Predictors and Outcomes of Mealtime Emotional Climate in Families With Preschoolers

Predictors and Outcomes of Mealtime Emotional Climate in Families With Preschoolers Abstract Objective Mealtime emotional climate (MEC) is related to parent feeding and mental health, and possibly to child food consumption. However, MEC has been inconsistently assessed with a variety of coding schemes and self-report instruments, and has not been examined longitudinally. This study aims to characterize MEC systematically using an observational, count-based coding scheme; identify whether parent feeding or mental health predict MEC; and examine whether MEC predicts child food consumption and weight. Methods A subsample of parents (n = 74) recruited from a larger study completed questionnaires when children were about 37 months, participated in a home visit to videotape a mealtime when children were about 41 months, and completed questionnaires again when children were about 51 months old. Maternal and child positive and negative emotions were coded from videotaped mealtimes. Observational data were submitted to cluster analyses, to identify dyads with similar emotion expression patterns, or MEC. Logistic regression was used to identify predictors of MEC, and Analysis of Covariance was used to examine differences between MEC groups. Results Dyads were characterized as either Positive Expressers (high positive, low negative emotion) or All Expressers (similar positive and negative emotion). Increased food involvement feeding practices were related to decreased likelihood of being an All Expresser. Positive Expressers reported that their children ate more healthy food, compared with All Expressers. Conclusions Observed MEC is driven by maternal emotion, and may predict child food consumption. Food involvement may promote positive MEC. Improving MEC may increase child consumption of healthy foods. body weight, child, emotions, family, longitudinal studies, meals, observational methodology, parents Introduction Increased consumption of healthy foods—such as fruits, vegetables, and lean protein sources—is linked to lower risk for chronic disease (Bazzano et al., 2002; Esmaillzadeh et al., 2006; Wang et al., 2014). Healthy food consumption declines from childhood to young adulthood (Demory-Luce et al., 2004) but is relatively stable from adolescence to adulthood, pointing to the importance of early habit formation and maintenance (Lien, Lytle, & Klepp, 2001). Parents control and shape the home food environment, where children spend the majority of their time during this critical developmental phase (American Academy of Pediatrics, 2003; Anzman, Rollins, & Birch, 2010; Fitzgibbon, Hayman, & Haire-Joshu, 2008; Krishnamoorthy, Hart, & Jelalian, 2006). Family mealtimes provide a window of opportunity to observe whether certain behaviors or interactions—from both parents and children—are linked to food consumption and weight outcomes (Fiese, Foley, & Spagnola, 2006). Increased family mealtime frequency is consistently related to healthy food consumption and lower risk for overweight in young children (Anderson & Whitaker, 2010; Fulkerson, Larson, Horning, & Neumark-Sztainer, 2014; Hammons & Fiese, 2011). Mealtime quality also predicts child healthy food consumption (Fiese, Hammons, & Grigsby-Toussaint, 2012). Interpersonal interactions, maternal emotional expression, mental health, and marital relationship quality have been examined previously as indicators of mealtime quality (Berge et al., 2014; Hughes et al., 2011; Lytle, Seifert, Greenstein, & McGovern, 2000; Rollins, Belue, & Francis, 2010). For example, increased maternal negative emotion expression (measured by assessing tone, body language, and facial expression) is linked to indulgent feeding styles and higher child weight (Hughes et al., 2011). Furthermore, maternal depression is related to observed controlling feeding practices (Haycraft, Farrow, & Blissett, 2013), and to self-report of fewer positive mealtime practices, such as being present while the child eats, child control of snacking, and food resource management (McCurdy, Gorman, Kisler, & Metallinos-Katsaras, 2014). Observational studies suggest that the mealtime emotional climate (MEC)—positive and negative emotion expression frequency—is linked to feeding and weight (Berge et al., 2014; Fosco & Grych, 2013; Hughes et al., 2011). For instance, families with more positive interpersonal dynamics (more warmth, higher relationship quality, and group enjoyment) have children with lower weight, as compared with families with more negative interpersonal dynamics (more hostility, food lecturing/moralizing, and indulgence; Berge et al., 2014). Furthermore, studies with clinical samples have found success using behavioral interventions focused on feeding to improve mealtime affective management (Janicke, Mitchell, Quittner, Piazza-Waggoner, & Stark, 2008). Collectively, findings suggest that emotional characteristics of mealtimes are modifiable, and may increase healthy food consumption or improve weight outcomes. However, measures of MEC vary, limiting our ability to compare across studies. Global observational schemes have been primarily used to examine MEC as a function of maternal (but not child) emotion expression (Hughes et al., 2011), or family- and dyad-relationship quality (Berge et al., 2014). Although global codes garner macro-level descriptions of behavior, it is difficult to measure changing dynamics using a global rating-based coding scheme (Chorney, McMurtry, Chambers, & Bakeman, 2015; Yoder & Symons, 2010). Self-report measures or global coding schemes may yield foundational information about emotional climate, but these measures do not always depict a thorough representation of emotional expression. For instance, it is logical to expect that emotional climate depends—in part—on the proportion of positive and negative emotions expressed by family members during the meal. Therefore, an observational count coding system may provide a better alternative. Nevertheless, to the best of our knowledge, no previous studies have used this approach to measure MEC. Additionally, mealtimes may be especially challenging for parents who use unhealthy feeding practices or with mental health problems (Hughes et al., 2011; McPhie, Skouteris, Daniels, & Jansen, 2014). Therefore, it is important to examine how parenting may influence MEC. Finally, all previous studies examining MEC have been cross-sectional; thus, longitudinal studies are needed. The current study has three aims designed to address these limitations and to contribute to the body of knowledge regarding promotion of health and health behaviors in young children. The first aim of the present study is to characterize MEC using an observational count coding scheme, by examining the proportional frequency of maternal and child emotional expression during home-based family mealtimes. Given the focus in the literature on maternal emotional expression as an indicator of mealtime climate, we will examine whether child emotional expression also contributes to MEC. Second, we will examine whether parent feeding or mental health symptoms predict MEC. Finally, we will examine whether MEC predicts child food consumption or weight longitudinally. This work will contribute to our current understanding of the socioemotional context of eating behavior, by examining how observed emotional expression during home-based mealtimes is related to preschool-aged children’s eating behaviors. Methods Participants Participants for this study were drawn from the larger Synergistic Theory and Research on Obesity and Nutrition Group (STRONG) Kids Panel Study. The larger STRONG Kids Study is a three-wave prospective panel survey designed to examine transdisciplinary predictors of child health behaviors and weight. Primary caregivers of preschoolers were recruited from childcare centers in East Central Illinois that were registered with the state Bureau of Child Care and Development, and enrolled at least 24 children (Harrison, Liechty, & The STRONG Kids Program, 2012). Children were about 37 months at Wave 1 (W1), 51 months at Wave 2 (W2), and 70 months at Wave 3 (W3). Parents completed self-report questionnaires, and children’s height and weight were measured by trained research assistants at each wave. Response rate varied by preschool, ranging from 60 to 95%. This study reports on a subsample (n = 74) of families in the STRONG Kids study that consented to participate in a 2-hr home visit when children were about 41 months old (SD = 5.23), after submitting surveys in W1. Response rate for the subsample was 52% of the eligible families that were contacted (n = 143). Home visit recruitment targeted families who had already completed surveys and measurements at W1 (Table I). Parents were offered $50 for remuneration for the home visit. This study was approved by the university institutional review board. Sample characteristics are reported in Table I. Table I. Sample Characteristics (n = 74) Demographic characteristics  n  %    n (M)  % (SD)  Range  Marital status      Employment status         Single  9  12   Employed  13  18     Married  59  80   Self-employed  32  43     Separated/divorced  3  4   Stay-at-home parent  3  4     Civil union  1  1   Student  3  4     Cohabiting  1  1   Retired  9  12    Race/ethnicitya       Out of work  11  15     Black or African American  7  10  Education         White  56  76   High school or less  3  4     Asian  3  4   Some college  16  22     Hispanic/Latino  8  11   College graduate  19  26     Mixed ethnicity  5  7   Postgraduate work  36  47    Household income      Age         $24,999 or less  9  12   W1 parent age  33.6  5.6  27.9   $25,000–$39,999  10  14   W1 child age (months)  37.2  6.2  29.0   $40,000–$69,999  11  15   W2 child age  51.7  7.3  39.2   $70,000–$99,999  22  30   W3 child age  70.3  7.2  34.0   $100,000 or more  19  26   Home visit child age  41  5.2  19.0  Child gender      Parent gender         Male  37  50   Male  7  9     Female  37  50   Female  67  91    Demographic characteristics  n  %    n (M)  % (SD)  Range  Marital status      Employment status         Single  9  12   Employed  13  18     Married  59  80   Self-employed  32  43     Separated/divorced  3  4   Stay-at-home parent  3  4     Civil union  1  1   Student  3  4     Cohabiting  1  1   Retired  9  12    Race/ethnicitya       Out of work  11  15     Black or African American  7  10  Education         White  56  76   High school or less  3  4     Asian  3  4   Some college  16  22     Hispanic/Latino  8  11   College graduate  19  26     Mixed ethnicity  5  7   Postgraduate work  36  47    Household income      Age         $24,999 or less  9  12   W1 parent age  33.6  5.6  27.9   $25,000–$39,999  10  14   W1 child age (months)  37.2  6.2  29.0   $40,000–$69,999  11  15   W2 child age  51.7  7.3  39.2   $70,000–$99,999  22  30   W3 child age  70.3  7.2  34.0   $100,000 or more  19  26   Home visit child age  41  5.2  19.0  Child gender      Parent gender         Male  37  50   Male  7  9     Female  37  50   Female  67  91    Note. W1 = Wave 1; W2 = Wave 2; W3 = Wave 3. a Race/ethnicity variables were not mutually exclusive, and a small number of participants indicated multiple ethnicities. Home Visit Procedures Parents provided informed consent for mealtimes to be video-recorded before and during home visits. Researchers stayed in the home before the mealtime to build rapport. Researchers left the home during the mealtime, and returned after the mealtime was finished, to prevent social desirability and observer biases from influencing behavior. Videotapes of mealtimes were immediately deidentified and uploaded onto a secure server. Observational Coding Procedures Training and Reliability Two research assistants were trained on the D.O.T.S. emotion coding system (Cole, Wiggins, Radzioch, & Pearl, 2007) until attaining adequate inter-rater reliability (Intra-class correlation ≥ .70) on the maternal and child positive and negative affect count codes. During this process, decision rules were established and disagreements were discussed. Observed agreement for 20% of the cases that were double coded after training were Intra-class correlation = .73 to .90. Coding Schemes The frequency of positive and negative affect from mothers and children was coded using the D.O.T.S emotion coding system (Cole et al., 2007). Videos were watched in separate sessions to code maternal and child affect. Behaviors, facial expressions, and vocalizations were coded. Examples of positive emotion expressions for mothers included smiling, speaking in enthusiastic tones, laughing, and expressing physical affection; whereas sighing, furrowing brows, or yelling were considered negative emotion expressions. Examples of positive emotion expressions for children included giggling, sustained smiling, and bouncing; and whining, quivering lips, or crying were considered negative emotion expressions. Mealtime Length Meal start time was indicated by food being placed on the table or—if the family did not eat at the table—by presence of food in the eating environment. Meal end time was indicated by the last time food is taken away from the target child, or the last time the target child leaves the table. Mealtime length is calculated according to meal start and end time. Measures Child Weight Child height and weight were measured at each time point by trained research assistants using a digital scale and stadiometer (Kuczmarski et al., 2000). Sex- and age-adjusted body mass index percentile (BMI-P) was calculated from W1, W2, and W3 data. Food Consumption Child food consumption was assessed using six items from the Early Childhood Longitudinal Study, Birth Cohort (ECLS-B; National Center for Educational Statistics, 2007). Parents reported how often their children ate fruits, vegetables, sugar-sweetened beverages, fruit juice, fast food, soy products, French fries, candy/sweets, and salty snacks over the past 7 days, accounting for meals and snacks eaten at home, at school, or in any other situation. Response options ranged from 0 to 7, with 0 as “my child did not eat/drink any _____ during the past 7 days,” 1 as “my child ate/drank ____ once a day,” 2 as “twice a day,” 3 as “three times a day,” 4 as “four or more times a day,” and 7 as “I don’t know,” which was treated as missing data. Responses that indicated 5 or 6 were recoded into fractions to represent consumption by times per day (5 was “1 to 3 times during the past 7 days” = 0.2857 times daily; 6 was “4 to 6 times during the past 7 days” = 0.7143 times daily). Food consumption variables were constructed based on a two-factor solution identified by a principal component analysis (PCA) with varimax rotation that included each of these items from the Wave 2 questionnaire (42% of variation explained). The rotated component matrix suggested that fresh fruit, vegetables, and soy-based foods comprised one component, whereas fruit juice, sugar sweetened beverages, fast food, sweets, French fries, and salty snacks comprised another component. All items loaded ≥0.3 on the identified components. Composite scores were constructed based on the findings of the PCA. A composite score for daily healthy food consumption was made by summing responses to items about fruit, vegetable, and soy consumption. A composite score for daily unhealthy food consumption was made by summing responses to items about sugar-sweetened