range	Actual Range
Expressive Negative	1.56 (0.43)	1.70 (0.33)	0.00–2.00	0.67–2.00	0 −2
Receptive Negative	1.77 (0.36)	1.92 (0.22)	0.00–2.00	0.50–2.00	0 −2
Stereotypical Situations	1.67 (0.45)	1.85 (0.32)	0.00–2.00	0.00–2.00	0 −2
Nonstereotypical Situations	1.70 (0.41)	1.84 (0.27)	0.00–2.00	0.00–2.00	0 −2

Table 3.

Correlations Among Affect Knowledge Test Subscales

	Scales	1.	2.	3.	4.	5.	6.	7.	8.
1.	T1 Expressive	…	.33***	.22***	.27***	.29***	.40***	.36***	.34***
2.	T1 Receptive		…	.34***	.39***	.24***	.35***	.28***	.29***
3.	T1 Stereotypical			…	.50***	.18**	19***	.27***	.27***
4.	T1 Nonstereotypical				…	.14*	.28***	.30***	.32***
5.	T2 Expressive					…	.27***	.21***	.16**
6.	T2 Receptive						…	.28***	.34***
7.	T2 Stereotypical							…	.45**
8.	T2 Nonstereotypical								…

^*p < .05,

^**p < .01,

^***p < .001.

In terms of intercorrelation of indices of teachers’ emotion socialization, Table 2 indicates notable associations among self-reports, especially within reactions and between reactions and both emotional expressiveness and beliefs. There were a few borderline or significant associations between teacher beliefs, reactions, and emotion terms used in picturebook reading, in expected directions. Table 3 corroborates the internal consistency, as well as the test-retest reliability, of the emotion knowledge measure, in that all associations were statistically significant.

Table 2.

Correlations Among Teacher Self-Report and Book-Reading Scales

Scales		1.	2.	3.	4.	5.	6.	7.	8.	9.	1.	11.	12.	13.	14.	15.	16.
1.	CEQ Positive	…	.06	.51***	.01	.03	−.36***	−.17	−.16	.48***	.48**	.53***	.09	.18+	−.07	.08	.04
2.	CEQ Negative		…	−.11	.25*	.10	.23*	.25*	.02	.01	−.10	−.03	.02	.06	.01	−.12	.15
3.	ERTS Coaching			…	−.07	−.05	−.24*	−.13	−.05	.54***	.52***	.63***	.10	.09	−.12	.18	−.03
4.	ERTS Dismissing				…	.02	.10	−.01	−.04	.34**	.10	.08	.11	.07	−.09	−.20+	.20+
5.	ERTS Punishing					…	.19+	.24*	.38***	.01	.28**	.10	−.33**	−.25*	.04	.10	−.14
6.	ERTS Laissez-Faire						…	.16	.20+	−.15	−.03	−.19+	.05	−.05	.03	−.07	.12
7.	TCCNES Punitive							…	.53***	−.18+	.04	−.20+	−.17	−.18	.14	.17	−.22*
8.	TCCNES Dismissing								…	.01	.25*	−.01	−.23+	−.23+	.06	.12	−.13
9.	TCCNES Emotion Encouragement									…	.59***	.72*	.16	.17	−.03	−.02	.16
10.	TCCNES Emotion-Focused										…	.69***	−.05	−.01	−.13	.19*	−.15
11.	TCCNES Problem-Focused											…	.22*	.20+	.04	.10	−.15
12.	Total Positive Words												…	.75***	.16	−.19+	.31**
13.	Total Negative Words													…	.09	−.13	.31**
14.	Percent Commenting														…	−.19+	−.10
15.	Percent Explaining															…	−.46***
16.	Percent Clarifying																…

⁺p < .10,

^*p < .05,

^**p < .01,

^***p < .001.

Unconditional multi-level models.

