The ubiquitous weekly spelling test assumes that words are best learned by memorisation and testing but is this the best way? This study compared two well-known approaches to spelling instruction, the rule based and visual memory approaches. A group of 55 seven-year-olds in two Year 3 classrooms was taught spelling in small groups for three lessons a week, 20-min per lesson, over ten weeks. In the first intervention, students learned statistically likely spelling strategies for vowel sounds, syllable breaking strategies, and the doubling rule. In the second intervention, students used a look, say, cover, write, check, fix strategy, listed words in alphabetical order, and wrote them in sentences. The control group completed non-spelling activities. Results showed that although both intervention groups learned to spell taught words better than the control group, the rule-based approach had greater transfer to spelling of new words for both proficient and less proficient spellers.