Investigating child-computer interactions within their contexts is vital for designing technology that caters to children's needs. However, determining what aspects of context are relevant for designing child-centric technology remains a challenge. We introduce EXPLORA, a multimodal, multistage online methodology comprising three pivotal stages: (1) building a teacher-apprentice relationship,(2) learning from child-teachers, and (3) assessing and reinforcing researcher-apprentice learning. Central to EXPLORA is the collection of attitudinal data through pre-observation interviews, offering researchers a deeper understanding of children's characteristics and contexts. This informs subsequent online observations, allowing researchers to focus on frequent interactions. Furthermore, researchers can validate preliminary assumptions with children. A means-ends analysis framework aids in the systematic analysis of data, shedding light on context, agency and homework-information searching processes children employ in their activities. To illustrate EXPLORA's capabilities, we present nine single case studies investigating Brazilian child-caregiver dyads' (children ages 9-11) use of technology in homework information-searching.