To assess the reliability of touch to predict fever in children.200 children who reported with fever formed the study material. Group I consisted of 100 children between 0-1 year of age and Group II consisted of 100 children between 6-12 years of age. Preterm, neonates under warming device, tachypnoeic and hypothermic were excluded from the study. The caregiver (CG) and the medical staffs (MS) response regarding presence or absence of fever by touch was noted in each child. Both were blinded to each other's response. Immediately temperature was recorded by calibrated rectal thermometer in Group I and calibrated axillary thermometer in Group II.The CG's touch had a sensitivity of 70.5% specificity of 40.9%, PPV of 38% NPV of 72.9%, PLR was 1.16 and NLR was 0.75. The MS's touch had a sensitivity of 78.0%, specificity of 63.6%, PPV of 38.0% NPV 84.8%, PLR of 2.08 and NLR of 0.36. There is over and under diagnosis of fever by both, the former being more by the CG reflecting the parental anxiety. The MS's touch is better to affirm or negative fever as compared to CG. The best site to palpate for presence of fever was abdomen, neck and forehead.Touch is not a valid screening test for fever. It is recommended that a thermometer must always be used by the MS to record fever and CG must be motivated for the same.