The measures of mathematical knowledge for teaching developed at the University of Michigan in the U.S., have been adapted and used in studies measuring teacher knowledge in several countries around the world. In the adaptation, many of these studies relied on comparisons of item parameters and none of them considered a comparison of raw data. In this article, we take advantage of having access to the raw data from the adaptation pilot studies of the same instrument in Norway and Slovakia (149 practicing elementary teachers in Norway, 134 practicing elementary teachers in Slovakia) that allowed us to compare item parameters and teachers' ability estimates on the same scale. Statistical analysis showed no significant difference in the mean scores between the Norwegian and the Slovak teachers in our samples and the paper provides further insight into the issues of cross-national adaptations of measures of teachers' knowledge and the limitations of the methods commonly applied in the item adaptation research. We show how item adaptations can be refined by combining robust quantitative methods with qualitative data, how decisions on adaptation of individual items depend on context and purpose of the adaptation, and how comparability and heterogeneity of samples affects interpretation of the results.