Poster #154 - Trust Development as an Evaluative Conditioning Process: Testing Contingency Effects

• In Event: 3-003 - Poster Session 09
  In Poster Session: PS 09 Section - Parenting & Parent-Child Relationships

Integrative Statement

Trust in parental support and subsequent support seeking is a hallmark of secure attachment and is thought to result from experiences with sensitive parents during distress. Trust development has been extensively studied in longitudinal research using broadband measures of sensitive parenting and secure attachment. However, little is known about the underlying learning mechanisms. Therefore, we tested the hypothesis that trust in caregivers is the result of an evaluative conditioning learning process in which trust develops when the probability is higher that the caregiver provides successful support. This is the contingency between the support of the caregiver during distress and the positive effect of this support on the distress.
To test this hypothesis, we developed a paradigm in which children have to find food for a hungry donkey and in which a novel caregiver helps. The experiment consists of a learning phase (10 trials in which the caregiver always helps) and a test phase (20 trials in which children have to call the same caregiver for help). Before the experiment, we measured children’s self-reported trust in their mother’s support. Before and after the learning phase, and after the test phase we measured children’s trust in the novel caregiver. In the test phase, we observed children’s inclination to seek the novel caregiver’s help.
In Study 1, 65 children (50.8% boys; Mage = 10.22, SDage = 1.05) were randomly assigned to a 50%, a 70%, or a 100% contingent learning phase condition. In the test phase, this caregiver was 100% contingent. In study 2, 170 children (49.7% boys; Mage = 9.58; SDage = 1.49) were randomly assigned to a 20%, an 80%, or a 100% contingent learning phase condition. The test phase’s caregiver was 20% or 100% contingent.
In both studies, trust in mother and trust in the novel caregiver were marginally significantly (1, r(62) = .24, p = .062) and significantly (2, r(170) = .16, p = .043) correlated. Study 1 (Table 1) showed that the learning phase contingency manipulation caused increases in trust in the 100% condition and decreases in trust as contingency decreased. Moreover, lower contingency decreased children’s support seeking behavior in the test phase. Finally, results show that 100% contingency in the test phase was linked with a significant catch-up in terms of increases in trust and support seeking behavior. However, the initial learning phase contingency manipulation retained an effect. Study 2 (Table 2) confirmed and extended Study 1, showing that 80% contingency in the learning phase was sufficient to significantly increase trust and that reducing the contingency from 100%/80% in the learning phase to 20% in the test phase fully erased the earlier learning effects.
Results supported the hypothesis that trust and support seeking result from an evaluative conditioning learning process. Moreover, it shows that novel learning experiences revise the level of trust, that less positive initial learning experiences have a longer impact on subsequent learning, while more positive initial learning experiences are more easily erased by more negative novel learning experiences. Implications for attachment theory and trust development are discussed.

Authors

• Guy Bosmans, Katholieke Universiteit Leuven
  Presenting Author

• Theodore E. A. Waters, New York University- Abu Dhabi
  Non-Presenting Author

• Simon De Winter, KU Leuven
  Non-Presenting Author

• Dirk Hermans, KU Leuven
  Non-Presenting Author