beverage, fruit juice, fast food, French fries, candy/sweet, and salty snack consumption (Table II). Healthy food consumption was correlated at W1 and W2 (r = .50, p < .001), and unhealthy food consumption was correlated at W1 and W2 (r = .49, p < .001). Healthy and unhealthy food consumption was not correlated at either W1 or W2. Table II. Model Variable Statistics by MEC Group Membership (n = 74) Variable  Positive Expressers (n = 59)   All Expressers (n = 15)   Full Sample (n = 74)   Kruskal–Wallis H testsa   M  SD  M  SD  M  SD  X2  p  Observed mother emotion                   Positive emotion ratio  0.92  0.10  0.41  0.22  0.82  0.24  36.16  <0.01   Negative emotion ratio  0.08  0.10  0.53  0.24  0.17  0.23  28.97  <0.01   Total emotions  18.97  9.50  12.40  8.68  17.64  9.66  7.80  0.01  Observed child emotion                   Positive emotion ratio  0.73  0.21  0.67  0.13  0.72  0.20  2.40  0.12   Negative emotion ratio  0.27  0.21  0.33  0.13  0.28  0.20  2.40  0.12   Total emotions  19.57  10.91  19.87  9.32  19.62  10.55  0.00  0.98  W1 parent BMI  25.75  5.11  26.65  6.01  25.93  5.27  0.12  0.73  Child BMI-P                   W1 child BMI-P  59.10  21.22  65.36  22.81  60.37  21.55  1.09  0.30   W2 child BMI-P  57.63  24.57  67.86  19.09  59.70  23.80  2.01  0.16   W3 child BMI-P  59.68  25.67  70.55  15.54  61.89  24.28  1.17  0.28  Food consumptionb                   W1 healthy food  3.80  1.31  3.09  1.29  3.66  1.33  3.03  0.08   W1 unhealthy food  3.02  1.43  3.38  1.32  3.09  1.41  1.32  0.25   W2 healthy food  3.60  1.50  2.66  1.12  3.41  1.47  5.67  0.02   W2 unhealthy food  2.68  1.49  2.67  0.79  2.68  1.37  0.21  0.65  Feeding practices (W1)                   Food Involvement  3.10  0.85  2.44  0.92  2.96  0.90  6.54  0.01   Restriction for Health  2.86  0.81  3.11  1.01  2.91  0.85  1.05  0.31   Restriction for Weight  1.62  0.54  1.66  0.45  1.63  0.52  0.44  0.51   Environment  3.97  0.61  3.67  0.63  3.91  0.62  2.20  0.14   Pressure to Eat  2.52  0.75  2.56  0.74  2.53  0.74  0.02  0.91   Monitoring  4.24  0.82  3.95  0.90  4.18  0.84  1.41  0.24   Modeling  3.86  0.75  3.59  0.81  3.81  0.77  1.84  0.18   Balance/Variety  4.41  0.56  4.31  0.75  4.39  0.60  0.06  0.81  Parent mental health (W1)                   Depressive symptoms  1.20  0.21  1.54  0.70  1.27  0.39  3.24  0.07   Stress symptoms  1.69  0.57  1.84  0.77  1.72  0.61  0.18  0.67  Total family members  3.59  0.77  4.13  0.64  3.70  0.08  5.79  0.02  Mealtime length  24.17  8.40  22.86  8.67  23.90  8.41  0.37  0.54  Variable  Positive Expressers (n = 59)   All Expressers (n = 15)   Full Sample (n = 74)   Kruskal–Wallis H testsa   M  SD  M  SD  M  SD  X2  p  Observed mother emotion                   Positive emotion ratio  0.92  0.10  0.41  0.22  0.82  0.24  36.16  <0.01   Negative emotion ratio  0.08  0.10  0.53  0.24  0.17  0.23  28.97  <0.01   Total emotions  18.97  9.50  12.40  8.68  17.64  9.66  7.80  0.01  Observed child emotion                   Positive emotion ratio  0.73  0.21  0.67  0.13  0.72  0.20  2.40  0.12   Negative emotion ratio  0.27  0.21  0.33  0.13  0.28  0.20  2.40  0.12   Total emotions  19.57  10.91  19.87  9.32  19.62  10.55  0.00  0.98  W1 parent BMI  25.75  5.11  26.65  6.01  25.93  5.27  0.12  0.73  Child BMI-P                   W1 child BMI-P  59.10  21.22  65.36  22.81  60.37  21.55  1.09  0.30   W2 child BMI-P  57.63  24.57  67.86  19.09  59.70  23.80  2.01  0.16   W3 child BMI-P  59.68  25.67  70.55  15.54  61.89  24.28  1.17  0.28  Food consumptionb                   W1 healthy food  3.80  1.31  3.09  1.29  3.66  1.33  3.03  0.08   W1 unhealthy food  3.02  1.43  3.38  1.32  3.09  1.41  1.32  0.25   W2 healthy food  3.60  1.50  2.66  1.12  3.41  1.47  5.67  0.02   W2 unhealthy food  2.68  1.49  2.67  0.79  2.68  1.37  0.21  0.65  Feeding practices (W1)                   Food Involvement  3.10  0.85  2.44  0.92  2.96  0.90  6.54  0.01   Restriction for Health  2.86  0.81  3.11  1.01  2.91  0.85  1.05  0.31   Restriction for Weight  1.62  0.54  1.66  0.45  1.63  0.52  0.44  0.51   Environment  3.97  0.61  3.67  0.63  3.91  0.62  2.20  0.14   Pressure to Eat  2.52  0.75  2.56  0.74  2.53  0.74  0.02  0.91   Monitoring  4.24  0.82  3.95  0.90  4.18  0.84  1.41  0.24   Modeling  3.86  0.75  3.59  0.81  3.81  0.77  1.84  0.18   Balance/Variety  4.41  0.56  4.31  0.75  4.39  0.60  0.06  0.81  Parent mental health (W1)                   Depressive symptoms  1.20  0.21  1.54  0.70  1.27  0.39  3.24  0.07   Stress symptoms  1.69  0.57  1.84  0.77  1.72  0.61  0.18  0.67  Total family members  3.59  0.77  4.13  0.64  3.70  0.08  5.79  0.02  Mealtime length  24.17  8.40  22.86  8.67  23.90  8.41  0.37  0.54  Note. BMI = body mass index; BMI-P = body mass index percentile; MEC = mealtime emotional climate; PCA = principal components analysis; W1 = Wave 1; W2 = Wave 2; W3 = Wave 3. a Kruskall–Wallis H tests examine mean differences between MEC groups on model variables. Bolded values indicate significant differences at p < .05. b Food consumption variables were constructed based on a two-factor solution identified by a PCA with varimax rotation on Wave 2 items (42% of variation explained). Rotated component matrix suggested that fresh fruit, vegetables, and soy-based foods comprised one factor, whereas fruit juice, sugary beverages, fast food, sweets, French fries, and salty snacks comprised another. Summed composite variables of healthy and unhealthy food consumption were constructed based on PCA findings. Feeding Practices Parent feeding practices were assessed at W1 and W2 using the Comprehensive Feeding Practices Questionnaire (CFPQ; Musher-Eizenman & Holub, 2007). The CFPQ consists of 49 items that correspond to 12 subscales: Child Control (α = 0.34), Emotion Regulation (α = 0.64), Balance/Variety (α = 0.72), Environment (α = 0.70), Food as Reward (α = 0.67), Child Food Involvement (α = 0.77), Modeling (α = 0.80), Monitoring (α = 0.87), Pressure to Eat (α = 0.71), Restriction for Health (α = 0.73), Restriction for Weight Control (α = 0.83), and Teaching about Nutrition (α = 0.51). Parents are asked to indicate how often they engaged in a particular feeding practice on a Likert scale from 1 (never) to 5 (always). Although subscale internal reliability was comparable with previously reported data (Musher-Eizenman & Holub, 2007), we elected to remove subscales with α < 0.70 (Food as Reward, Emotion Regulation, Child Control, and Teaching about Nutrition) to reduce measurement error. Maternal Mental Health Maternal mental health symptoms were assessed at W1 and W2 by the Depression, Anxiety, and Stress Scale (DASS-21; Henry & Crawford, 2005). The DASS consists of 21 items in a Likert-SCALE format. Depressive and Stress symptom subscales (seven items each) had adequate reliability in this sample (α = 0.84 and 0.87, respectively), but the Anxiety subscale did not (α = 0.62), and so was not included in analyses to reduce measurement error. Scores on the Depressive symptoms subscale ranged from 1.00 to 3.57 at W1, and 1.00 to 2.43 at W2. Scores on the Stress symptoms subscale ranged from 1.00 to 4.00 at W1, and 1.00 to 3.29 at W2. Covariates Demographic characteristics—including race/ethnicity, income, education, employment status, gender, age, and number of family members in the household—were self-reported by parents at W1 and W2 and included as controls in relevant analyses. Parent body mass index (BMI) was calculated from self-reported height and weight, and was included as a control in all analyses. Mealtime length was calculated from videotaped observations, and included as a control. Analysis Strategy Missing Data Missingness ranged from 0 to 16%. However, about 24% (n = 18) of W3 child BMI-P data were missing. We examined differences between cases missing and not missing W3 child BMI data. Salty snack consumption at W2 was associated with greater likelihood of missing W3 child BMI, but data were otherwise not missing systematically (Little’s Missing Completely At Random [MCAR] test X2 [df] = 1225.62 [8279], p = 1.00). Therefore, we applied multiple imputation with expectation maximization algorithms >20 iterations. Demographic variables were modeled as predictors only, and all other variables were modeled as both predictors and imputable outcomes. Imputed data sets were then aggregated. Analyses were also run on raw data with missing variables, with no significant differences between findings; therefore, we present results from imputed data to leverage the full sample. Additionally, given that observational coding focused on maternal–child interactions, all analyses were also conducted on a sample excluding the seven male primary caregivers who completed surveys at W1. All analyses were run in SPSS 24.0 (IBM Analytics, 2016) and probability levels were set at .05. Aim 1 To characterize MEC, we first created ratio variables by dividing the number of positive and negative emotions by the total number of emotions expressed by an individual, yielding four variables describing the proportion of positive and negative emotions expressed by mothers and children. Ratio variables were submitted to a Ward’s hierarchical cluster analysis using a squared Euclidean distance clustering method. After analyzing the agglomeration schedule and the dendrogram, a two-cluster solution was deemed the best fit for these data. To confirm this solution, we performed an additional K-means cluster analysis, specifying a two-cluster solution. The solution converged in 12 iterations, and the clusters identified by Ward’s method were not different from the clusters identified by the K-means method. Because K-means methods are sensitive to outliers (Steinley, 2006), cluster placement identified by Ward’s method was used in final analyses. Table II describes cluster characteristics. Aim 2 To examine predictors of MEC, we examined the effects of parent feeding practices and mental health on MEC group membership first using Kruskal–Wallis H tests (Table II), and then in binary logistic regression (Table III). Table III. Effects of Demographics, Depressive Symptoms, and Feeding Practices on Observed MEC Cluster Membershipa Predictors  Model 1   Model 2   β  SE  OR  95% CI   β  SE  OR  95% CI   LLCI  ULCI  LLCI  ULCI  Block 1: Demographic variables                       Constant  −5.17  3.05  0.01      −2.64  3.80  0.07       Child genderb  −0.43  0.65  0.65  0.18  2.32  −0.74  0.75  0.48  0.11  2.07   Parent genderb  0.41  1.22  1.51  0.14  16.33  0.20  1.36  1.22  0.09  17.40   Child age (months)  −0.01  0.05  0.99  0.89  1.11  0.01  0.06  1.01  0.89  1.14   Income  −0.24  0.24  0.78  0.49  1.26  −0.31  0.28  0.73  0.43  1.26   Total family members  1.40  0.60  4.03  1.25  13.03  0.83  0.61  2.31  0.70  7.60   Mealtime length  −0.03  0.04  0.97  0.90  1.06  −0.03  0.05  0.98  0.89  1.07  Block 2: Model variables                       W1 DASS: Depressive symptoms            1.90  0.98  6.65  0.98  45.10   W1 CFPQ: Food Involvement            −1.06  0.47  0.35  0.14  0.87  −2 log likelihood  64.51  53.49  Model X2  X2 = 8.71, df = 6, p = .19  X2 = 19.73, df = 8, p = .011  Nagelkerke R2  17.9%  37.7%  Hosmer and Lemeshow test  X2 = 3.64, df = 8, p = .88  X2 = 11.18, df = 8, p = .19  Classification accuracy  81.7%  88.7%  Predictors  Model 1   Model 2   β  SE  OR  95% CI   β  SE  OR  95% CI   LLCI  ULCI  LLCI  ULCI  Block 1: Demographic variables                       Constant  −5.17  3.05  0.01      −2.64  3.80  0.07       Child genderb  −0.43  0.65  0.65  0.18  2.32  −0.74  0.75  0.48  0.11  2.07   Parent genderb  0.41  1.22  1.51  0.14  16.33  0.20  1.36  1.22  0.09  17.40   Child age (months)  −0.01  0.05  0.99  0.89  1.11  0.01  0.06  1.01  0.89  1.14   Income  −0.24  0.24  0.78  0.49  1.26  −0.31  0.28  0.73  0.43  1.26   Total family members  1.40  0.60  4.03  1.25  13.03  0.83  0.61  2.31  0.70  7.60   Mealtime length  −0.03  0.04  0.97  0.90  1.06  −0.03  0.05  0.98  0.89  1.07  Block 2: Model variables                       W1 DASS: Depressive symptoms            1.90  0.98  6.65  0.98  45.10   W1 CFPQ: Food Involvement            −1.06  0.47  0.35  0.14  0.87  −2 log likelihood  64.51  53.49  Model X2  X2 = 8.71, df = 6, p = .19  X2 = 19.73, df = 8, p = .011  Nagelkerke R2  17.9%  37.7%  Hosmer and Lemeshow test  X2 = 3.64, df = 8, p = .88  X2 = 11.18, df = 8, p = .19  Classification accuracy  81.7%  88.7%  Note. CFPQ = Comprehensive Feeding Practices Questionnaire; CI = confidence interval; DASS = Depression, Anxiety, and Stress Scale; LLCI = lower-level confidence interval; MEC = mealtime emotional climate; OR = odds ratio; ULCI = upper-level confidence interval. Bolded lines indicate statistically significant findings. a Positive Expressers coded as 0, All Expressers coded as 1. Higher OR indicates greater likelihood of being an All Expresser. b For both parent and child gender, reference category was male. Aim 3 To examine whether MEC influences child food consumption at W2 or weight outcomes at W2 and W3. Preliminary analyses of mean differences were conducted with Kruskal–Wallis H tests, and post hoc analyses of covariance (ANCOVA) were used to differences in mean food consumption across MEC groups (Table IV). Covariates included child gender, parent gender, income, total number of family members, mealtime length, and change in the outcome variable from W1 to W2. Analyses met error variance equality assumptions. Table IV. Effects of MEC on Child Healthy Food Consumption, Controlling for Demographics and Change in Healthy Food Consumption From Wave 1 (W1) to Wave 2 (W2) Predictors  β  SE  η2  95% CI unstandardized β   Estimated marginal means (SE)a   LLCI  ULCI  Positive Expressers  All Expressers  Covariates                 Parent gender  0.13  0.48  0.03  −0.32  1.64       Child gender  −0.15  0.27  0.04  −0.99  0.09       Child age  −0.13  0.02  0.02  −0.07  0.02       Total family members  0.01  0.21  0.03  −0.16  0.69       Income  −0.18  0.11  0.05  −0.42  0.03       Meal length  0.17  0.02  0.04  −0.01  0.06       Change in healthy food consumption (W1–W2)  0.60  0.10  0.37  0.42  0.83      MEC  0.28  0.35  0.12  0.33  1.70  3.65 (0.15)  2.64 (0.30)  Predictors  β  SE  η2  95% CI unstandardized β   Estimated marginal means (SE)a   LLCI  ULCI  Positive Expressers  All Expressers  Covariates                 Parent gender  0.13  0.48  0.03  −0.32  1.64       Child gender  −0.15  0.27  0.04  −0.99  0.09       Child age  −0.13  0.02  0.02  −0.07  0.02       Total family members  0.01  0.21  0.03  −0.16  0.69       Income  −0.18  0.11  0.05  −0.42  0.03       Meal length  0.17  0.02  0.04  −0.01  0.06       Change in healthy food consumption (W1–W2)  0.60  0.10  0.37  0.42  0.83      MEC  0.28  0.35  0.12  0.33  1.70  3.65 (0.15)  2.64 (0.30)  Note. CI = confidence interval; LLCI = lower-level confidence interval; MEC = mealtime emotional climate; ULCI = upper-level confidence interval. Model R2 = .48, adjusted R2 = .41, F(1, 62) = 8.67, p < .01. Pairwise comparison mean difference between All Expressers and Positive Expressers on W2 healthy food consumption (mean difference = −1.01, SE = .35, p < .01). Bolded lines indicate statistically significant findings. a Estimated marginal means are adjusted by covariates in the model. All predictors have one degree of freedom. Effect sizes for ANCOVAs are reported as partial Eta squared (η2). Results Summary Statistics Descriptive statistics for model variables are reported in Table II. Mothers expressed positive emotions