Identification procedures for autism spectrum disorders in paediatric settings include screening and routine clinical surveillance. Screening tools are perceived as challenging, and a preference for direct observation has been reported in Ecuadorian paediatric settings. Augmentative observational procedures could prompt the application of screening tools, increase professionals' sense of self-efficacy and support referral decision-making. Following the recommendations that identification procedures should be tested in a group of children without autism from local populations, we tested a set of observational tasks on 125 children from 12 to 59 months of age in various socio-economic settings, in semi-experimental conditions and on a group of 33 non-autistic children in a paediatric-setting condition. Results indicated that the social referencing task did not produce the expected response in most children in semi-experimental conditions. Responses to the elicitation of a gesture were significantly influenced by socio-economic status and geographic area. Three other tested tasks (response to name, eye contact, response to joint attention) produced the expected responses in semi-experimental and clinical conditions, suggesting that they could be used in daily routines if fidelity to training is ensured. We discuss the implication of those findings in clinical practice and professional education in an Ecuadorian context.