Background: Some children with severe learning difficulties fail to begin word recognition. For these children there is a need for an effective and appropriate pedagogy. However, conflicting advice can be found regarding this derived from teaching approaches that are not based on a shared understanding of how reading develops or the skills that the non-reader needs to master. Purpose: In this research, three techniques for teaching word recognition in this context are described and compared: (1) the handle technique, (2) morphing method and (3) word alone. It also discusses whether it is appropriate for such small-scale research to influence pedagogy. Programme description: The handle technique uses an abstract mnemonic cue used to teach word recognition, and previous research indicates it is more successful than the presentation of words alone. The morphing method transforms a word into a photographic picture and a previous study suggested that it might also be more effective that presenting words alone. Sample: Six children between 11 and 13 years of age were selected. The criterion for selection was being unable to recognise any words from the British Ability Scales Reading Test. All the children attended a school for children with severe learning difficulties. Design and methods: A three-condition related design was used. The order in which the conditions were presented was counterbalanced and each child was taught 12 words, four words in each experimental condition. The children encountered each of the three methods and overall each word was taught via each method. Within conditions (teaching methods), the presentation of words was randomised. The number of words that the children could read (without cues) before each session was recorded, following the presentation of the uncued words in a random order. The difference in the number of words recognised between the three conditions was considered using a non-parametric statistical analysis. Results: The results suggest that the handle approach might be a more effective method of teaching word recognition. Conclusion: Research in this area is necessarily small in scale. However, it is ongoing and cumulative, and can give insights into potentially beneficial changes in classroom practice. (Contains 2 tables and 2 figures.)