about 15.1 (SD = 9.37, range = 46) times on average over the course of the mealtime, and negative emotions about 2.5 (SD = 3.32, range = 18.5) times. Children expressed positive emotions about 14.6 times (SD = 9.2, range = 49.5) on average over the course of the mealtime, and negative emotions about 5.03 (SD = 3.94, range = 19.0) times. There were no significant or substantive differences in findings of any analyses when the seven male primary caregivers who completed surveys—but were not coded, as the coding scheme focused on mothers—were excluded or included. Aim 1: Characterize MEC Cluster analyses identified two distinct types of dyads. The first cluster accounted for 79% (n = 59) of the dyads in this sample, and was characterized by relatively high positive emotions in mothers and children, and relatively low negative emotions in mothers (Positive Expressers). The second cluster accounted for 20% (n = 15) of the dyads in this sample, and was characterized by a similar amount of positive and negative emotions in parents and children (All Expressers). Group characteristics are reported in Table II. Maternal emotions seemed to drive cluster placement. Kruskal–Wallis H tests found significant differences between groups on maternal positive emotion ratios and maternal negative emotion ratios. Although there were trends toward Positive Expresser dyads having children with higher positive and lower negative emotion ratios as compared with All Expressers, these differences were not statistically significant. Aim 2: Identify Predictors of MEC Kruskal–Wallis H tests (Table II) found significant differences between Positive Expressers and All Expressers on Food Involvement feeding practices and maternal depressive symptoms (Table II). Therefore, these variables were examined as potential predictors of MEC in logistic regression analyses. In the first block of logistic regression analyses, we examined demographics as independent variables, and MEC group membership as a dependent variable (Table III). Positive Expressers were coded as 0, and All Expressers were coded as 1. As the total number of family members increased, so too did the likelihood of being an All Expresser. No other variables had a significant effect. In the second block, we submitted the potential predictor variables identified in Kruskal–Wallis H tests to the logistic regression as independent variables, again examining MEC group membership as a dependent variable. Model fit and classification accuracy improved in this step. As food involvement increased, the likelihood of being an All Expresser decreased significantly. No other variables significantly predicted MEC group membership. Aim 3: Examine Outcomes of MEC Kruskal–Wallis H tests were used to examine mean differences on child outcome variables based on MEC group membership. The Positive Expressers group had significantly higher healthy food consumption at W2, as compared with All Expressers (Table II). There were no significant differences between groups on unhealthy food consumption or child BMI-P. Finally, ANCOVA was used to examine whether differences on W2 healthy food consumption between Positive and All Expressers persisted, controlling for covariates (Table IV). Controlling for child and parent gender, child age, total family members, income, mealtime length, and change in healthy food consumption from W1 to W2, Positive Expressers reported 1.01 (SE = 0.35, p = .005, 95% confidence interval  = 0.33, 1.70, R2 = .48; adjusted R2 = .41) more daily servings of healthy food at W2, compared with All Expressers. Discussion This study expands the scientific literature on the link between emotions and food in the family context in several ways. First, this is the first study to characterize family MEC using a count-based coding scheme and proportion scores, and to examine longitudinal predictors and outcomes of family MEC. Using this novel method, we found that MEC was determined primarily by maternal, not child, emotions. Dyads were categorized as either Positive Expressers (expressing high levels of positive emotion, and little negative emotion) or All Expressers (about equal levels of positive and negative emotion). Second, we aimed to examine whether feeding practices or maternal mental health were predictors of family MEC, given the prior research finding associations between parent emotional expression, depression, feeding, and mealtime practices (Haycraft, Farrow, & Blissett, 2013; Hughes et al., 2011; McCurdy et al., 2014). Food involvement feeding practices—but not maternal mental health or other feeding practices—predicted lower likelihood of being an All Expresser. Finally, this study examined whether family MEC predicted indicators of children’s health, including food consumption and weight outcomes. Positive Expressers reported that their children ate significantly more healthy food at W2, as compared with All Expressers, but there were no differences between groups by weight outcomes. Although interpretations of these results must be tempered by acknowledging some methodological limitations, they point to several avenues for future research and practice. Overall, findings suggest that MEC should be investigated as a potential, modifiable family-level factor to target in efforts to increase child consumption of healthy foods. Although cluster analyses accounted for maternal and child emotional expression, MEC groups had significantly different emotion profiles for mothers—but not children. Previous studies have found that family–child and parent–child interpersonal dynamics were linked to reduced child overweight, but that child–sibling interpersonal dynamics had no effect (Berge et al., 2014). Mothers’ emotions may have a larger effect on MEC overall because they exert more control over and spend more time engaged in orchestrating family routines during early childhood (Kotila, Schoppe-Sullivan, & Dush, 2013). Alternatively, children may have been more affected by the recording equipment during the mealtime than parents, tempering their negative emotional expression. It was surprising that a cluster of mothers and children with high negative emotion and low positive emotion did not emerge. Again, this may be because of social desirability biases, or the small and homogenous sample in the current study. We also found that parent report of food involvement feeding practices—but no other feeding practices, and not maternal mental health—was related to decreased likelihood of being an All Expresser. In a longitudinal study of 10–12-year-old Australian children, child involvement in food preparation, shopping, planning, or cleanup was not related to family mealtime frequency, dietary intake, or weight outcomes (Leech, McNaughton, Crawford, Campbell, Pearson, & Timperio, 2014). However, in a study among (n = 394) families with children aged between 18 months and 5 years, parent food involvement (or the importance, time, and value parents allocate to food) was associated with both child and parent fruit and vegetable consumption (Ohly et al., 2013). The Food Involvement subscale of the CFPQ includes three questions, all centered on the child’s involvement in food preparation, planning, and shopping (Musher-Eizenman & Holub, 2007). It is surprising that our results are somewhat contrary to the findings from Leech and colleagues’ (2014) study. However, involving younger children in food preparation and planning may present an opportunity for families to develop positive rapport around food and mealtimes. Indeed, a qualitative study with adult women found that negative affect was associated with lower food involvement for themselves (Jarman et al., 2012). Together, these findings suggest that the MEC may be determined, in part, by the degree to which parents are involved with food themselves, and the degree to which they involve their children with food. However, future research is needed to ascertain (1) whether parent report of child food involvement is indicative of true food involvement; (2) whether these effects are true only in young children; and (3) whether modifying food involvement could improve the MEC, and, ultimately, child food consumption. In contrast to the literature linking maternal depression to feeding and mealtime practices (Haycraft et al., 2013; McCurdy et al., 2014), we did not find associations with MEC group membership. On the one hand, it is possible that the sample was too small to detect differences in maternal mental health. Indeed, this sample’s range for stress and depressive symptoms at W2 is smaller than the range reported in a larger community sample of adults that completed the DASS (Crawford & Henry, 2003). Furthermore, although maternal depressive symptoms may serve as an indicator for maternal depression, the DASS cannot be used to diagnose individuals with depression and may not be comparable with previous studies using diagnostic assessments (Henry & Crawford, 2005). On the other hand, another study also found that parental depressive symptoms were not associated with mothers’ report of family relationship quality or family expressiveness, but was associated with use of fewer controlling parenting practices (Foster et al., 2008). Alternatively, it may be that family MEC is less associated with individual factors (such as maternal mental health), and more with factors related to the parent–child relationship (such as parenting and feeding practices), or the family system as a whole (such as food insecurity or household chaos; Fiese, Gundersen, Koester, & Jones, 2016). Finally, we found that parents in the Positive Expressers group reported greater child consumption of healthy food at W2, than All Expressers. There were no differences for unhealthy food consumption or child BMI-P. Previous studies found that parents of overweight children engaged in more hostility and food-related lecturing, as compared with families with non-overweight children (Berge et al., 2014). This may facilitate a counterproductive cycle, in which child overweight prompts hostility or food-related lecturing during mealtimes, which in turn may decrease healthy food consumption and eventually have an impact on weight (Francis, Hofer, & Birch, 2001; Webber, Cooke, Hill, & Wardle, 2010). Alternatively, if parents foster a more positive MEC, they may increase healthy food consumption by making the mealtime a more pleasant environment. Although other cross-sectional and longitudinal studies have found associations between parent–child interaction quality and risk for obesity and excessive weight gain (Anderson & Keim, 2016), this study found no association between family MEC and child weight outcomes. One possible explanation for this null finding is that the effects of MEC on weight as children age may be cumulative. For instance, although one study on families with children aged 6–12 years did find an association between the emotional atmosphere and risk for overweight (Berge et al., 2014), a narrative review found evidence to suggest that parent–child interaction quality in early childhood or infancy was more strongly and consistently associated with obesity risk during middle childhood, than during early childhood (Anderson & Keim, 2016). Links between parent–child interaction quality and weight outcomes were weak or nonsignificant for children under 5 years old, but associations seemed to increase in strength and consistency as children aged. Therefore, the effects of MEC on child weight may not be apparent until after the preschool years. Future research is needed to examine the longitudinal effects of MEC on child weight using observations across several time points. The current study has several strengths, including use of direct, home-based observations, a count-based coding scheme, and a longitudinal design. However, some limitations bear noting. First, it is possible that family members altered their behavior in response to the observation, although some dyads clearly expressed more negative emotions during the mealtime than others, suggesting that—while some families may have changed their behaviors according to social desirability—many did not. Second, measures assessing feeding practices, parent mental health, and child eating behaviors used self- or parent-report questionnaires, rendering these constructs susceptible to bias. In particular, parent report of child food consumption has been noted previously to have low internal reliability, and because there are currently no validated observational assessments of food consumption for children in the home environment, these methodological limitations may provide fruitful avenues for future research. Finally, our findings are based on a relatively small, homogenous community sample of mostly White mothers and their children, and should be replicated in a larger, more diverse sample of families. Additionally, although several male primary caregivers completed surveys, because we only coded maternal emotion during mealtimes, we cannot generalize our findings to fathers, and so focus our discussion on mothers. Clinical and Research Implications There are several implications from the findings of this study for practitioners and researchers. Results from this study will allow physicians and care providers to improve communication with parents about the health benefits of family mealtimes, by showing specific associations between family MEC and child food consumption. Given the challenging realities of mealtimes (e.g., food insecurity; picky eating) for many families (Malhotra, Herman, Wright, Bruton, Fisher, & Whitaker, 2013), it would be unreasonable to simply suggest that parents should express more positive emotion at meals. Instead, practitioners may find it helpful to provide parents with information about several evidence-based strategies that may reduce parent stress and negative affect during mealtimes, and indirectly improve child food consumption. For example, increasing food involvement, practicing child behavior management (Janicke et al., 2008; Stark et al., 2011), establishing routines and schedules (Jones, Fiese, & STRONG Kids Team, 2014), eliminating distractions (FitzPatrick, Edmunds, & Dennison, 2007), and engaging in direct, clear communication during mealtimes may all promote a more positive MEC (Fiese et al., 2006). Providers may be able to communicate more clearly about the benefits of family meals, by increasing awareness about the link between family MEC and child food consumption. This study identified also several avenues for future research. First, these findings should be replicated among a sample with more fathers, and in a larger and more diverse population, using multiple observations of family mealtimes in a longitudinal design. Second, studies using other measures of child food consumption (24-hr recalls, dietary intake diaries, and repeated food frequency questionnaires) are needed to validate links between family MEC and child food consumption. Third, investigators should examine how other family members or family factors (e.g., routines; distractions; communication) contribute to MEC, and ultimately, child health and well-being. Fourth, by using a coding and analytic approach that emphasizes behaviors on a moment-to-moment basis, researchers may be able to better understand the antecedents and consequences of positive and negative affect at the mealtime, thus deepening our understanding of the contributors to MEC. Finally, these preliminary