Respiratory testing with rapid antigen tests (RATs) in children under 5 years of age may be uncomfortable and presents specific challenges to testing due to compliance and procedural distress. The aim of this study was to investigate sensitivity and feasibility of self-collected nasal and saliva RAT tests compared with a combined nose and throat (CTN) swab PCR in children under 5. Children aged between 1month and 5 years, with confirmed COVID-19 or who were a household contact within 7 days were included. A saliva RAT, nasal RAT and CTN swab were collected by the parent. SARS-CoV-2 cycle threshold (Ct) values for CTN tested by PCR were compared with saliva and nasal RAT results. Parent preference for method of sample was recorded. Forty-one children were recruited with median age of 1.5 (interquartile range 0.7-4.0) years. Only 22/41 (54%) of parents were able to successfully collect a saliva RAT from their child. Sensitivity of the nasal RAT and saliva RAT was 0.889 (95% confidence interval (CI) 0.739-0.969) and 0.158 (95% CI 0.034-0.396), respectively. Upper limit of nasal RAT detection by CTN Ct value was higher than saliva (36.05 vs. 27.29). While saliva RAT was rated most comfortable, nasal RAT was rated the preferred specimen by parents for future testing, due to saliva collection difficulties and time taken. Rapid antigen testing with nasal RAT is a more feasible and sensitive method for SARS-CoV-2 detection in young children compared with saliva RAT.