The variance explained at Level 1 for each model are presented in Tables 4 to 7. The ICC in the unconditional model showed that class membership accounted for a significant amount of variance in emotion knowledge (ICC = .16, p < .001), though still less than variance explained at the child level, suggesting that HLM is appropriate. Thus, classroom membership was an important factor in predicting children’s outcomes.

Table 4.

Hierarchical Linear Modeling Analyses: Teacher Emotions Predicting AKT-S

CEQ Predictors	AKT-S
Fixed Effects, Level 2
Intercept	0.14
Risk	−0.99
Positive Expressiveness	−0.05
Negative Expressiveness	0.02
Risk X Positive Expressiveness	0.13*
Risk X Negative Expressiveness	−0.03
Level-1 % Variance Explained	40.8+

⁺p < .10,

^*p < .05.

Note. CEQ = Classroom Emotion Questionnaire. AKT= total AKT-S score.

Table 7.

Hierarchical Linear Modeling Analyses: Teacher Beliefs Predicting AKT-S

TESTT Predictors	AKT-S
Fixed Effects, Level 2
Intercept	0.43
Risk	−2.15*
Coaching	−0.03
Dismissing	0.00
Punishing	0.01
Laissez-Faire	0.00
Risk X Coaching	0.28**
Risk X Dismissing	0.01
Risk X Punishing	0.06
Risk X Laissez Faire	0.04
Level-1 % Variance Explained	40.8+

⁺p < .10,

^*p < .05,

^**p < .01.

Note. TESST = Teacher Emotion Socialization Self-Test. AKT = total AKT-S score.

Full models.

Level-2 findings are shown in Tables 4 through 7 and Figure 1. Before addressing our problem questions, we briefly confirm earlier research via reporting on covariate contributions (e.g., Denham et al., 2015). Thus, in Level 1, child age (in months) and T1 AKT-S scores positively predicted T2 emotion knowledge in all equations (β_age = 0.01, p < .01, and β_premeasure = 00.45 to .49, ps < .001). As for child gender, its regression coefficient bordered on significance in two equations and was significant in the equation involving teachers’ emotional expressiveness (β_gender = 0.09, p < .05); being a girl predicted greater T2 emotion knowledge. Variance explained at Level 1 is also presented in Table 4 through 7.

Figure 1.

Interactions of Classroom Risk Status with Total Emotion Knowledge as Y-Axis

Considering Level 2 findings, there were overall few main effects of emotion socialization, except for teachers’ minimizing reactions positively predicting T2 emotion knowledge, with the opposite pattern for punitive reactions. Classroom socioeconomic risk was generally negatively associated with T2 emotion knowledge, but only significantly so in conjunction with teachers’ reactions and beliefs.

In contrast, classroom socioeconomic risk moderated several findings (see Figure 1). In high-risk classes, positive emotional expressiveness, teachers’ use of negative emotion words in picturebook reading, and their belief in an emotion coaching teaching style were positively associated with T2 emotion knowledge, even with age, gender, and T1 emotion knowledge partialled. Contributions of punitive and problem-solving reactions to children’s negative emotions were also moderated by classroom risk status: Punitive reactions were negatively associated with T2 emotion knowledge only for children in low-risk classrooms, and problem-solving reactions were positively related to T2 emotion knowledge only for children in high-risk classrooms. The positive reactions aggregate also was associated with greater T2 emotion knowledge for children in high-risk classrooms (β= 0.16, p < .05); the negative reactions aggregate was positively associated with T2 emotion knowledge for children in high-risk classrooms, but negatively for children in low-risk classrooms (β= 0.22, p < .05). Finally, teachers’ use of positive emotion words and clarifying emotion language during picturebook reading were negatively associated with T2 emotion knowledge in high-risk classrooms, whereas use of negative terms during picturebook reading was positively associated with end-of-year emotion knowledge in these classrooms.