findings suggest that more positive parent emotionality at meals may promote children’s healthy food consumption. Parenting interventions targeting emotion socialization have been remarkably successful at improving both parent and child emotional awareness and regulation (Havighurst, Wilson, Harley, Prior, & Kehoe, 2010; Wilson, Havighurst, & Harley, 2012). Therefore, researchers should consider assessing whether parent emotion socialization training could impact family MEC and subsequent child food consumption. To conclude, MEC is coconstructed by mothers and children (with more influence from mothers), and may impact child healthy food consumption. Children who experience more positive mealtime interactions may eat a greater amount of healthy food than children who experience mealtimes with about equal amounts of positive and negative emotion. These novel findings open up additional avenues of research, and suggest that further research on mealtime quality, associations between emotional expression and child food consumption, and emotion socialization training for parents may be needed. By helping parents engage in more positive mealtime interactions with their young children, practitioners and researchers may be able to increase children’s healthy food consumption. Funding This research was supported in part by grants from the Agriculture and Food Research Initiative of the United States Department of Agriculture (USDA) under the Illinois Transdisciplinary Obesity Prevention Program grant (2011-67001-30101) to the Division of Nutritional Sciences at the University of Illinois; the Illinois Council for Agriculture Research to Kristen Harrison (PI); the University of Illinois Health and Wellness Initiative to Barbara Fiese and Sharon Donovan; and the USDA Hatch (793-328) Program to Barbara Fiese (PI) and USDA Hatch to Kelly Bost. Conflicts of interest: None declared. References American Academy of Pediatrics. ( 2003). Prevention of pediatric overweight and obesity. Pediatrics , 112, 424– 430. Retrieved from www.aap.org CrossRef Search ADS PubMed  Anderson S. E., Keim S. A. ( 2016). Parent–child interaction, self-regulation, and obesity prevention in early childhood. Current Obesity Reports , 5, 192– 200. doi: 10.1007/s13679-016-0208-9 Google Scholar CrossRef Search ADS PubMed  Anderson S. E., Whitaker R. C. ( 2010). Household routines and obesity in US preschool-aged children. Pediatrics , 125, 420– 428. doi: 10.1542/peds.2009-0417. Google Scholar CrossRef Search ADS PubMed  Anzman S. L., Rollins B. Y., Birch L. L. ( 2010). Parental influence on children’s early eating environments and obesity risk: Implications for prevention. International Journal of Obesity , 34, 1116– 1124. doi: 10.1038/ijo.2010.43 Google Scholar CrossRef Search ADS PubMed  Bazzano L. A., He J., Ogden L. G., Loria C. M., Vupputuri S., Myers L., Whelton P. K. ( 2002). Fruit and vegetable intake and risk of cardiovascular disease in US adults: The first National Health and Nutrition Examination Survey Epidemiologic follow-up study. The American Journal of Clinical Nutrition , 76, 93– 99. Retrieved from www.ajcn.nutrition.org Google Scholar CrossRef Search ADS PubMed  Berge J. M., Rowley S., Trofholz A., Hanson C., Rueter M., MacLehose R. F., Neumark-Sztainer D. ( 2014). Childhood obesity and interpersonal dynamics during family meals. Pediatrics , 134, 923– 932. doi: 10.1542/peds.2014-1936 Google Scholar CrossRef Search ADS PubMed  Chorney J. M., McMurtry C. M., Chambers C. T., Bakeman R. ( 2015). Developing and modifying behavioral coding schemes in pediatric psychology: A practical guide. Journal of Pediatric Psychology , 40, 154– 164. doi: 10.1093/jpepsy/jsu099 Google Scholar CrossRef Search ADS PubMed  Cole P., Wiggins C., Radzioch A., Pearl A. ( 2007). D.O.T.S. Emotion coding system. Unpublished Manuscript, Department of Psychology, Pennsylvania State University. Crawford J. R., Henry J. D. ( 2003). The Depression Anxiety Stress Scales (DASS): Normative data and latent structure in a large non‐clinical sample. British Journal of Clinical Psychology , 42, 111– 131. Google Scholar CrossRef Search ADS PubMed  Demory-Luce D., Morales M., Nicklas T., Baranowski T., Zakeri I., Berenson G. ( 2004). Changes in food group consumption patterns from childhood to young adulthood: The Bogalusa Heart Study. Journal of the American Dietetic Association , 104, 1684– 1691. doi: 10.1016/j.jada.2004.07.026 Google Scholar CrossRef Search ADS PubMed  Esmaillzadeh A., Kimiagar M., Mehrabi Y., Azadbakht L., Hu F. B., Willett W. C. ( 2006). Fruit and vegetable intakes, C-reactive protein, and the metabolic syndrome. The American Journal of Clinical Nutrition , 84, 1489– 1497. Retrieved from www.ajcn.nutrition.org Google Scholar PubMed  Fiese B. H., Foley K. P., Spagnola M. ( 2006). Routine and ritual elements in family mealtimes: Contexts for child well-being and family identity. New Directions for Child and Adolescent Development , 2006, 67– 89. doi:10.1002/cd.156 Google Scholar CrossRef Search ADS   Fiese B. H., Gundersen C., Koester B., Jones B. ( 2016). Family chaos and lack of mealtime planning is associated with food insecurity in low income households. Economics and Human Biology , 21, 147– 155. doi: 10.1016/j.ehb.2016.01.004 Google Scholar CrossRef Search ADS PubMed  Fiese B. H., Hammons A., Grigsby-Toussaint D. ( 2012). Family mealtimes: A contextual approach to understanding childhood obesity. Economics and Human Biology , 10, 365– 374. doi: 10.1016/j.ehb.2012.04.004 Google Scholar CrossRef Search ADS PubMed  Fitzgibbon M., Hayman L., Haire-Joshu D. ( 2008). Childhood obesity: A policy statement of the society of behavioral medicine. Society of Behavioral Medicine. Retrieved from http://www.sbm.org FitzPatrick E., Edmunds L. S., Dennison B. A. ( 2007). Positive effects of family dinner are undone by television viewing. Journal of the American Dietetic Association , 107, 666– 671. doi: 10.1016/j.jada.2007.01.014 Google Scholar CrossRef Search ADS PubMed  Fosco G. M., Grych J. H. ( 2013). Capturing the family context of emotion regulation: A family systems model comparison approach. Journal of Family Issues , 34, 557– 578. doi: 10.1177/0192513X12445889 Google Scholar CrossRef Search ADS   Foster C. E., Webster M. C., Weissman M. M., Pilowsky D. J., Wickramaratne P. J., Rush A. J., Hughes C. W., Garber J., Malloy E., Cerda G., Kornstein S. G., Alpert J. E., Wisniewski S. R., Trivedi M. H., Fava M., King C. A. ( 2008). Course and severity of maternal depression: Associations with family functioning and child adjustment. Journal of Youth and Adolescence , 37, 906– 916. doi: 10.1007/s10964-007-9216-0 Google Scholar CrossRef Search ADS PubMed  Francis L. A., Hofer S. M., Birch L. L. ( 2001). Predictors of maternal child-feeding style: Maternal and child characteristics. Appetite , 37, 231– 243. doi: 10.1006/appe.2001.0427 Google Scholar CrossRef Search ADS PubMed  Fulkerson J. A., Larson N., Horning M., Neumark-Sztainer D. ( 2014). A review of associations between family or shared meal frequency and dietary and weight status outcomes across the lifespan. Journal of Nutrition Education and Behavior , 46, 2– 19. doi: 10.1016/j.jneb.2013.07.012 Google Scholar CrossRef Search ADS PubMed  Hammons A. J., Fiese B. H. ( 2011). Is frequency of shared family meals related to the nutritional health of children and adolescents? Pediatrics , 127, e1565– e1574. doi: 10.1542/peds.2010-1440 Google Scholar CrossRef Search ADS PubMed  Harrison K., Liechty J. M.; The STRONG Kids Program. ( 2012). US preschoolers’ media exposure and dietary habits: The primacy of television and the limits of parental mediation. Journal of Children and Media , 6, 18– 36. doi: 10.1080/17482798.2011.633402 Google Scholar CrossRef Search ADS   Havighurst S. S., Wilson K. R., Harley A. E., Prior M. R., Kehoe C. ( 2010). Tuning in to Kids: Improving emotion socialization practices in parents of preschool children—findings from a community trial. Journal of Child Psychology and Psychiatry , 51, 1342– 1350. doi: 10.1111/j.1469-7610.2010.02303.x Google Scholar CrossRef Search ADS PubMed  Haycraft E., Farrow C., Blissett J. ( 2013). Maternal symptoms of depression are related to observations of controlling feeding practices in mothers of young children. Journal of Family Psychology , 27, 159– 64. doi: 10.1037/a0031110 Google Scholar CrossRef Search ADS PubMed  Henry J. D., Crawford J. R. ( 2005). The short-form version of the Depression Anxiety Stress Scales (DASS-21): Construct validity and normative data in a large non-clinical sample. British Journal of Clinical Psychology , 44, 227– 239. doi: 10.1348/014466505X29657 Google Scholar CrossRef Search ADS PubMed  Hughes S. O., Power T. G., Papaioannou M. A., Cross M. B., Nicklas T. A., Hall S. K., Shewchuk R. M. ( 2011). Emotional climate, feeding practices, and feeding styles: An observational analysis of the dinner meal in Head Start families. International Journal of Behavioral Nutrition and Physical Activity , 8, 60. doi:10.1186/1479-5868-8-60 Google Scholar CrossRef Search ADS PubMed  IBM Analytics. ( 2016, March 17). Retrieved from http://www.ibm.com/analytics/us/en/technology/spss/ Retrieved 18 April 2016. Janicke D. M., Mitchell M. J., Quittner A. L., Piazza-Waggoner C., Stark L. J. ( 2008). The impact of behavioral intervention on family interactions at mealtime in pediatric cystic fibrosis. Children's Health Care , 37, 49– 66. doi: 10.1080/02739610701766891 Google Scholar CrossRef Search ADS   Jarman M., Lawrence W., Ntani G., Tinati T., Pease A., Black C., Baird J., Barker M.; SIH Study Group. ( 2012). Low levels of food involvement and negative affect reduce the quality of diet in women of lower educational attainment. Journal of Human Nutrition and Dietetics , 25, 444– 452. doi: 10.1111/j.1365-277X.2012.01250.x Google Scholar CrossRef Search ADS PubMed  Jones B. L., Fiese B. H.; STRONG Kids Team. ( 2014). Parent routines, child routines, and family demographics associated with obesity in parents and preschool-aged children. Frontiers in Psychology , 5, 374. doi: 10.3389/fpsyg.2014.00374 Google Scholar PubMed  Kotila L. E., Schoppe-Sullivan S. J., Dush C. M. ( 2013). Time in parenting activities in dual‐earner families at the transition to parenthood. Family Relations , 62, 795– 807. doi: 10.1111/fare.12037 Google Scholar CrossRef Search ADS PubMed  Krishnamoorthy J. S., Hart C., Jelalian E. ( 2006). The epidemic of childhood obesity: Review of research and implications for public policy  (Vol. 20). Ann Arbor, MI: Society for Research in Child Development. Retrieved from http://eric.ed.gov Kuczmarski R. J., Ogden C. L., Grummer-Strawn L. M., Flegal K. M., Guo S. S., Wei R., Mei Z., Curtin L. R., Roche A. F., Johnson C. L. ( 2000). CDC growth charts: United States. Advance Data , 314, 1– 27. Retrieved from http://www.cdc.gov Leech R. M., McNaughton S. A., Crawford D. A., Campbell K. J., Pearson N., Timperio A. ( 2014). Family food involvement and frequency of family dinner meals among Australian children aged 10–12 years. Cross-sectional and longitudinal associations with dietary patterns. Appetite , 75, 64– 70. doi: 10.1016/j.appet.2013.12.021 Google Scholar CrossRef Search ADS PubMed  Lien N., Lytle L. A., Klepp K.-I. ( 2001). Stability in consumption of fruit, vegetables, and sugary foods in a cohort from age 14 to age 21. Preventive Medicine , 33, 217– 226. doi: 10.1006/pmed.2001.0874 Google Scholar CrossRef Search ADS PubMed  Lytle L. A., Seifert S., Greenstein J., McGovern P. ( 2000). How do children’s eating patterns and food choices change over time? Results from a cohort study. American Journal of Health Promotion , 14, 222– 228. doi: 10.4278/0890-1171-14.4.222 Google Scholar CrossRef Search ADS PubMed  Malhotra K., Herman A. N., Wright G., Bruton Y., Fisher J. O., Whitaker R. C. ( 2013). Perceived benefits and challenges for low-income mothers of having family meals with preschool-aged children: Childhood memories matter. Journal of the Academy of Nutrition and Dietetics , 113, 1484– 1493. doi: 10.1016/j.jand.2013.07.028 Google Scholar CrossRef Search ADS PubMed  McCurdy K., Gorman K. S., Kisler T., Metallinos-Katsaras E. ( 2014). Associations between family food behaviors, maternal depression, and child weight among low-income children. Appetite , 79, 97– 105. doi: 10.1016/j.appet.2014.04.015 Google Scholar CrossRef Search ADS PubMed  McPhie S., Skouteris H., Daniels L., Jansen E. ( 2014). Maternal correlates of maternal child feeding practices: A systematic review. Maternal and Child Nutrition , 10, 18– 43. doi: 10.1111/j.1740-8709.2012.00452.x Google Scholar CrossRef Search ADS PubMed  Musher-Eizenman D., Holub S. ( 2007). Comprehensive feeding practices questionnaire: Validation of a new measure of parental feeding practices. Journal of Pediatric Psychology , 32, 960– 972. doi: 10.1093/jpepsy/jsm037 Google Scholar CrossRef Search ADS PubMed  National Center for Educational Statistics. ( 2007). Early Childhood Longitudinal Program (ECLS)—Overview. Retrieved from https://nces.ed.gov/ Ohly H., Pealing J., Hayter A. K., Pettinger C., Pikhart H., Watt R. G., Rees G. ( 2013). Parental food involvement predicts parent and child intakes of fruits and vegetables. Appetite , 69, 8– 14. doi: 10.1016/j.appet.2013.05.003 Google Scholar CrossRef Search ADS PubMed  Rollins B. Y., Belue R. Z., Francis L. A. ( 2010). The beneficial effect of family meals on obesity differs by race, sex, and household education: The National Survey of Children’s Health, 2003-2004. Journal of the American Dietetic Association , 110, 1335– 1339. doi: 10.1016/j.jada.2010.06.004 Google Scholar CrossRef Search ADS PubMed  Stark L. J., Spear S., Boles R., Kuhl E., Ratcliff M., Scharf C., Bolling C., Rausch J. ( 2011). A pilot randomized controlled trial of a clinic and home‐based behavioral intervention to decrease obesity in preschoolers. Obesity , 19, 134– 141. doi: 10.1038/oby.2010.87 Google Scholar CrossRef Search ADS PubMed  Steinley D. ( 2006). K-means clustering: A half-century synthesis. British Journal of Mathematical and Statistical Psychology , 59, 1– 34. doi: 10.1348/000711005X48266 Google Scholar CrossRef Search ADS PubMed  Wang X., Ouyang Y., Liu J., Zhu M., Zhao G., Bao W., Hu F. B. ( 2014). Fruit and vegetable consumption and mortality from all causes, cardiovascular disease, and cancer: Systematic review and dose-response meta-analysis of prospective cohort studies. BMJ , 349, g4490. doi: 10.1136/bmj.g4490 Google Scholar CrossRef Search ADS PubMed  Webber L., Cooke L., Hill C., Wardle J. ( 2010). Associations between children’s appetitive traits and maternal feeding practices. Journal of the American Dietetic Association , 110, 1718– 1722. doi: 10.1016/j.jada.2010.08.007 Google Scholar CrossRef Search ADS PubMed  Wilson K. R., Havighurst S. S., Harley A. E. ( 2012). Tuning in to kids: An effectiveness trial of a parenting program targeting emotion socialization of preschoolers. Journal of Family Psychology , 26, 56– 65. doi: 10.1037/a0026480 Google Scholar CrossRef Search ADS PubMed  Yoder P., Symons F. ( 2010). Observational measurement of behavior . New York, NY: Springer Publishing Company. © The Author 2017. Published by Oxford University Press on behalf of the Society of Pediatric Psychology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Pediatric Psychology Oxford University Press

Predictors and Outcomes of Mealtime Emotional Climate in Families With Preschoolers

Loading next page...
