Eating, feeding, and feeling: emotional responsiveness mediates longitudinal associations between maternal binge eating, feeding practices, and child weight

Table 2 Direct effects of maternal binge eating (BE) on feeding practices (CFPQ), and indirect effects of BE on feeding practices through responses to child negative emotion (CCNES) using bias-corrected bootstrapped confidence intervals (CI) and full-information maximum likelihood estimation ^a,b

	Direct effect		Indirect effect		Total effect
	BE→CFPQ		BE→CCNES→CFPQ		Total effect
IV: Maternal BE	B (SE)	95 % CI	B (SE)	95 % CI	B (SE)	95 % CI
M: CCNES Distress Responses
DV: CFPQ Emotion regulation	.015 (.035)	(-.051, .084)	.016 (.011)*	(.003, .050)	.031 (.032)	(-.026, .096)
DV: CFPQ Food reward	.079 (.078)	(-.079, .218)	.031 (.016)*	(.010, .088)	.110 (.081)	(-.044, .267)
DV: CFPQ Restriction for health	.130 (.078)*	(.012, .290)	.039 (.020)*	(.013, .105)	.169 (.082)*	(.046, .347)
DV: CFPQ Pressure to eat	-.114 (.058)*	(-.263, -.026)	.024 (.017)*	(.003, .069)	-.090 (.051)*	(-.214, -.008)
DV: CFPQ Balance/variety	.052 (.044)	(-.032, .092)	-.016 (.012)*	(-.049, -.004)	.036 (.040)	(-.049, .107)
DV: CFPQ Involvement	.024 (.060)	(-.095, .143)	-.023 (.016)*	(-.083, -.003)	.001 (.062)	(-.122, .120)
M: CCNES Problem-focused Responses
DV: CFPQ Balance/variety	.063 (.040)	(-.004, .154)	-.030 (.026)	(-.107, .006)	.033 (.036)	(-.046, .097)
DV: CFPQ Involvement	.037 (.060)	(-.072, .166)	-.029 (.021)*	(-.100, -.003)	.008 (.063)	(-.117, .128)
DV: CFPQ Modeling	-.043 (.061)	(-.163, .078)	-.034 (.027)	(-.112, .002)	-.078 (.067)	(-.112, .002)
M: CCNES Emotion-focused Responses
DV: CFPQ Balance/variety	.065 (.043)	(-.009, .157)	-.033 (.018)*	(-.094, -.010)	.032 (.038)	(-.045, .104)
DV: CFPQ Modeling	-.045 (.064)	(-.173, .080)	-.034 (.020)*	(-.096, -.009)	-.079 (.065)	(-.214, .044)

Note. IV Independent Variable, M Mediator, DV Dependent Variable, CFPQ Comprehensive Feeding Practices Questionnaire, CCNES Coping with Children’s Negative Emotions Scale, BE Binge eating
^aAll models controlled for maternal BMI at Wave 1 and 2, and change in feeding practices from Wave 1 to Wave 2. Pressure to Eat feeding practices were also related to parent age, child age, child BMI percentile, and parent race/ethnicity and so these were included as additional controls in analyses estimating effects on Pressure to Eat only. Only pathways in which the IV was associated with the mediator, and the mediator was associated with the DV were tested, in accordance with criteria for testing mediation (Hayes et al., 2013)
^bIn order to account for missingness on exogenous covariates, all control variables were brought into the model in Mplus. There were few differences in results between models with control variables and without control variables in the model. Therefore, conservative findings with covariates in the model and no missingness are presented

ISSN: 1479-5868