Discussion

These findings shed light on how teacher emotion socialization contributes to individual differences in preschoolers’ emotion knowledge. Teacher emotion socialization, especially as moderated by classroom socioeconomic risk, often predicted emotion knowledge, independent of child covariates, including children’s beginning-of-year emotion knowledge. Importantly, only one finding resembled the parent emotion socialization literature, and one appeared unique to the preschool environment; most were specific to classrooms where children experienced socioeconomic risk. The pattern of moderation findings also highlights the importance of context in teachers’ emotion socialization, especially since most studies with parents and the few that exist with teachers heretofore have been undertaken with children not living in poverty.

Main Effect Contributions of Teacher Emotion Socialization

The pattern of teacher predictors showed only one similarity with parental socialization findings in the literature, following the tenets of emotion socialization theory. As is true for families, lack of punitive reactions facilitated development of children’s emotion knowledge (Perlman, Camras, & Pelphrey, 2008). However, another finding ran counter to those generally found in the parent emotion socialization literature (Denham, Mitchell-Copeland, Strandberg, Auerbach, & Blair, 1997), and may be unique to the classroom context: Children displayed greater emotion knowledge when teachers reacted with minimization of their negative emotions.

Why might this pattern of findings occur? In considering potential mechanisms at work supporting our findings we should think about the socialization messages children are likely receiving. Teachers may use punitive reactions to quell children’s emotional outbursts, because dealing with multiple emotional preschoolers at any one time - over weeks - is not easy. Young children, when faced with a teacher who punishes their own and others’ emotions, might be aroused by these interactions (as they may be with parents; Eisenberg et al., 1998), rendering them somewhat avoidant of emotional information surrounding expression or experience of emotion.

In contrast, teachers’ minimization seemed to convey a different message than that documented in the literature for parents. When teachers minimize emotionality in the classroom (e.g., telling children that they are overreacting or behaving immaturely), such behavior may focus children’s attention on their own and others’ emotions, with subsequent attentiveness to teachers’ emotion socialization and consequent gains in emotion knowledge. Perhaps when parents use minimization, preschoolers feel disparaged, but in a classroom where a premium is placed on behavioral regulation of the group, the message’s meaning is different, less hurtful and more of a signal that “regulating your emotions in school is important”. Whether these teacher reactions, so often considered nonoptimal for families, continue to have salutary effects requires longer-term longitudinal investigation. Further, the contribution of teachers’ minimizing reactions might differ for toddlers or other aspects of social-emotional behavior such as prosocial behavior or compliance (King & La Paro, 2018; Morris et al., 2013); a fuller picture is important.

Interactive Contributions of Teacher Emotion Socialization

Significant teacher contributions to children’s emotion knowledge often were obtained for high-risk classrooms only. Arguably, these children could benefit most from teachers’ emotion socialization (see also Garner & Parker, 2018). Classroom positive emotional expressiveness, problem-solving reactions to emotions and an aggregate of positive reactions to children’s emotions, the use of negative emotion words in picturebook reading, and beliefs in a coaching socialization style were all positively associated with end-of-year emotion knowledge in high-risk classrooms. Unlike earlier developing knowledge of positive emotions, negative emotion knowledge develops throughout preschool (Denham & Couchoud, 1990). Teachers’ references to negative emotions during book reading thus may be especially useful to children living at socioeconomic risk, given their needs in the area of emotion knowledge and their exposure to negative emotion (Garner, 2006; Raver et al., 2015).

Overall, then, when teachers in high-risk classrooms reported valuing and enacting positive aspects of emotion socialization, children’s emotion knowledge was fostered. Why might these techniques be especially salutary for these children living at socioeconomic risk? As suggested by several of our findings where risk was negatively associated with end-of-year emotion knowledge, it may be that the emotion knowledge of children at low socioeconomic risk is already progressing somewhat independently of teachers’ emotion socialization input.

Further, considering that both children and teachers in these high-risk classrooms are more likely to be African American than in the low-risk classrooms, an examination of these findings from the perspective of ethnicity and culture is warranted. Although emotion socialization has been most often studied with European American families and teachers, it has been noted that African American teachers may be especially open to emotions, accentuating emotion socialization in their teaching (Parker et al., 2012). In earlier work (Denham, Bassett, & Miller, 2017), we found African American teachers to be more likely to endorse these self-reported positive emotion socialization attributes. Our current findings go further to find that these very practices support emerging emotion knowledge for children in high-risk classrooms.