 
/lp/ou_press/predictors-and-outcomes-of-mealtime-emotional-climate-in-families-with-yyYAI6oIQa
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of the Society of Pediatric Psychology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
0146-8693
eISSN
1465-735X
D.O.I.
10.1093/jpepsy/jsx109
Publisher site
See Article on Publisher Site

Abstract

Abstract Objective Mealtime emotional climate (MEC) is related to parent feeding and mental health, and possibly to child food consumption. However, MEC has been inconsistently assessed with a variety of coding schemes and self-report instruments, and has not been examined longitudinally. This study aims to characterize MEC systematically using an observational, count-based coding scheme; identify whether parent feeding or mental health predict MEC; and examine whether MEC predicts child food consumption and weight. Methods A subsample of parents (n = 74) recruited from a larger study completed questionnaires when children were about 37 months, participated in a home visit to videotape a mealtime when children were about 41 months, and completed questionnaires again when children were about 51 months old. Maternal and child positive and negative emotions were coded from videotaped mealtimes. Observational data were submitted to cluster analyses, to identify dyads with similar emotion expression patterns, or MEC. Logistic regression was used to identify predictors of MEC, and Analysis of Covariance was used to examine differences between MEC groups. Results Dyads were characterized as either Positive Expressers (high positive, low negative emotion) or All Expressers (similar positive and negative emotion). Increased food involvement feeding practices were related to decreased likelihood of being an All Expresser. Positive Expressers reported that their children ate more healthy food, compared with All Expressers. Conclusions Observed MEC is driven by maternal emotion, and may predict child food consumption. Food involvement may promote positive MEC. Improving MEC may increase child consumption of healthy foods. body weight, child, emotions, family, longitudinal studies, meals, observational methodology, parents Introduction Increased consumption of healthy foods—such as fruits, vegetables, and lean protein sources—is linked to lower risk for chronic disease (Bazzano et al., 2002; Esmaillzadeh et al., 2006; Wang et al., 2014). Healthy food consumption declines from childhood to young adulthood (Demory-Luce et al., 2004) but is relatively stable from adolescence to adulthood, pointing to the importance of early habit formation and maintenance (Lien, Lytle, & Klepp, 2001). Parents control and shape the home food environment, where children spend the majority of their time during this critical developmental phase (American Academy of Pediatrics, 2003; Anzman, Rollins, & Birch, 2010; Fitzgibbon, Hayman, & Haire-Joshu, 2008; Krishnamoorthy, Hart, & Jelalian, 2006). Family mealtimes provide a window of opportunity to observe whether certain behaviors or interactions—from both parents and children—are linked to food consumption and weight outcomes (Fiese, Foley, & Spagnola, 2006). Increased family mealtime frequency is consistently related to healthy food consumption and lower risk for overweight in young children (Anderson & Whitaker, 2010; Fulkerson, Larson, Horning, & Neumark-Sztainer, 2014; Hammons & Fiese, 2011). Mealtime quality also predicts child healthy food consumption (Fiese, Hammons, & Grigsby-Toussaint, 2012). Interpersonal interactions, maternal emotional expression, mental health, and marital relationship quality have been examined previously as indicators of mealtime quality (Berge et al., 2014; Hughes et al., 2011; Lytle, Seifert, Greenstein, & McGovern, 2000; Rollins, Belue, & Francis, 2010). For example, increased maternal negative emotion expression (measured by assessing tone, body language, and facial expression) is linked to indulgent feeding styles and higher child weight (Hughes et al., 2011). Furthermore, maternal depression is related to observed controlling feeding practices (Haycraft, Farrow, & Blissett, 2013), and to self-report of fewer positive mealtime practices, such as being present while the child eats, child control of snacking, and food resource management (McCurdy, Gorman, Kisler, & Metallinos-Katsaras, 2014). Observational studies suggest that the mealtime emotional climate (MEC)—positive and negative emotion expression frequency—is linked to feeding and weight (Berge et al., 2014; Fosco & Grych, 2013; Hughes et al., 2011). For instance, families with more positive interpersonal dynamics (more warmth, higher relationship quality, and group enjoyment) have children with lower weight, as compared with families with more negative interpersonal dynamics (more hostility, food lecturing/moralizing, and indulgence; Berge et al., 2014). Furthermore, studies with clinical samples have found success using behavioral interventions focused on feeding to improve mealtime affective management (Janicke, Mitchell, Quittner, Piazza-Waggoner, & Stark, 2008). Collectively, findings suggest that emotional characteristics of mealtimes are modifiable, and may increase healthy food consumption or improve weight outcomes. However, measures of MEC vary, limiting our ability to compare across studies. Global observational schemes have been primarily used to examine MEC as a function of maternal (but not child) emotion expression (Hughes et al., 2011), or family- and dyad-relationship quality (Berge et al., 2014). Although global codes garner macro-level descriptions of behavior, it is difficult to measure changing dynamics using a global rating-based coding scheme (Chorney, McMurtry, Chambers, & Bakeman, 2015; Yoder & Symons, 2010). Self-report measures or global coding schemes may yield foundational information about emotional climate, but these measures do not always depict a thorough representation of emotional expression. For instance, it is logical to expect that emotional climate depends—in part—on the proportion of positive and negative emotions expressed by family members during the meal. Therefore, an observational count coding system may provide a better alternative. Nevertheless, to the best of our knowledge, no previous studies have used this approach to measure MEC. Additionally, mealtimes may be especially challenging for parents who use unhealthy feeding practices or with mental health problems (Hughes et al., 2011; McPhie, Skouteris, Daniels, & Jansen, 2014). Therefore, it is important to examine how parenting may influence MEC. Finally, all previous studies examining MEC have been cross-sectional; thus, longitudinal studies are needed. The current study has three aims designed to address these limitations and to contribute to the body of knowledge regarding promotion of health and health behaviors in young children. The first aim of the present study is to characterize MEC using an observational count coding scheme, by examining the proportional frequency of maternal and child emotional expression during home-based family mealtimes. Given the focus in the literature on maternal emotional expression as an indicator of mealtime climate, we will examine whether child emotional expression also contributes to MEC. Second, we will examine whether parent feeding or mental health symptoms predict MEC. Finally, we will examine whether MEC predicts child food consumption or weight longitudinally. This work will contribute to our current understanding of the socioemotional context of eating behavior, by examining how observed emotional expression during home-based mealtimes is related to preschool-aged children’s eating behaviors. Methods Participants Participants for this study were drawn from the larger Synergistic Theory and Research on Obesity and Nutrition Group (STRONG) Kids Panel Study. The larger STRONG Kids Study is a three-wave prospective panel survey designed to examine transdisciplinary predictors of child health behaviors and weight. Primary caregivers of preschoolers were recruited from childcare centers in East Central Illinois that were registered with the state Bureau of Child Care and Development, and enrolled at least 24 children (Harrison, Liechty, & The STRONG Kids Program, 2012). Children were about 37 months at Wave 1 (W1), 51 months at Wave 2 (W2), and 70 months at Wave 3 (W3). Parents completed self-report questionnaires, and children’s height and weight were measured by trained research assistants at each wave. Response rate varied by preschool, ranging from 60 to 95%. This study reports on a subsample (n = 74) of families in the STRONG Kids study that consented to participate in a 2-hr home visit when children were about 41 months old (SD = 5.23), after submitting surveys in W1. Response rate for the subsample was 52% of the eligible families that were contacted (n = 143). Home visit recruitment targeted families who had already completed surveys and measurements at W1 (Table I). Parents were offered $50 for remuneration for the home visit. This study was approved by the university institutional review board. Sample characteristics are reported in Table I. Table I. Sample Characteristics (n = 74) Demographic characteristics  n  %    n (M)  % (SD)  Range  Marital status      Employment status         Single  9  12   Employed  13  18     Married  59  80   Self-employed  32  43     Separated/divorced  3  4   Stay-at-home parent  3  4     Civil union  1  1   Student  3  4     Cohabiting  1  1   Retired  9  12    Race/ethnicitya       Out of work  11  15     Black or African American  7  10  Education         White  56  76   High school or less  3  4     Asian  3  4   Some college  16  22     Hispanic/Latino  8  11   College graduate  19  26     Mixed ethnicity  5  7   Postgraduate work  36  47    Household income      Age         $24,999 or less  9  12   W1 parent age  33.6  5.6  27.9   $25,000–$39,999  10  14   W1 child age (months)  37.2  6.2  29.0   $40,000–$69,999  11  15   W2 child age  51.7  7.3  39.2   $70,000–$99,999  22  30   W3 child age  70.3  7.2  34.0   $100,000 or more  19  26   Home visit child age  41  5.2  19.0  Child gender      Parent gender         Male  37  50   Male  7  9     Female  37  50   Female  67  91    Demographic characteristics  n  %    n (M)  % (SD)  Range  Marital status      Employment status         Single  9  12   Employed  13  18     Married  59  80   Self-employed  32  43     Separated/divorced  3  4   Stay-at-home parent  3  4     Civil union  1  1   Student  3  4     Cohabiting  1  1   Retired  9  12    Race/ethnicitya       Out of work  11  15     Black or African American  7  10  Education         White  56  76   High school or less  3  4     Asian  3  4   Some college  16  22     Hispanic/Latino  8  11   College graduate  19  26     Mixed ethnicity  5  7   Postgraduate work  36  47    Household income      Age         $24,999 or less  9  12   W1 parent age  33.6  5.6  27.9   $25,000–$39,999  10  14   W1 child age (months)  37.2  6.2  29.0   $40,000–$69,999  11  15   W2 child age  51.7  7.3  39.2   $70,000–$99,999  22  30   W3 child age  70.3  7.2  34.0   $100,000 or more  19  26   Home visit child age  41  5.2  19.0  Child gender      Parent gender         Male  37  50   Male  7  9     Female  37  50   Female  67  91    Note. W1 = Wave 1; W2 = Wave 2; W3 = Wave 3. a Race/ethnicity variables were not mutually exclusive, and a small number of participants indicated multiple ethnicities. Home Visit Procedures Parents provided informed consent for mealtimes to be video-recorded before and during home visits. Researchers stayed in the home before the mealtime to build rapport. Researchers left the home during the mealtime, and returned after the mealtime was finished, to prevent social desirability and observer biases from influencing behavior. Videotapes of mealtimes were immediately deidentified and uploaded onto a secure server. Observational Coding Procedures Training and Reliability Two research assistants were trained on the D.O.T.S. emotion coding system (Cole, Wiggins, Radzioch, & Pearl, 2007) until attaining adequate inter-rater reliability (Intra-class correlation ≥ .70) on the maternal and child positive and negative affect count codes. During this process, decision rules were established and disagreements were discussed. Observed agreement for 20% of the cases that were double coded after training were Intra-class correlation = .73 to .90. Coding Schemes The frequency of positive and negative affect from mothers and children was coded using the D.O.T.S emotion coding system (Cole et al., 2007). Videos were watched in separate sessions to code maternal and child affect. Behaviors, facial expressions, and vocalizations were coded. Examples of positive emotion expressions for mothers included smiling, speaking in enthusiastic tones, laughing, and expressing physical affection; whereas sighing, furrowing brows, or yelling were considered negative emotion expressions. Examples of positive emotion expressions for children included giggling, sustained smiling, and bouncing; and whining, quivering lips, or crying were considered negative emotion expressions. Mealtime Length Meal start time was indicated by food being placed on the table or—if the family did not eat at the table—by presence of food in the eating environment. Meal end time was indicated by the last time food is taken away from the target child, or the last time the target child leaves the table. Mealtime length is calculated according to meal start and end time. Measures Child Weight Child height and weight were measured at each time point by trained research assistants using a digital scale and stadiometer (Kuczmarski et al., 2000). Sex- and age-adjusted body mass index percentile (BMI-P) was calculated from W1, W2, and W3 data. Food Consumption Child food consumption was assessed using six items from the Early Childhood Longitudinal Study, Birth Cohort (ECLS-B; National Center for Educational Statistics, 2007). Parents reported how often their children ate fruits, vegetables, sugar-sweetened beverages, fruit juice, fast food, soy products, French fries, candy/sweets, and salty snacks over the past 7 days, accounting for meals and snacks eaten at home, at school, or in any other situation. Response options ranged from 0 to 7, with 0 as “my child did not eat/drink any _____ during the past 7 days,” 1 as “my child ate/drank ____ once a day,” 2 as “twice a day,” 3 as “three times a day,” 4 as “four or more times a day,” and 7 as “I don’t know,” which was treated as missing data. Responses that indicated 5 or 6 were recoded into fractions to represent consumption by times per day (5 was “1 to 3 times during the past 7 days” = 0.2857 times daily; 6 was “4 to 6 times during the past 7 days” = 0.7143 times daily). Food consumption variables were constructed based on a two-factor solution identified by a principal component analysis (PCA) with varimax rotation that included each of these items from the Wave 2 questionnaire (42% of variation explained). The rotated component matrix suggested that fresh fruit, vegetables, and soy-based foods comprised one component, whereas fruit juice, sugar sweetened beverages, fast food, sweets, French fries, and salty snacks comprised another component. All items loaded ≥0.3 on the identified components. Composite scores were constructed based on the findings of the PCA. A composite score for daily healthy food consumption was made by summing responses to items about