The thesis examined the cultural belief systems of mothers related to early mother-child interactions in New Zealand and Malaysia and within the Malaysian multicultural subgroups. The overall aim was to examine reported and observed interaction patterns of these cultural groups and to determine the relationship between specific measures of mothers’ verbal productivity and interaction behaviours with children’s verbal productivity. Two studies with different methodologies were designed. The first study used a survey to examine the reported beliefs and practices related to mother-child interactions of New Zealand and Malaysian mothers. A total of 284 mothers with children between 2- and 4-years old completed a written questionnaire on aspects related to talking with children, family and child-rearing values, and teaching and learning patterns in young children. Chi-square analyses revealed New Zealand and Malaysian mothers differed significantly (p<0.001) on 39 of 60 belief and practice statements. Subgroup analysis revealed Malaysian mothers of Malay, Chinese and Indian ethnicities were more similar than different in their reported beliefs and practices; however, there were some statistically significant group differences. In the second study, a subsample of 48 Malaysian mothers was video-recorded interacting with their children. Mother-child dyadic language samples were recorded, transcribed and analysed. Measures of verbal productivity and interaction behaviours were coded and statistically analysed. The findings revealed most mothers, irrespective of their ethnicities, used response-control utterances when interacting with their children. Correlation analyses revealed mostly positive relationships between measures of mothers’ and children’s verbal productivity, and mostly negative relationship between measures of mothers’ interaction behaviours and children’s verbal productivity. Results are discussed with reference to the literature, Bronfenbrenner’s ecological model and language-learning models. The findings have implications for Speech-Language Therapists who use intervention programs that recommend the use of language-modelling utterances to develop children’s language skills. A five-point guideline to navigate belief systems when working with diverse populations is proposed.