In the German HeLp-study (2007/08), a longitudinal intervention and control study, orthography data of more than 1000 fifth graders has been collected by using a systematic orthography test based on the graphematic linguistic theory and on a differentiated competence model. Within the data, an intervention group and a control group have been differentiated. The data has been assessed on three different time points in grade five which is for the intervention group before, during and after the intervention. It will be shown how the complex data can be analyzed and compared within the two groups and the three time points with the help of the Item Response Theory (IRT) in order to gain a maximum insight into the students' development of differenciated orthography skills. The IRT analysis was done on a whole-word level and then differentiated into four subskills. Thereby, the distinct development of subskills can be further analyzed. On top of that, anchor items were compared. We want to present the methodology and also demonstrate, interpret and discuss selected results of the study. It will be shown in how far the differentiation of subskills is important as well as how the two groups differ in their development.