fruit, vegetable, and soy consumption. A composite score for daily unhealthy food consumption was made by summing responses to items about sugar-sweetened beverage, fruit juice, fast food, French fries, candy/sweet, and salty snack consumption (Table II). Healthy food consumption was correlated at W1 and W2 (r = .50, p < .001), and unhealthy food consumption was correlated at W1 and W2 (r = .49, p < .001). Healthy and unhealthy food consumption was not correlated at either W1 or W2. Table II. Model Variable Statistics by MEC Group Membership (n = 74) Variable  Positive Expressers (n = 59)   All Expressers (n = 15)   Full Sample (n = 74)   Kruskal–Wallis H testsa   M  SD  M  SD  M  SD  X2  p  Observed mother emotion                   Positive emotion ratio  0.92  0.10  0.41  0.22  0.82  0.24  36.16  <0.01   Negative emotion ratio  0.08  0.10  0.53  0.24  0.17  0.23  28.97  <0.01   Total emotions  18.97  9.50  12.40  8.68  17.64  9.66  7.80  0.01  Observed child emotion                   Positive emotion ratio  0.73  0.21  0.67  0.13  0.72  0.20  2.40  0.12   Negative emotion ratio  0.27  0.21  0.33  0.13  0.28  0.20  2.40  0.12   Total emotions  19.57  10.91  19.87  9.32  19.62  10.55  0.00  0.98  W1 parent BMI  25.75  5.11  26.65  6.01  25.93  5.27  0.12  0.73  Child BMI-P                   W1 child BMI-P  59.10  21.22  65.36  22.81  60.37  21.55  1.09  0.30   W2 child BMI-P  57.63  24.57  67.86  19.09  59.70  23.80  2.01  0.16   W3 child BMI-P  59.68  25.67  70.55  15.54  61.89  24.28  1.17  0.28  Food consumptionb                   W1 healthy food  3.80  1.31  3.09  1.29  3.66  1.33  3.03  0.08   W1 unhealthy food  3.02  1.43  3.38  1.32  3.09  1.41  1.32  0.25   W2 healthy food  3.60  1.50  2.66  1.12  3.41  1.47  5.67  0.02   W2 unhealthy food  2.68  1.49  2.67  0.79  2.68  1.37  0.21  0.65  Feeding practices (W1)                   Food Involvement  3.10  0.85  2.44  0.92  2.96  0.90  6.54  0.01   Restriction for Health  2.86  0.81  3.11  1.01  2.91  0.85  1.05  0.31   Restriction for Weight  1.62  0.54  1.66  0.45  1.63  0.52  0.44  0.51   Environment  3.97  0.61  3.67  0.63  3.91  0.62  2.20  0.14   Pressure to Eat  2.52  0.75  2.56  0.74  2.53  0.74  0.02  0.91   Monitoring  4.24  0.82  3.95  0.90  4.18  0.84  1.41  0.24   Modeling  3.86  0.75  3.59  0.81  3.81  0.77  1.84  0.18   Balance/Variety  4.41  0.56  4.31  0.75  4.39  0.60  0.06  0.81  Parent mental health (W1)                   Depressive symptoms  1.20  0.21  1.54  0.70  1.27  0.39  3.24  0.07   Stress symptoms  1.69  0.57  1.84  0.77  1.72  0.61  0.18  0.67  Total family members  3.59  0.77  4.13  0.64  3.70  0.08  5.79  0.02  Mealtime length  24.17  8.40  22.86  8.67  23.90  8.41  0.37  0.54  Variable  Positive Expressers (n = 59)   All Expressers (n = 15)   Full Sample (n = 74)   Kruskal–Wallis H testsa   M  SD  M  SD  M  SD  X2  p  Observed mother emotion                   Positive emotion ratio  0.92  0.10  0.41  0.22  0.82  0.24  36.16  <0.01   Negative emotion ratio  0.08  0.10  0.53  0.24  0.17  0.23  28.97  <0.01   Total emotions  18.97  9.50  12.40  8.68  17.64  9.66  7.80  0.01  Observed child emotion                   Positive emotion ratio  0.73  0.21  0.67  0.13  0.72  0.20  2.40  0.12   Negative emotion ratio  0.27  0.21  0.33  0.13  0.28  0.20  2.40  0.12   Total emotions  19.57  10.91  19.87  9.32  19.62  10.55  0.00  0.98  W1 parent BMI  25.75  5.11  26.65  6.01  25.93  5.27  0.12  0.73  Child BMI-P                   W1 child BMI-P  59.10  21.22  65.36  22.81  60.37  21.55  1.09  0.30   W2 child BMI-P  57.63  24.57  67.86  19.09  59.70  23.80  2.01  0.16   W3 child BMI-P  59.68  25.67  70.55  15.54  61.89  24.28  1.17  0.28  Food consumptionb                   W1 healthy food  3.80  1.31  3.09  1.29  3.66  1.33  3.03  0.08   W1 unhealthy food  3.02  1.43  3.38  1.32  3.09  1.41  1.32  0.25   W2 healthy food  3.60  1.50  2.66  1.12  3.41  1.47  5.67  0.02   W2 unhealthy food  2.68  1.49  2.67  0.79  2.68  1.37  0.21  0.65  Feeding practices (W1)                   Food Involvement  3.10  0.85  2.44  0.92  2.96  0.90  6.54  0.01   Restriction for Health  2.86  0.81  3.11  1.01  2.91  0.85  1.05  0.31   Restriction for Weight  1.62  0.54  1.66  0.45  1.63  0.52  0.44  0.51   Environment  3.97  0.61  3.67  0.63  3.91  0.62  2.20  0.14   Pressure to Eat  2.52  0.75  2.56  0.74  2.53  0.74  0.02  0.91   Monitoring  4.24  0.82  3.95  0.90  4.18  0.84  1.41  0.24   Modeling  3.86  0.75  3.59  0.81  3.81  0.77  1.84  0.18   Balance/Variety  4.41  0.56  4.31  0.75  4.39  0.60  0.06  0.81  Parent mental health (W1)                   Depressive symptoms  1.20  0.21  1.54  0.70  1.27  0.39  3.24  0.07   Stress symptoms  1.69  0.57  1.84  0.77  1.72  0.61  0.18  0.67  Total family members  3.59  0.77  4.13  0.64  3.70  0.08  5.79  0.02  Mealtime length  24.17  8.40  22.86  8.67  23.90  8.41  0.37  0.54  Note. BMI = body mass index; BMI-P = body mass index percentile; MEC = mealtime emotional climate; PCA = principal components analysis; W1 = Wave 1; W2 = Wave 2; W3 = Wave 3. a Kruskall–Wallis H tests examine mean differences between MEC groups on model variables. Bolded values indicate significant differences at p < .05. b Food consumption variables were constructed based on a two-factor solution identified by a PCA with varimax rotation on Wave 2 items (42% of variation explained). Rotated component matrix suggested that fresh fruit, vegetables, and soy-based foods comprised one factor, whereas fruit juice, sugary beverages, fast food, sweets, French fries, and salty snacks comprised another. Summed composite variables of healthy and unhealthy food consumption were constructed based on PCA findings. Feeding Practices Parent feeding practices were assessed at W1 and W2 using the Comprehensive Feeding Practices Questionnaire (CFPQ; Musher-Eizenman & Holub, 2007). The CFPQ consists of 49 items that correspond to 12 subscales: Child Control (α = 0.34), Emotion Regulation (α = 0.64), Balance/Variety (α = 0.72), Environment (α = 0.70), Food as Reward (α = 0.67), Child Food Involvement (α = 0.77), Modeling (α = 0.80), Monitoring (α = 0.87), Pressure to Eat (α = 0.71), Restriction for Health (α = 0.73), Restriction for Weight Control (α = 0.83), and Teaching about Nutrition (α = 0.51). Parents are asked to indicate how often they engaged in a particular feeding practice on a Likert scale from 1 (never) to 5 (always). Although subscale internal reliability was comparable with previously reported data (Musher-Eizenman & Holub, 2007), we elected to remove subscales with α < 0.70 (Food as Reward, Emotion Regulation, Child Control, and Teaching about Nutrition) to reduce measurement error. Maternal Mental Health Maternal mental health symptoms were assessed at W1 and W2 by the Depression, Anxiety, and Stress Scale (DASS-21; Henry & Crawford, 2005). The DASS consists of 21 items in a Likert-SCALE format. Depressive and Stress symptom subscales (seven items each) had adequate reliability in this sample (α = 0.84 and 0.87, respectively), but the Anxiety subscale did not (α = 0.62), and so was not included in analyses to reduce measurement error. Scores on the Depressive symptoms subscale ranged from 1.00 to 3.57 at W1, and 1.00 to 2.43 at W2. Scores on the Stress symptoms subscale ranged from 1.00 to 4.00 at W1, and 1.00 to 3.29 at W2. Covariates Demographic characteristics—including race/ethnicity, income, education, employment status, gender, age, and number of family members in the household—were self-reported by parents at W1 and W2 and included as controls in relevant analyses. Parent body mass index (BMI) was calculated from self-reported height and weight, and was included as a control in all analyses. Mealtime length was calculated from videotaped observations, and included as a control. Analysis Strategy Missing Data Missingness ranged from 0 to 16%. However, about 24% (n = 18) of W3 child BMI-P data were missing. We examined differences between cases missing and not missing W3 child BMI data. Salty snack consumption at W2 was associated with greater likelihood of missing W3 child BMI, but data were otherwise not missing systematically (Little’s Missing Completely At Random [MCAR] test X2 [df] = 1225.62 [8279], p = 1.00). Therefore, we applied multiple imputation with expectation maximization algorithms >20 iterations. Demographic variables were modeled as predictors only, and all other variables were modeled as both predictors and imputable outcomes. Imputed data sets were then aggregated. Analyses were also run on raw data with missing variables, with no significant differences between findings; therefore, we present results from imputed data to leverage the full sample. Additionally, given that observational coding focused on maternal–child interactions, all analyses were also conducted on a sample excluding the seven male primary caregivers who completed surveys at W1. All analyses were run in SPSS 24.0 (IBM Analytics, 2016) and probability levels were set at .05. Aim 1 To characterize MEC, we first created ratio variables by dividing the number of positive and negative emotions by the total number of emotions expressed by an individual, yielding four variables describing the proportion of positive and negative emotions expressed by mothers and children. Ratio variables were submitted to a Ward’s hierarchical cluster analysis using a squared Euclidean distance clustering method. After analyzing the agglomeration schedule and the dendrogram, a two-cluster solution was deemed the best fit for these data. To confirm this solution, we performed an additional K-means cluster analysis, specifying a two-cluster solution. The solution converged in 12 iterations, and the clusters identified by Ward’s method were not different from the clusters identified by the K-means method. Because K-means methods are sensitive to outliers (Steinley, 2006), cluster placement identified by Ward’s method was used in final analyses. Table II describes cluster characteristics. Aim 2 To examine predictors of MEC, we examined the effects of parent feeding practices and mental health on MEC group membership first using Kruskal–Wallis H tests (Table II), and then in binary logistic regression (Table III). Table III. Effects of Demographics, Depressive Symptoms, and Feeding Practices on Observed MEC Cluster Membershipa Predictors  Model 1   Model 2   β  SE  OR  95% CI   β  SE  OR  95% CI   LLCI  ULCI  LLCI  ULCI  Block 1: Demographic variables                       Constant  −5.17  3.05  0.01      −2.64  3.80  0.07       Child genderb  −0.43  0.65  0.65  0.18  2.32  −0.74  0.75  0.48  0.11  2.07   Parent genderb  0.41  1.22  1.51  0.14  16.33  0.20  1.36  1.22  0.09  17.40   Child age (months)  −0.01  0.05  0.99  0.89  1.11  0.01  0.06  1.01  0.89  1.14   Income  −0.24  0.24  0.78  0.49  1.26  −0.31  0.28  0.73  0.43  1.26   Total family members  1.40  0.60  4.03  1.25  13.03  0.83  0.61  2.31  0.70  7.60   Mealtime length  −0.03  0.04  0.97  0.90  1.06  −0.03  0.05  0.98  0.89  1.07  Block 2: Model variables                       W1 DASS: Depressive symptoms            1.90  0.98  6.65  0.98  45.10   W1 CFPQ: Food Involvement            −1.06  0.47  0.35  0.14  0.87  −2 log likelihood  64.51  53.49  Model X2  X2 = 8.71, df = 6, p = .19  X2 = 19.73, df = 8, p = .011  Nagelkerke R2  17.9%  37.7%  Hosmer and Lemeshow test  X2 = 3.64, df = 8, p = .88  X2 = 11.18, df = 8, p = .19  Classification accuracy  81.7%  88.7%  Predictors  Model 1   Model 2   β  SE  OR  95% CI   β  SE  OR  95% CI   LLCI  ULCI  LLCI  ULCI  Block 1: Demographic variables                       Constant  −5.17  3.05  0.01      −2.64  3.80  0.07       Child genderb  −0.43  0.65  0.65  0.18  2.32  −0.74  0.75  0.48  0.11  2.07   Parent genderb  0.41  1.22  1.51  0.14  16.33  0.20  1.36  1.22  0.09  17.40   Child age (months)  −0.01  0.05  0.99  0.89  1.11  0.01  0.06  1.01  0.89  1.14   Income  −0.24  0.24  0.78  0.49  1.26  −0.31  0.28  0.73  0.43  1.26   Total family members  1.40  0.60  4.03  1.25  13.03  0.83  0.61  2.31  0.70  7.60   Mealtime length  −0.03  0.04  0.97  0.90  1.06  −0.03  0.05  0.98  0.89  1.07  Block 2: Model variables                       W1 DASS: Depressive symptoms            1.90  0.98  6.65  0.98  45.10   W1 CFPQ: Food Involvement            −1.06  0.47  0.35  0.14  0.87  −2 log likelihood  64.51  53.49  Model X2  X2 = 8.71, df = 6, p = .19  X2 = 19.73, df = 8, p = .011  Nagelkerke R2  17.9%  37.7%  Hosmer and Lemeshow test  X2 = 3.64, df = 8, p = .88  X2 = 11.18, df = 8, p = .19  Classification accuracy  81.7%  88.7%  Note. CFPQ = Comprehensive Feeding Practices Questionnaire; CI = confidence interval; DASS = Depression, Anxiety, and Stress Scale; LLCI = lower-level confidence interval; MEC = mealtime emotional climate; OR = odds ratio; ULCI = upper-level confidence interval. Bolded lines indicate statistically significant findings. a Positive Expressers coded as 0, All Expressers coded as 1. Higher OR indicates greater likelihood of being an All Expresser. b For both parent and child gender, reference category was male. Aim 3 To examine whether MEC influences child food consumption at W2 or weight outcomes at W2 and W3. Preliminary analyses of mean differences were conducted with Kruskal–Wallis H tests, and post hoc analyses of covariance (ANCOVA) were used to differences in mean food consumption across MEC groups (Table IV). Covariates included child gender, parent gender, income, total number of family members, mealtime length, and change in the outcome variable from W1 to W2. Analyses met error variance equality assumptions. Table IV. Effects of MEC on Child Healthy Food Consumption, Controlling for Demographics and Change in Healthy Food Consumption From Wave 1 (W1) to Wave 2 (W2) Predictors  β  SE  η2  95% CI unstandardized β   Estimated marginal means (SE)a   LLCI  ULCI  Positive Expressers  All Expressers  Covariates                 Parent gender  0.13  0.48  0.03  −0.32  1.64       Child gender  −0.15  0.27  0.04  −0.99  0.09       Child age  −0.13  0.02  0.02  −0.07  0.02       Total family members  0.01  0.21  0.03  −0.16  0.69       Income  −0.18  0.11  0.05  −0.42  0.03       Meal length  0.17  0.02  0.04  −0.01  0.06       Change in healthy food consumption (W1–W2)  0.60  0.10  0.37  0.42  0.83      MEC  0.28  0.35  0.12  0.33  1.70  3.65 (0.15)  2.64 (0.30)  Predictors  β  SE  η2  95% CI unstandardized β   Estimated marginal means (SE)a   LLCI  ULCI  Positive Expressers  All Expressers  Covariates                 Parent gender  0.13  0.48  0.03  −0.32  1.64       Child gender  −0.15  0.27  0.04  −0.99  0.09       Child age  −0.13  0.02  0.02  −0.07  0.02       Total family members  0.01  0.21  0.03  −0.16  0.69       Income  −0.18  0.11  0.05  −0.42  0.03       Meal length  0.17  0.02  0.04  −0.01  0.06       Change in healthy food consumption (W1–W2)  0.60  0.10  0.37  0.42  0.83      MEC  0.28  0.35  0.12  0.33  1.70  3.65 (0.15)  2.64 (0.30)  Note. CI = confidence interval; LLCI = lower-level confidence interval; MEC = mealtime emotional climate; ULCI = upper-level confidence interval. Model R2 = .48, adjusted R2 = .41, F(1, 62) = 8.67, p < .01. Pairwise comparison mean difference between All Expressers and Positive Expressers on W2 healthy food consumption (mean difference = −1.01, SE = .35, p < .01). Bolded lines indicate statistically significant findings. a Estimated marginal means are adjusted by covariates in the model. All predictors have one degree of freedom. Effect sizes for ANCOVAs are reported as partial Eta squared (η2). Results