A contrasting but important interactive effect also was found. Although teacher-reported punitive reactions to children’s emotions were associated with lower emotion knowledge in low-risk classrooms, these reactions were not deleterious to the emotion knowledge of children in high-risk classrooms. In fact, the “nonsupportive” reaction aggregate was positively associated with end-of-year emotion knowledge for children in the high-risk classrooms. Thus, African American teachers’ focus on emotions may require a more nuanced view of adaptive emotion socialization (Labella, 2018; Morelen & Thomassin, 2013), in which “celebration and restriction of children’s emotion coexist closely…, perhaps reflecting the joint influences of traditional Afro-cultural values and the historical context of slavery and discrimination” (Labella, 2018, p. 1).

There is support for this assertion in the parent emotion socialization literature. For example, African American mothers emphasize, especially with their sons, the negative social consequences of showing negative emotions in a discriminatory society. They report more “nonsupportive” and less “supportive” attitudes towards the emotionality of their children than do European American mothers, perhaps using emotionally strict reactions to keep children safe (Nelson, Leerkes, O’Brien, Calkins, & Marcovitch, 2012; see also Parker et al., 2012). In recent observational research (Denham & Bassett, 2019), we also found that largely African American teachers of high-risk classrooms reacted behaviorally to children’s emotions in ways that historically have been considered both “supportive” and “nonsupportive” of the development of emotional competence, with both making positive contributions to children’s observed positive social-emotional behaviors in this context. Taken together, these dovetailing findings from both our studies suggest that the meaning of reactions traditionally termed “nonsupportive” may have differed in a racially and culturally divergent context. Adaptive emotion socialization practices may look different for different racial, ethnic, cultural, and/or income groups.

Given these considerations, a very complex picture can be painted of these teachers valuing positive emotion socialization, but also performing some unique behaviors necessary and useful in their specific context. So, children in high-risk classrooms are receiving a complex message about emotions, one that includes both punitive and minimizing reactions to their emotions, as well as the “supportive” reactions and practices already mentioned. This message may be something like this: “I am cared for. Emotions are ok - I can show them and learn about them, but I need to regulate them, too.” Thus, a unified model of ethnic and emotion socialization, including more inclusive terminology, is required in future research (Dunbar et al., 2017). “Positive” and “negative”, or “supportive” and “nonsupportive” emotion socialization can be considered ethnically-bound terms, not universally applicable to individuals from differing ethnic and socioeconomic backgrounds. Researchers of emotion socialization must consider carefully the terminology and logic models underlying their predictions; partnering with early childhood educators, researchers also must become culturally competent. Value-laden terminology such as “supportive” and “nonsupportive” reactions should be re-named, perhaps as “adaptive” and “nonadaptive”, with definitions varying by context (e.g., including race).

Limitations and Future Research

There are, as with any research, methodological and analytical limitations that bear on conclusions from our findings. First, we were enjoined from asking questions about family income. This injunction is not unusual; perhaps, however, knowing fuller socioeconomic information on actual income, chaos level in the home, and material hardship could add to our understanding in future research. In fact, adding individual child socioeconomic risk, perhaps using several indices such as parents’ education and income and community aspects such as neighborhood safety could be important to include, especially given the important moderation by socioeconomic risk found here. Finally, having more balanced comparison groups of low- and high-socioeconomic risk groups could yield more statistically stable findings.

