Abstract

When learning to read, children are regularly exposed to words with novel symbols and spellings in stories and lists. I investigated the differential effect of self-teaching in and out of context, as well as the moderating effect of stimulus set size, word length, and word regularity, on the formation of orthographic representations. In Study 1, adult participants practiced reading regularised spellings of real words that were written in an artificial orthography across four experiments (N = 114). Participants trained with either a small set of words or with a large set of words. The first dependent variable was word reading accuracy measured during training (over 6 trials) and 3-days later at post-test. The second and third dependent variables were spelling recognition and production also measured 3-days later at post-test. Generalization of learning was also explored. Results indicate that training in context bolstered reading accuracy during and after training. Yet the highest spelling scores were noted following training in isolation. Further, self-teaching a large set of diverse words in isolation increased reading accuracy, closing the gap with context. Overall, self-teaching in context helped establish orthographic representations that supported reading accuracy. Whereas reading in isolation more highly specified orthographic representations, enabling spelling accuracy—especially when training with a large set of words. Study 2 explored whether these effects could be obtained with children. Children read regular and irregular sets of words of varying length written in standard English orthography across two experiments (N = 64). Reading practice took place in both context and isolation and stimulus set size was experimentally manipulated. Children participants were then asked to complete a variety of reading and spelling retention tasks at post-test. Reading in context supported reading accuracy during training and 1 week later. Like adults, children spelled more words correctly from the isolation training condition. The greatest isolation advantage was for exception words. We also noted that when children had fewer opportunities to map orthographic representations onto pronunciations, training in context was more advantageous than training in isolation. This signals the scope and limitations of reading in context and highlights the faciliatory effect of reading amount.