Summary Statistics Descriptive statistics for model variables are reported in Table II. Mothers expressed positive emotions about 15.1 (SD = 9.37, range = 46) times on average over the course of the mealtime, and negative emotions about 2.5 (SD = 3.32, range = 18.5) times. Children expressed positive emotions about 14.6 times (SD = 9.2, range = 49.5) on average over the course of the mealtime, and negative emotions about 5.03 (SD = 3.94, range = 19.0) times. There were no significant or substantive differences in findings of any analyses when the seven male primary caregivers who completed surveys—but were not coded, as the coding scheme focused on mothers—were excluded or included. Aim 1: Characterize MEC Cluster analyses identified two distinct types of dyads. The first cluster accounted for 79% (n = 59) of the dyads in this sample, and was characterized by relatively high positive emotions in mothers and children, and relatively low negative emotions in mothers (Positive Expressers). The second cluster accounted for 20% (n = 15) of the dyads in this sample, and was characterized by a similar amount of positive and negative emotions in parents and children (All Expressers). Group characteristics are reported in Table II. Maternal emotions seemed to drive cluster placement. Kruskal–Wallis H tests found significant differences between groups on maternal positive emotion ratios and maternal negative emotion ratios. Although there were trends toward Positive Expresser dyads having children with higher positive and lower negative emotion ratios as compared with All Expressers, these differences were not statistically significant. Aim 2: Identify Predictors of MEC Kruskal–Wallis H tests (Table II) found significant differences between Positive Expressers and All Expressers on Food Involvement feeding practices and maternal depressive symptoms (Table II). Therefore, these variables were examined as potential predictors of MEC in logistic regression analyses. In the first block of logistic regression analyses, we examined demographics as independent variables, and MEC group membership as a dependent variable (Table III). Positive Expressers were coded as 0, and All Expressers were coded as 1. As the total number of family members increased, so too did the likelihood of being an All Expresser. No other variables had a significant effect. In the second block, we submitted the potential predictor variables identified in Kruskal–Wallis H tests to the logistic regression as independent variables, again examining MEC group membership as a dependent variable. Model fit and classification accuracy improved in this step. As food involvement increased, the likelihood of being an All Expresser decreased significantly. No other variables significantly predicted MEC group membership. Aim 3: Examine Outcomes of MEC Kruskal–Wallis H tests were used to examine mean differences on child outcome variables based on MEC group membership. The Positive Expressers group had significantly higher healthy food consumption at W2, as compared with All Expressers (Table II). There were no significant differences between groups on unhealthy food consumption or child BMI-P. Finally, ANCOVA was used to examine whether differences on W2 healthy food consumption between Positive and All Expressers persisted, controlling for covariates (Table IV). Controlling for child and parent gender, child age, total family members, income, mealtime length, and change in healthy food consumption from W1 to W2, Positive Expressers reported 1.01 (SE = 0.35, p = .005, 95% confidence interval  = 0.33, 1.70, R2 = .48; adjusted R2 = .41) more daily servings of healthy food at W2, compared with All Expressers. Discussion This study expands the scientific literature on the link between emotions and food in the family context in several ways. First, this is the first study to characterize family MEC using a count-based coding scheme and proportion scores, and to examine longitudinal predictors and outcomes of family MEC. Using this novel method, we found that MEC was determined primarily by maternal, not child, emotions. Dyads were categorized as either Positive Expressers (expressing high levels of positive emotion, and little negative emotion) or All Expressers (about equal levels of positive and negative emotion). Second, we aimed to examine whether feeding practices or maternal mental health were predictors of family MEC, given the prior research finding associations between parent emotional expression, depression, feeding, and mealtime practices (Haycraft, Farrow, & Blissett, 2013; Hughes et al., 2011; McCurdy et al., 2014). Food involvement feeding practices—but not maternal mental health or other feeding practices—predicted lower likelihood of being an All Expresser. Finally, this study examined whether family MEC predicted indicators of children’s health, including food consumption and weight outcomes. Positive Expressers reported that their children ate significantly more healthy food at W2, as compared with All Expressers, but there were no differences between groups by weight outcomes. Although interpretations of these results must be tempered by acknowledging some methodological limitations, they point to several avenues for future research and practice. Overall, findings suggest that MEC should be investigated as a potential, modifiable family-level factor to target in efforts to increase child consumption of healthy foods. Although cluster analyses accounted for maternal and child emotional expression, MEC groups had significantly different emotion profiles for mothers—but not children. Previous studies have found that family–child and parent–child interpersonal dynamics were linked to reduced child overweight, but that child–sibling interpersonal dynamics had no effect (Berge et al., 2014). Mothers’ emotions may have a larger effect on MEC overall because they exert more control over and spend more time engaged in orchestrating family routines during early childhood (Kotila, Schoppe-Sullivan, & Dush, 2013). Alternatively, children may have been more affected by the recording equipment during the mealtime than parents, tempering their negative emotional expression. It was surprising that a cluster of mothers and children with high negative emotion and low positive emotion did not emerge. Again, this may be because of social desirability biases, or the small and homogenous sample in the current study. We also found that parent report of food involvement feeding practices—but no other feeding practices, and not maternal mental health—was related to decreased likelihood of being an All Expresser. In a longitudinal study of 10–12-year-old Australian children, child involvement in food preparation, shopping, planning, or cleanup was not related to family mealtime frequency, dietary intake, or weight outcomes (Leech, McNaughton, Crawford, Campbell, Pearson, & Timperio, 2014). However, in a study among (n = 394) families with children aged between 18 months and 5 years, parent food involvement (or the importance, time, and value parents allocate to food) was associated with both child and parent fruit and vegetable consumption (Ohly et al., 2013). The Food Involvement subscale of the CFPQ includes three questions, all centered on the child’s involvement in food preparation, planning, and shopping (Musher-Eizenman & Holub, 2007). It is surprising that our results are somewhat contrary to the findings from Leech and colleagues’ (2014) study. However, involving younger children in food preparation and planning may present an opportunity for families to develop positive rapport around food and mealtimes. Indeed, a qualitative study with adult women found that negative affect was associated with lower food involvement for themselves (Jarman et al., 2012). Together, these findings suggest that the MEC may be determined, in part, by the degree to which parents are involved with food themselves, and the degree to which they involve their children with food. However, future research is needed to ascertain (1) whether parent report of child food involvement is indicative of true food involvement; (2) whether these effects are true only in young children; and (3) whether modifying food involvement could improve the MEC, and, ultimately, child food consumption. In contrast to the literature linking maternal depression to feeding and mealtime practices (Haycraft et al., 2013; McCurdy et al., 2014), we did not find associations with MEC group membership. On the one hand, it is possible that the sample was too small to detect differences in maternal mental health. Indeed, this sample’s range for stress and depressive symptoms at W2 is smaller than the range reported in a larger community sample of adults that completed the DASS (Crawford & Henry, 2003). Furthermore, although maternal depressive symptoms may serve as an indicator for maternal depression, the DASS cannot be used to diagnose individuals with depression and may not be comparable with previous studies using diagnostic assessments (Henry & Crawford, 2005). On the other hand, another study also found that parental depressive symptoms were not associated with mothers’ report of family relationship quality or family expressiveness, but was associated with use of fewer controlling parenting practices (Foster et al., 2008). Alternatively, it may be that family MEC is less associated with individual factors (such as maternal mental health), and more with factors related to the parent–child relationship (such as parenting and feeding practices), or the family system as a whole (such as food insecurity or household chaos; Fiese, Gundersen, Koester, & Jones, 2016). Finally, we found that parents in the Positive Expressers group reported greater child consumption of healthy food at W2, than All Expressers. There were no differences for unhealthy food consumption or child BMI-P. Previous studies found that parents of overweight children engaged in more hostility and food-related lecturing, as compared with families with non-overweight children (Berge et al., 2014). This may facilitate a counterproductive cycle, in which child overweight prompts hostility or food-related lecturing during mealtimes, which in turn may decrease healthy food consumption and eventually have an impact on weight (Francis, Hofer, & Birch, 2001; Webber, Cooke, Hill, & Wardle, 2010). Alternatively, if parents foster a more positive MEC, they may increase healthy food consumption by making the mealtime a more pleasant environment. Although other cross-sectional and longitudinal studies have found associations between parent–child interaction quality and risk for obesity and excessive weight gain (Anderson & Keim, 2016), this study found no association between family MEC and child weight outcomes. One possible explanation for this null finding is that the effects of MEC on weight as children age may be cumulative. For instance, although one study on families with children aged 6–12 years did find an association between the emotional atmosphere and risk for overweight (Berge et al., 2014), a narrative review found evidence to suggest that parent–child interaction quality in early childhood or infancy was more strongly and consistently associated with obesity risk during middle childhood, than during early childhood (Anderson & Keim, 2016). Links between parent–child interaction quality and weight outcomes were weak or nonsignificant for children under 5 years old, but associations seemed to increase in strength and consistency as children aged. Therefore, the effects of MEC on child weight may not be apparent until after the preschool years. Future research is needed to examine the longitudinal effects of MEC on child weight using observations across several time points. The current study has several strengths, including use of direct, home-based observations, a count-based coding scheme, and a longitudinal design. However, some limitations bear noting. First, it is possible that family members altered their behavior in response to the observation, although some dyads clearly expressed more negative emotions during the mealtime than others, suggesting that—while some families may have changed their behaviors according to social desirability—many did not. Second, measures assessing feeding practices, parent mental health, and child eating behaviors used self- or parent-report questionnaires, rendering these constructs susceptible to bias. In particular, parent report of child food consumption has been noted previously to have low internal reliability, and because there are currently no validated observational assessments of food consumption for children in the home environment, these methodological limitations may provide fruitful avenues for future research. Finally, our findings are based on a relatively small, homogenous community sample of mostly White mothers and their children, and should be replicated in a larger, more diverse sample of families. Additionally, although several male primary caregivers completed surveys, because we only coded maternal emotion during mealtimes, we cannot generalize our findings to fathers, and so focus our discussion on mothers. Clinical and Research Implications There are several implications from the findings of this study for practitioners and researchers. Results from this study will allow physicians and care providers to improve communication with parents about the health benefits of family mealtimes, by showing specific associations between family MEC and child food consumption. Given the challenging realities of mealtimes (e.g., food insecurity; picky eating) for many families (Malhotra, Herman, Wright, Bruton, Fisher, & Whitaker, 2013), it would be unreasonable to simply suggest that parents should express more positive emotion at meals. Instead, practitioners may find it helpful to provide parents with information about several evidence-based strategies that may reduce parent stress and negative affect during mealtimes, and indirectly improve child food consumption. For example, increasing food involvement, practicing child behavior management (Janicke et al., 2008; Stark et al., 2011), establishing routines and schedules (Jones, Fiese, & STRONG Kids Team, 2014), eliminating distractions (FitzPatrick, Edmunds, & Dennison, 2007), and engaging in direct, clear communication during mealtimes may all promote a more positive MEC (Fiese et al., 2006). Providers may be able to communicate more clearly about the benefits of family meals, by increasing awareness about the link between family MEC and child food consumption. This study identified also several avenues for future research. First, these findings should be replicated among a sample with more fathers, and in a larger and more diverse population, using multiple observations of family mealtimes in a longitudinal design. Second, studies using other measures of child food consumption (24-hr recalls, dietary intake diaries, and repeated food frequency questionnaires) are needed to validate links between family MEC and child food consumption. Third, investigators should examine how other family members or family factors (e.g., routines; distractions; communication) contribute to MEC, and ultimately, child health and well-being. Fourth, by using a coding and analytic approach that emphasizes behaviors on a moment-to-moment basis, researchers may be able to better understand the antecedents and consequences of positive and negative affect at the