Software inspection facilitates product improvement in early phases of software development by detecting defects in various types of documents, e.g., requirements and design specifications. Empirical study reports show that usage-based reading (UBR) techniques can focus inspectors on most important use cases. However, the impact of inspector qualification and learning effects in the context of inspecting a set of documents in several sessions is still not well understood. This paper contributes a model for investigating the impact of inspector capability and learning effects on inspection effectiveness and efficiency in a large-scale empirical study in an academic context. Main findings of the study are (a) the inspection technique UBR better supported the performance inspectors with lower experience in sequential inspection cycles (learning effect) and (b) when inspecting objects of similar complexity significant improvements of defect detection performance could be measured.