**4. Discussion**

Responsive caregiving is a dimension of parenting in the early childhood years that is consistently related to subsequent cognitive and socioemotional aspects of functioning [2,47,48]. This aspect of caregiving is modifiable. Results from randomized controlled trials in developed and developing countries show that it is possible to improve this aspect of parenting, that doing so improves children's developmental outcomes, and that it may be the modifiable risk factor with the strongest effect on later brain development [3,49,50]. Although marked improvements to long-term child outcomes have now been demonstrated for parental responsivity interventions [4], moving from efficacy to effectiveness has proven to be a substantial challenge. If scalability is to be achieved, it is essential to have an instrument that can measure parental responsivity. Unfortunately, however, observational measures are the only ones that show good validity and most previous measures take a long time to administer and code. On the other hand, the literature suggests that adaptation of instruments that have shown reliability and validity evidence is more advantageous in LMIC than novel instrument development [19].

The cross-cultural adaptation of RIFL-P involved gold standard methodology to avoid cultural bias and the measure was successfully adapted to Portuguese and the Brazilian context. The adapted RIFL-P measure showed strong psychometric properties, mirroring the findings from the original instrument. As it is efficient, only taking eight minutes to assess and code, it is a measure that can be used at the population level. Since the sample was based on a stratified sample from a population cohort in one city of Brazil and refusal rates from the cohort were low, a high level of representativeness was achieved. A recent global effort led by the World Health Organization and UNICEF to monitor the Nurturing Care Framework's responsive caregiving component only provided proxy indicators (e.g., parental mental health, childcare availability, parental support) [51]. To make real progress in this realm, it is imperative to assess responsive interactions at the population level with measures, such as RIFL-P, that are psychometrically strong and quick to train, administer, and code. An online course in three languages is available to researchers or parenting professionals free of charge by contacting the first or last authors.

Four findings are worthy of further comment. First, the mean score for responsive interactions in Brazil was lower than that found in Canada. This was expected given the results of earlier studies in which the parenting of Latino mothers has been found to be more directive and controlling [52], with a lack of emphasis on understanding the child's internal goals and need for autonomy [53]. Of course, cultures differ in their values, customs, and beliefs [54], with the result that some parental behaviors receive less attention than others [55]. Regarding childrearing practices, Latino mothers tend to value obedience and politeness, physically guide their toddler´s actions, attribute less importance to children´s autonomy, and report greater use of discipline as a teaching method when compared to European American mothers of the same socioeconomic status [56]. Given that parental responsivity (although lower in Brazil) was found to relate to children's developmental outcomes, this does suggest that improving parental responsivity should be an important policy goal in Brazil. Second, in this study, Brazilian mothers were more responsive to girls than to boys, according to the RIFL-P scores. This finding should be cautiously interpreted as the same gender difference was not observed based on PICCOLO scores and a meta-analysis of parenting as a function of gender reported no significant differences [57]. An examination of this in future studies in Brazil is encouraged. Third, for purposes of criterion validity, we examined relationships with child outcomes, as well as other indices of parenting and contextual risk. Findings showed that both parenting measures were significantly related (i.e., RIFL-P and PICCOLO scores), as well as RIFL-P scores and a measure of parenting stimulation [44], and the strongest relationships with child outcomes were observed for the cognitive and language domains, which is in line with previous findings for the RIFL-P in a Canadian sample [15]. RIFL-P scores, both in Brazil and Canada, have been found to be associated with contextual risk (inversely). Fourth, as in previous work, low socioeconomic status (SES) was found in the current study to be associated with lower levels of parental responsivity. The association between socioeconomics and parenting has been widely reported [58] with evidence that economic hardship leads to parental emotional distress which impairs parenting [59]. It is also notable that economic hardship does not operate on its own. SES clusters risk [60] such that children are exposed to multiple challenges that compromise development (e.g., maternal depression, unemployment, domestic violence, poor neighborhoods) and parenting [61].
