Ellen Roche
LightNews LightNews
banner
ellroche.bsky.social
Ellen Roche
@ellroche.bsky.social
2.6K followers 3.5K following 310 posts
** searching for fall 2026 postdoc **

Science — how do early experiences shape emotional development & regulation? 🧠

Democracy — how do emotion regulation/co-regulation support our highest ideals (freedom, justice, peace, etc)? 🌈

ellencroche.com
Posts Replies Media Videos
ellroche.bsky.social
Thanks!
January 31, 2026 at 1:40 AM
ellroche.bsky.social
I'm excited about this finding because babies grow up in complex relational ecosystems of multiple familial and non-familial caregivers. Real-time consistency *between* those caregivers is probably playing an important role in development, which we completely miss when we narrowly focus on dyads.
January 30, 2026 at 4:28 PM
ellroche.bsky.social
Ultimately even with a logistic model controlling for the individual sensitivity of both parent & grandparent, child age, & affective state in the prior block, we found that mutual sensitivity predicted infant affective state at the 30s level (p < .001). No significant findings in JVA model.
Hypothesis 2: mutual sensitivity predicts infant across-screen JVA In our JVA model we tested the relationship between mutual sensitivity and infant joint attention with their across-screen conversational partner, controlling for infant age. Here, we hypothesize that greater mutual sensitivity will be associated with a greater likelihood of successful infant engagement in across-screen joint visual attention. The image below depicts an infant’s bid for attention to the dog on the grandparent’s side of the screen, an example of across-screen JVA> Results H1: Mutual sensitivity predicts affective state Even after controlling for the individual sensitivity of each caregiver and child age, and after resolving issues of collinearity, mutual sensitivity significantly predicted infant affective state. (b = 0.78, p < .001). A 1-point increase in the distance between caregiver sensitivity was associated with a 22-percent decreased likelihood that the infant would be in the positive, alert affective state. Block order and the infant’s affective state in the preceding block were associated with infant affective state in the current block. H2: Mutual sensitivity predicts successful across-screen JVA When controlling for infant age, parent sensitivity, and grandparent sensitivity, mutual sensitivity was not significantly associated with successful across-screen JVA. link to the full paper in infancy: https://onlinelibrary.wiley.com/doi/10.1111/infa.70068
January 30, 2026 at 4:27 PM
ellroche.bsky.social
I've wanted to look at mutual sensitivity (consistency between caregivers in their sensitivity) for years, and this analysis was buoyed by encouragement at @affectscience.bsky.social 2023 after a flash talk exploring the method. Thanks to @tomhollenstein.bsky.social for state space inspiration!
What is mutual sensitivity between two caregivers? A growing body of research suggests that consistency in the caregiving environment predicts adaptive child development (see Ugarte & Hastings, 2023). Most studies have focused on consistency within a single caregiver over time. However, children do not grow up in isolated parent-child dyads. Is real-time consistency across two caregivers uniquely predictive of infant affective state (H1) and joint visual attention (H2)? This figure depicts real-time state spaces combining parent and grandparent sensitivity over three sessions within one family. Our mutual sensitivity variable was calculated by taking the square root of the distance between parent and grandparent. Eg., a distance of 0 meant the two caregivers shared a sensitivity score within a thirty second block. Importantly, our mutual sensitivity measure was agnostic to the degree of sensitivity of either caregiver. We collected data during the COVID-19 pandemic, inviting families with infants who had grandparents at a distance to record three Zoom conversations between the infant (see ages below) and the grandparent. 43 families submitted 3 video chats; 4 submitted only 2. Our final models included 2741 30s affect blocks and 660 JVA observations in multilevel logistic regressions. Hypothesis 1: Mutual sensitivity predicts infant affective state Prior theories and studies have linked young children’s affective states with language development, attentional control, and joint attention, but findings diverge on the role of affective valence, and rarely combine both affective valence and behavioral arousal. We chose a theoretically and empirically driven multidimensional affective state characterized by neutral-to-positive valence (see Bloom, 1993; Kubicek & Emde, 2012; Laake and Bridgett, 2014) and low-medium arousal, hypothesizing that mutual sensitivity would predict whether or not infants occupied this state space, even after controlling for the individual sensitivity of each caregiver and child age. Each infant occupies a distinct landscape of states.
January 30, 2026 at 4:25 PM
ellroche.bsky.social
Ultimately even with a logistic model controlling for the individual sensitivity of both parent & grandparent, child age, & affective state in the prior block, we found that mutual sensitivity predicted infant affective state at the 30s level (p < .0001). No significant findings in JVA model.
January 30, 2026 at 4:17 PM
ellroche.bsky.social
This is fantastic, congrats!
January 20, 2026 at 11:34 PM
ellroche.bsky.social
I’m excited to meet you in person please say hello!
September 6, 2025 at 12:49 PM