mealtime, thus deepening our understanding of the contributors to MEC. Finally, these preliminary findings suggest that more positive parent emotionality at meals may promote children’s healthy food consumption. Parenting interventions targeting emotion socialization have been remarkably successful at improving both parent and child emotional awareness and regulation (Havighurst, Wilson, Harley, Prior, & Kehoe, 2010; Wilson, Havighurst, & Harley, 2012). Therefore, researchers should consider assessing whether parent emotion socialization training could impact family MEC and subsequent child food consumption. To conclude, MEC is coconstructed by mothers and children (with more influence from mothers), and may impact child healthy food consumption. Children who experience more positive mealtime interactions may eat a greater amount of healthy food than children who experience mealtimes with about equal amounts of positive and negative emotion. These novel findings open up additional avenues of research, and suggest that further research on mealtime quality, associations between emotional expression and child food consumption, and emotion socialization training for parents may be needed. By helping parents engage in more positive mealtime interactions with their young children, practitioners and researchers may be able to increase children’s healthy food consumption. Funding This research was supported in part by grants from the Agriculture and Food Research Initiative of the United States Department of Agriculture (USDA) under the Illinois Transdisciplinary Obesity Prevention Program grant (2011-67001-30101) to the Division of Nutritional Sciences at the University of Illinois; the Illinois Council for Agriculture Research to Kristen Harrison (PI); the University of Illinois Health and Wellness Initiative to Barbara Fiese and Sharon Donovan; and the USDA Hatch (793-328) Program to Barbara Fiese (PI) and USDA Hatch to Kelly Bost. Conflicts of interest: None declared. References American Academy of Pediatrics. ( 2003). Prevention of pediatric overweight and obesity. Pediatrics , 112, 424– 430. Retrieved from www.aap.org CrossRef Search ADS PubMed  Anderson S. E., Keim S. A. ( 2016). Parent–child interaction, self-regulation, and obesity prevention in early childhood. Current Obesity Reports , 5, 192– 200. doi: 10.1007/s13679-016-0208-9 Google Scholar CrossRef Search ADS PubMed  Anderson S. E., Whitaker R. C. ( 2010). Household routines and obesity in US preschool-aged children. Pediatrics , 125, 420– 428. doi: 10.1542/peds.2009-0417. Google Scholar CrossRef Search ADS PubMed  Anzman S. L., Rollins B. Y., Birch L. L. ( 2010). Parental influence on children’s early eating environments and obesity risk: Implications for prevention. International Journal of Obesity , 34, 1116– 1124. doi: 10.1038/ijo.2010.43 Google Scholar CrossRef Search ADS PubMed  Bazzano L. A., He J., Ogden L. G., Loria C. M., Vupputuri S., Myers L., Whelton P. K. ( 2002). Fruit and vegetable intake and risk of cardiovascular disease in US adults: The first National Health and Nutrition Examination Survey Epidemiologic follow-up study. The American Journal of Clinical Nutrition , 76, 93– 99. Retrieved from www.ajcn.nutrition.org Google Scholar CrossRef Search ADS PubMed  Berge J. M., Rowley S., Trofholz A., Hanson C., Rueter M., MacLehose R. F., Neumark-Sztainer D. ( 2014). Childhood obesity and interpersonal dynamics during family meals. Pediatrics , 134, 923– 932. doi: 10.1542/peds.2014-1936 Google Scholar CrossRef Search ADS PubMed  Chorney J. M., McMurtry C. M., Chambers C. T., Bakeman R. ( 2015). Developing and modifying behavioral coding schemes in pediatric psychology: A practical guide. Journal of Pediatric Psychology , 40, 154– 164. doi: 10.1093/jpepsy/jsu099 Google Scholar CrossRef Search ADS PubMed  Cole P., Wiggins C., Radzioch A., Pearl A. ( 2007). D.O.T.S. Emotion coding system. Unpublished Manuscript, Department of Psychology, Pennsylvania State University. Crawford J. R., Henry J. D. ( 2003). The Depression Anxiety Stress Scales (DASS): Normative data and latent structure in a large non‐clinical sample. British Journal of Clinical Psychology , 42, 111– 131. Google Scholar CrossRef Search ADS PubMed  Demory-Luce D., Morales M., Nicklas T., Baranowski T., Zakeri I., Berenson G. ( 2004). Changes in food group consumption patterns from childhood to young adulthood: The Bogalusa Heart Study. Journal of the American Dietetic Association , 104, 1684– 1691. doi: 10.1016/j.jada.2004.07.026 Google Scholar CrossRef Search ADS PubMed  Esmaillzadeh A., Kimiagar M., Mehrabi Y., Azadbakht L., Hu F. B., Willett W. C. ( 2006). Fruit and vegetable intakes, C-reactive protein, and the metabolic syndrome. The American Journal of Clinical Nutrition , 84, 1489– 1497. Retrieved from www.ajcn.nutrition.org Google Scholar PubMed  Fiese B. H., Foley K. P., Spagnola M. ( 2006). Routine and ritual elements in family mealtimes: Contexts for child well-being and family identity. New Directions for Child and Adolescent Development , 2006, 67– 89. doi:10.1002/cd.156 Google Scholar CrossRef Search ADS   Fiese B. H., Gundersen C., Koester B., Jones B. ( 2016). Family chaos and lack of mealtime planning is associated with food insecurity in low income households. Economics and Human Biology , 21, 147– 155. doi: 10.1016/j.ehb.2016.01.004 Google Scholar CrossRef Search ADS PubMed  Fiese B. H., Hammons A., Grigsby-Toussaint D. ( 2012). Family mealtimes: A contextual approach to understanding childhood obesity. Economics and Human Biology , 10, 365– 374. doi: 10.1016/j.ehb.2012.04.004 Google Scholar CrossRef Search ADS PubMed  Fitzgibbon M., Hayman L., Haire-Joshu D. ( 2008). Childhood obesity: A policy statement of the society of behavioral medicine. Society of Behavioral Medicine. Retrieved from http://www.sbm.org FitzPatrick E., Edmunds L. S., Dennison B. A. ( 2007). Positive effects of family dinner are undone by television viewing. Journal of the American Dietetic Association , 107, 666– 671. doi: 10.1016/j.jada.2007.01.014 Google Scholar CrossRef Search ADS PubMed  Fosco G. M., Grych J. H. ( 2013). Capturing the family context of emotion regulation: A family systems model comparison approach. Journal of Family Issues , 34, 557– 578. doi: 10.1177/0192513X12445889 Google Scholar CrossRef Search ADS   Foster C. E., Webster M. C., Weissman M. M., Pilowsky D. J., Wickramaratne P. J., Rush A. J., Hughes C. W., Garber J., Malloy E., Cerda G., Kornstein S. G., Alpert J. E., Wisniewski S. R., Trivedi M. H., Fava M., King C. A. ( 2008). Course and severity of maternal depression: Associations with family functioning and child adjustment. Journal of Youth and Adolescence , 37, 906– 916. doi: 10.1007/s10964-007-9216-0 Google Scholar CrossRef Search ADS PubMed  Francis L. A., Hofer S. M., Birch L. L. ( 2001). Predictors of maternal child-feeding style: Maternal and child characteristics. Appetite , 37, 231– 243. doi: 10.1006/appe.2001.0427 Google Scholar CrossRef Search ADS PubMed  Fulkerson J. A., Larson N., Horning M., Neumark-Sztainer D. ( 2014). A review of associations between family or shared meal frequency and dietary and weight status outcomes across the lifespan. Journal of Nutrition Education and Behavior , 46, 2– 19. doi: 10.1016/j.jneb.2013.07.012 Google Scholar CrossRef Search ADS PubMed  Hammons A. J., Fiese B. H. ( 2011). Is frequency of shared family meals related to the nutritional health of children and adolescents? Pediatrics , 127, e1565– e1574. doi: 10.1542/peds.2010-1440 Google Scholar CrossRef Search ADS PubMed  Harrison K., Liechty J. M.; The STRONG Kids Program. ( 2012). US preschoolers’ media exposure and dietary habits: The primacy of television and the limits of parental mediation. Journal of Children and Media , 6, 18– 36. doi: 10.1080/17482798.2011.633402 Google Scholar CrossRef Search ADS   Havighurst S. S., Wilson K. R., Harley A. E., Prior M. R., Kehoe C. ( 2010). Tuning in to Kids: Improving emotion socialization practices in parents of preschool children—findings from a community trial. Journal of Child Psychology and Psychiatry , 51, 1342– 1350. doi: 10.1111/j.1469-7610.2010.02303.x Google Scholar CrossRef Search ADS PubMed  Haycraft E., Farrow C., Blissett J. ( 2013). Maternal symptoms of depression are related to observations of controlling feeding practices in mothers of young children. Journal of Family Psychology , 27, 159– 64. doi: 10.1037/a0031110 Google Scholar CrossRef Search ADS PubMed  Henry J. D., Crawford J. R. ( 2005). The short-form version of the Depression Anxiety Stress Scales (DASS-21): Construct validity and normative data in a large non-clinical sample. British Journal of Clinical Psychology , 44, 227– 239. doi: 10.1348/014466505X29657 Google Scholar CrossRef Search ADS PubMed  Hughes S. O., Power T. G., Papaioannou M. A., Cross M. B., Nicklas T. A., Hall S. K., Shewchuk R. M. ( 2011). Emotional climate, feeding practices, and feeding styles: An observational analysis of the dinner meal in Head Start families. International Journal of Behavioral Nutrition and Physical Activity , 8, 60. doi:10.1186/1479-5868-8-60 Google Scholar CrossRef Search ADS PubMed  IBM Analytics. ( 2016, March 17). Retrieved from http://www.ibm.com/analytics/us/en/technology/spss/ Retrieved 18 April 2016. Janicke D. M., Mitchell M. J., Quittner A. L., Piazza-Waggoner C., Stark L. J. ( 2008). The impact of behavioral intervention on family interactions at mealtime in pediatric cystic fibrosis. Children's Health Care , 37, 49– 66. doi: 10.1080/02739610701766891 Google Scholar CrossRef Search ADS   Jarman M., Lawrence W., Ntani G., Tinati T., Pease A., Black C., Baird J., Barker M.; SIH Study Group. ( 2012). Low levels of food involvement and negative affect reduce the quality of diet in women of lower educational attainment. Journal of Human Nutrition and Dietetics , 25, 444– 452. doi: 10.1111/j.1365-277X.2012.01250.x Google Scholar CrossRef Search ADS PubMed  Jones B. L., Fiese B. H.; STRONG Kids Team. ( 2014). Parent routines, child routines, and family demographics associated with obesity in parents and preschool-aged children. Frontiers in Psychology , 5, 374. doi: 10.3389/fpsyg.2014.00374 Google Scholar PubMed  Kotila L. E., Schoppe-Sullivan S. J., Dush C. M. ( 2013). Time in parenting activities in dual‐earner families at the transition to parenthood. Family Relations , 62, 795– 807. doi: 10.1111/fare.12037 Google Scholar CrossRef Search ADS PubMed  Krishnamoorthy J. S., Hart C., Jelalian E. ( 2006). The epidemic of childhood obesity: Review of research and implications for public policy  (Vol. 20). Ann Arbor, MI: Society for Research in Child Development. Retrieved from http://eric.ed.gov Kuczmarski R. J., Ogden C. L., Grummer-Strawn L. M., Flegal K. M., Guo S. S., Wei R., Mei Z., Curtin L. R., Roche A. F., Johnson C. L. ( 2000). CDC growth charts: United States. Advance Data , 314, 1– 27. Retrieved from http://www.cdc.gov Leech R. M., McNaughton S. A., Crawford D. A., Campbell K. J., Pearson N., Timperio A. ( 2014). Family food involvement and frequency of family dinner meals among Australian children aged 10–12 years. Cross-sectional and longitudinal associations with dietary patterns. Appetite , 75, 64– 70. doi: 10.1016/j.appet.2013.12.021 Google Scholar CrossRef Search ADS PubMed  Lien N., Lytle L. A., Klepp K.-I. ( 2001). Stability in consumption of fruit, vegetables, and sugary foods in a cohort from age 14 to age 21. Preventive Medicine , 33, 217– 226. doi: 10.1006/pmed.2001.0874 Google Scholar CrossRef Search ADS PubMed  Lytle L. A., Seifert S., Greenstein J., McGovern P. ( 2000). How do children’s eating patterns and food choices change over time? Results from a cohort study. American Journal of Health Promotion , 14, 222– 228. doi: 10.4278/0890-1171-14.4.222 Google Scholar CrossRef Search ADS PubMed  Malhotra K., Herman A. N., Wright G., Bruton Y., Fisher J. O., Whitaker R. C. ( 2013). Perceived benefits and challenges for low-income mothers of having family meals with preschool-aged children: Childhood memories matter. Journal of the Academy of Nutrition and Dietetics , 113, 1484– 1493. doi: 10.1016/j.jand.2013.07.028 Google Scholar CrossRef Search ADS PubMed  McCurdy K., Gorman K. S., Kisler T., Metallinos-Katsaras E. ( 2014). Associations between family food behaviors, maternal depression, and child weight among low-income children. Appetite , 79, 97– 105. doi: 10.1016/j.appet.2014.04.015 Google Scholar CrossRef Search ADS PubMed  McPhie S., Skouteris H., Daniels L., Jansen E. ( 2014). Maternal correlates of maternal child feeding practices: A systematic review. Maternal and Child Nutrition , 10, 18– 43. doi: 10.1111/j.1740-8709.2012.00452.x Google Scholar CrossRef Search ADS PubMed  Musher-Eizenman D., Holub S. ( 2007). Comprehensive feeding practices questionnaire: Validation of a new measure of parental feeding practices. Journal of Pediatric Psychology , 32, 960– 972. doi: 10.1093/jpepsy/jsm037 Google Scholar CrossRef Search ADS PubMed  National Center for Educational Statistics. ( 2007). Early Childhood Longitudinal Program (ECLS)—Overview. Retrieved from https://nces.ed.gov/ Ohly H., Pealing J., Hayter A. K., Pettinger C., Pikhart H., Watt R. G., Rees G. ( 2013). Parental food involvement predicts parent and child intakes of fruits and vegetables. Appetite , 69, 8– 14. doi: 10.1016/j.appet.2013.05.003 Google Scholar CrossRef Search ADS PubMed  Rollins B. Y., Belue R. Z., Francis L. A. ( 2010). The beneficial effect of family meals on obesity differs by race, sex, and household education: The National Survey of Children’s Health, 2003-2004. Journal of the American Dietetic Association , 110, 1335– 1339. doi: 10.1016/j.jada.2010.06.004 Google Scholar CrossRef Search ADS PubMed  Stark L. J., Spear S., Boles R., Kuhl E., Ratcliff M., Scharf C., Bolling C., Rausch J. ( 2011). A pilot randomized controlled trial of a clinic and home‐based behavioral intervention to decrease obesity in preschoolers. Obesity , 19, 134– 141. doi: 10.1038/oby.2010.87 Google Scholar CrossRef Search ADS PubMed  Steinley D. ( 2006). K-means clustering: A half-century synthesis. British Journal of Mathematical and Statistical Psychology , 59, 1– 34. doi: 10.1348/000711005X48266 Google Scholar CrossRef Search ADS PubMed  Wang X., Ouyang Y., Liu J., Zhu M., Zhao G., Bao W., Hu F. B. ( 2014). Fruit and vegetable consumption and mortality from all causes, cardiovascular disease, and cancer: Systematic review and dose-response meta-analysis of prospective cohort studies. BMJ , 349, g4490. doi: 10.1136/bmj.g4490 Google Scholar CrossRef Search ADS PubMed  Webber L., Cooke L., Hill C., Wardle J. ( 2010). Associations between children’s appetitive traits and maternal feeding practices. Journal of the American Dietetic Association , 110, 1718– 1722. doi: 10.1016/j.jada.2010.08.007 Google Scholar CrossRef Search ADS PubMed  Wilson K. R., Havighurst S. S., Harley A. E. ( 2012). Tuning in to kids: An effectiveness trial of a parenting program targeting emotion socialization of preschoolers. Journal of Family Psychology , 26, 56– 65. doi: 10.1037/a0026480 Google Scholar CrossRef Search ADS PubMed  Yoder P., Symons F. ( 2010). Observational measurement of behavior . New York, NY: Springer Publishing Company. © The Author 2017. Published by Oxford University Press on behalf of the Society of Pediatric Psychology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

Journal of Pediatric PsychologyOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off