Poster #96 - Variable- and Person-Centered Approaches for Modeling Toxic Stress: Associations with parent executive function and sensitivity

• In Event: 1-061 - Poster Session 02
  In Poster Session: PS 02 Section - Family Context & Processes

Integrative Statement

Introduction: Families living around the poverty line are disproportionately exposed to a multitude of poverty-related stressors in their environments that can undermine positive parenting behaviors. An enduring question concerns how to statistically model the multiple and often co-occurring stressors that low-income families face. In particular, whether it is the accumulation of poverty-related risk factors or the particular patterns of risk factors that “matters” for parents in poverty is not well understood. The current analysis extends this literature by generating both variable-centered cumulative risk and person-centered latent class models of risk in order to explore whether the accumulation of risk or specific patterns of risks, respectively, are associated with variation in levels of sensitive parenting behaviors in a sample of primarily low-income and urban-residing families. Consistent with recent cognitive models of parenting, the analysis also explored whether parent executive function (EF) statistically moderated the association between poverty-related risk exposure and sensitive parenting.

Methods: Data were collected from 259 primarily Hispanic/Latino parents during home visits. Cumulative risk and latent class models were generated using six indicators of poverty-related risk: income-to-needs, depression, intimate partner violence (IPV; separate measures of physical violence and psychological violence), household crowding, and marital status. Parents were administered a computerized Hearts and Flowers task to assess EF. Parents were also filmed interacting with their child in a play activity, from which levels of sensitivity were coded. Hypotheses: We expected that higher cumulative risk would be associated with lower levels of sensitivity overall. We also expected that high parent EF would buffer the negative association between cumulative risk and sensitivity. In latent class analyses, we expected to observe theoretically meaningful classes of families in the data, who varied according to their particular patterns of risk exposure. We expected that levels of sensitivity might vary between latent classes. We also tested whether parent EF moderated the association between class membership and sensitivity.

Results and Discussion: We observed no association between cumulative risk and sensitivity (β = -0.01, p = 0.86). There was an interaction between cumulative risk and parent EF (β = -0.12, p = 0.06) suggesting a trend in which higher cumulative risk was associated with lower sensitivity among those parents with lower EF. In latent class analyses, a 3-class model fit the data best (Figure 1). Parents belonged to one of three latent classes: a Poor and Crowded class, a Violence and Depressed class, and an Average Risk class. We found no between-class differences in levels of sensitivity, on average (χ2(2) = 1.28, p = 0.52). There was, however, an interaction between class membership and parent EF predicting sensitivity (χ2(2) = 10.65, p < 0.01), suggesting lower sensitivity among parents in the Violence and Depressed class who also exhibited lower EF (i.e. longer latencies; Figure 2). Findings suggest a possible protective role of high parent EF for parents who are exposed to IPV and depression. In general, the findings increase our understanding of the co-occurrence and consequences of multiple-risk exposures for low-income families and have implications for intervention programs.

Authors

• Eric Finegood, New York University
  Presenting Author

• Cybele Raver, New York University
  Non-Presenting Author

• Clancy Blair, New York University
  Non-Presenting Author