Second, in the current study we did not include race/ethnicity of teacher or child in our analyses because classroom risk status also formed a reasonable proxy marker for this attribute. Further, given the nested-nature of our data (i.e., children within classrooms), dividing our sample according to the teachers’ race would have run counter to common multilevel modeling approaches due to the smaller number of African-American teachers. Although this study was not designed to look specifically at racial differences, we value the unique cultural influences that African-American teachers bring to their classrooms (Garner, 2006). Thus, future investigations could include the race/ethnicity of both children and teachers, examining unique and joint contributions, and consider contributions of teachers from diverse cultural groups.

Third, further consideration of the mechanisms behind these findings, particularly the interactions of emotion socialization and socioeconomic risk, is also appropriate. Designs with more detail about specific teacher-child emotional transactions (e.g., learning context, verbalizations involved, specific children’s responses) could be useful toward this goal. Mixed-method designs also could be useful, especially regarding views on emotion socialization techniques in nonEuropean American teachers, triangulating qualitative information on teachers’ emotion socialization with quantitative information like that reported here. Other ways in which a more detailed picture may be flushed out include, for example: (a) determining whether different patterns of emotion socialization support end-of-year emotion knowledge for children low versus high in beginning-of-year emotion knowledge (see Bassett et al., in press); (b) delineating how parental and teacher emotion socialization might compare and contrast, and even interact; and (c) proceeding from our current correlational findings, examining whether and, if so, how, emotion socialization beliefs form the structural substrate for observed and reported emotion socialization modeling, reacting, and teaching practices.

Fourth and finally, the picturebook storytelling context may be a particularly rich one for the acquisition of emotion knowledge and bears further research explicitly targeting its contribution to emotion knowledge. We know that parents can elicit toddlers’ and preschoolers’ emotion language in such a setting with beneficial effects on their social-emotional behavior (Drummond, Paul, Waugh, Hammond, & Brownell, 2014) and emotion knowledge (Martin & Green, 2005). Moreover, in comparison with other settings, such as joint play, more elaborative discussion of emotion transpires during book reading - probably because the plot of even a wordless book, replete with illustrations of emotions, can allow adults to highlight emotions in a way most other contexts cannot. The freedom of storytelling rather than reading also may lend itself to wider-ranging emotion discussions (Ziv et al., 2015). Further, children can identify with and share the emotions of a story character within this setting, applying story characters’ experiences to their own emotional challenges (Garner & Parker, 2018). Given that training in social-emotional learning is lacking in teacher preparation (Schonert-Reichl, Kitil, & Hanson-Peterson, 2017), teacher preparation programs’ pre-existing focus on book reading might provide an opportunity for more specific promotion of emotion knowledge (Garner & Parker, 2018). If replicated, current findings suggesting that, for children at socioeconomic risk, positive emotion knowledge outcomes are related to teachers’ focus on negative emotion terms, rather than the developmentally more elementary positive emotion terms and clarifying language, could guide teachers’ book reading efforts.

Potential Applications

Even given our preliminary interpretation of these findings, some suggestions can be made for optimizing preschool teacher training and practice. Many early childhood teachers are intuitively aware of the importance of their own as well as children’s emotions to learning and wellbeing, and closely attend to these issues in the classroom. However, this is not always the case; there are, as we found, differences in teachers’ enactment of adaptive practices (Zembylas, 2007; Zinsser, Denham, Curby, & Shewark, 2015; Zinsser, Shewark, Denham, & Curby, 2014). Thus, teachers and their supportive administrators, as well as pre-service teachers, could profit from attention to and training in these issues (Garner, 2010; Waajid, Garner, & Owen, 2013).

First, ways that teachers deal with their own emotional lives - perceiving emotions of self and others, using emotions to facilitate cognition and action, understanding emotions, and managing them - undoubtedly contribute to their socialization of children’s emotion knowledge. A more emotionally aware teacher could more usefully talk about emotions with children. Further, preschool teachers’ emotional competence is related to their reactions to children’s emotions; lack of emotional awareness has been associated especially with nonoptimal emotion socialization techniques (Ersay, 2007, 2015).

