Background: Two major conceptions of learning exist: reproducing new material and transforming material to make meaning. Teachers' understandings of what learning is probably influence their teaching practices and student academic performance. Aims: To validate a short scale derived from Tait, Entwistle, & McCune's (1998) ASSIST inventory and to determine and compare the strength of agreement New Zealand and Queensland primary and secondary teachers had for both conceptions. Samples: Two survey studies with three populations provided valid data: 235 NZ Primary teachers in 2001, and 784 Primary and 614 Secondary Queensland teachers in 2003. A survey of 81 NZ secondary teachers in 2000 did not have enough participants to generate stable estimates. Methods: Five items defining learning were administered using a six point, positively-packed agreement rating scale. Data were analysed with MMLE confirmatory factor analysis with oblimin rotation. Multiple models were compared and results from the best fitting model (CFI and TLI greater than 0.90; RMSEA less than 0.08) for all three samples are reported. Cohen's "d" effect size was used to determine significance of differences in conceptions mean scores. Results and Conclusions: Psychometric properties of the two scales were good. Conceptions of learning were structured as two inter-correlated factors related to transforming and reproducing conceptions of learning. All samples agreed more with the transforming than the reproducing conception of learning; however, there were small group differences in mean scores. The response scale and items generated sufficient variation to detect differences in teachers' attitudes towards transforming and reproducing conceptions of learning. (Contains 2 tables, 3 figures, and 2 footnotes.)