The aim of our study was to establish bench-mark values for speech intelligibility in noise for children. We also considered which of the two dichotic discrimination tests, Feldmann's or Uttenweiler's, would be most suitable for use within this age group. We examined 102 children aged between 7 and 10 years (second year, primary school), with an average age of 8 years. After confirming normal hearing acuity (examination of the ear, tympanometry and audiogram), speech perception was tested. This was performed with and without noise using the Göttingen Audiometric Speech Test for Children II and the monosyllables of the Freiburger Speech Intelligibility Test. The percentage of word intelligibility was recorded. Furthermore, dichotic hearing was tested using a combination of Feldmann and Uttenweiler dichotic discrimination tests. We rated the percentage of correct word pair repetitions with the correct article. The results showed that the best method for testing speech perception in noise for this age group was the Göttingen Test II for Children. Speech perception below 70% for word intelligibility should be considered as pathologic. For testing dichotic hearing, Uttenweiler's dichotic discrimination test for children was most suitable. In this test values below 80% should be considered as pathologic.