Second, many professional development techniques could contribute to advances in early childhood educators’ modeling of emotions, contingent reactions, teaching, and positive beliefs about emotions. Jennings and Greenberg (2009) have suggested ways to promote teachers’ emotional competence, including: (a) mindfulness training to maintain positivity, be willing to accurately express emotions, and modulate understandable negative emotions (Jennings, 2015; Kemeny et al., 2012; Shewark et al., 2018; Zinsser, Denham, Curby, & Shewark, 2015; Zinsser, Shewark, Denham, & Curby, 2014); (b) reflective supervision to understand their own emotions and gain access to a broader emotion vocabulary, increasing their ease in discussing feelings (Gilkerson, 2004); (c) stress reduction to aid in reacting optimally to children’s emotions (Buettner, Jeon, Hur, & Garcia, 2016); and (d) direct training. Regarding direct training, Kremenitzer (2005; Kremenitzer & Miller, 2008) gives excellent, concrete suggestions on how teachers can become aware of their own emotional competence and its effects on children, especially via “emotional intelligence journaling”. These techniques could also help teachers to focus on valuing teacher-child emotion conversations and sustaining interchanges about emotions in classroom activities. As well, they could help promote emotion dialogues with children about ongoing classroom interactions. In all these considerations, the culturally/ethnically bound nature of emotion socialization must be addressed sensitively.

Conclusion

Our research is among the first to examine mechanisms of teacher emotion socialization in their contribution to young children’s emotion knowledge, including the important moderating impact that socioeconomic risk can have on the development of emotion knowledge. We have found ways in which such teacher-reported behaviors and beliefs are important for the development of emotional competence, ways that may differ from those documented in the parent emotion socialization literature, and ways that may be context-dependent. As noted by Jones and Bouffard (2012), all these varying contributions constitute everyday strategies based on kernels of evidence, “essential ingredients” compared to the “brands” of curricula. Continuing to pinpoint and elucidate these behaviors can benefit both teachers and children in the crucial promotion of emotion knowledge for both.

Table 5.

Hierarchical Linear Modeling Analyses: Teacher Reactions Predicting AKT-S

TCCNES Predictors	AKT-S
Fixed Effects, Level 2
Intercept	0.21
Risk	−1.66***
Punitive Reactions	−0.12*
Minimizing Reactions	0.14*
Emotion Encouragement	−0.00
Emotion-Focused Reactions	0.02
Problem-Focused Reactions	−0.06
Risk X Punitive Reactions	0.18+
Risk X Minimizing Reactions	−0.02
Risk X Emotion Encouragement	0.08
Risk X Emotion-Focused Reactions	0.01
Risk X Problem-Focused Reactions	0.13+
Level-1 % Variance Explained	40.8+

⁺p < .10,

^*p < .05,

^**p < .01,

^***p < .001.

Note. TCCNES = Teachers’ Contingent Reactions to Children’s Negative Emotions Scale. AKT = total AKT-S score.

Table 6.

Hierarchical Linear Modeling Analyses: Teacher Emotion Language During Bookreading Predicting AKT-S

Teacher Book Reading Predictors	AKT-S
Fixed Effects, Level 2
Intercept	−0.11
Risk	0.32
Total Positive Words	0.00
Total Negative Words	0.00
Percent Comments	0.00
Percent Explanations	0.00
Percent Clarifying	0.00
Risk X Positive Words	−0.04**
Risk X Negative Words	0.02+
Risk X Comments	0.00
Risk X Explanations	−0.01
Risk X Clarifying	−0.02*
Level-1 % Variance Explained	36.7*

⁺p < .10,

^*p < .05,

^**p < .01.

Note. AKT = total AKT-S score.

Highlights.

Teachers’:

• Minimizing reactions contributed positively to preschoolers’ emotion knowledge (EK)

• Punitive reactions to emotion contributed negatively to preschoolers’ EK

• Talk about negative emotions predicted EK for preschoolers at risk

• Positive expressiveness and emotion coaching predicted EK for preschoolers at risk

• Punitive reactions predicted EK for preschoolers not at risk