Abstract
The present study examined conversations of 164 mothers from seven different cultural contexts when reminiscing with their 3-year-old children. We chose samples based on their sociodemographic profiles, which represented three different cultural models: (1) autonomy (urban middle-class families from Western societies), (2) relatedness (rural farming families from non-Western societies), and (3) autonomy-relatedness (urban middle-class families from non-Western societies). The results showed that mothers from the autonomous contexts predominantly adopted an elaborative-evaluative reminiscing style (variable-oriented approach) and pattern (person-oriented approach). Mothers from the relational contexts mainly adopted a repetitive reminiscing style and pattern. There was greater heterogeneity for style variables among mothers from autonomous-relational contexts; in addition, the person-oriented approach revealed that the majority of mothers from autonomous-relational contexts showed hybrid style patterns. Thus, the cultural models, and their respective orientations towards autonomy and relatedness, were reflected in the way mothers reminisced with their children. The children's provision of memory elaborations was high in the autonomous contexts, low in the relational contexts, and moderate in the autonomous-relational contexts. Across contexts, maternal evaluations prompted children to contribute memory elaborations. Maternal elaborations were an additional predictor for children's memory, but only for families from the relational cultural model.
ACKNOWLEDGMENTS
We thank the German Research Council (KE 263/46-1 bis 46-4) and the following research assistants: María Andrea Araya, Patricia Soley, Darshana Desai, Yogesh Ramdasani, Rashi Gupta, Pooja Vadehra, Pooja Bhargava, Deepa Gupta, Eunice Kiyeinimo Yovsi, and Sister Fai Emma. We also thank the Deepak Foundation, India.
Notes
Note. Indexed letters indicate results of simple main effects testing (using LSD, p < .05); samples not sharing indexed letters differ significantly from each other.
Note. Indexed letters indicate results of simple main effects testing (using LSD, p < .05); samples not sharing indexed letters differ significantly from each other.
Note. SR = Standardized residuals.
Note. **p < .01 (two-tailed).
1We ran additional regression analyses in which we entered sociodemographic variables in a first step prior to entering maternal reminiscing variables, but in no case did the sociodemographic variables predict unique variance in children's memory elaborations after entry of the maternal reminiscing variables.
2The results did not change when excluding the urban Nso sample from the regression analysis.