Abstract: The emergence of archaeology as an independent discipline during the early 1950s directly resulted in the establishment of three major archaeological journals – Wenwu (1950), Kaogu Xuebao (1951), and Kaogu (1955) – as outlets for information gathered in fieldwork; full site reports followed suit. Nowadays, we have roughly three dozen archaeological periodicals and surely more than a thousand monographs covering all areas and periods of Chinese (pre)history at our disposal. Said publications are our primary sources of information. However, the fact that they are anything but primary sources in a strictly methodological sense hardly gets acknowledged. In reality, excavation reports – preliminary as well as monographs – most often only provide a sample of actual data collected from archaeological sites. Consequently, we are constantly dealing with deliberate choices of editors on what particular information to divulge. This paper shall demonstrate that nature and quality of findings are the main decisive factors in this process. For instance, even looted tombs dating from the Zhanguo and Han periods yielding manuscripts generally take precedence over undisturbed graves discovered at the same cemetery simply because they contained manuscripts. Many conclusions concerning such burials are therefore based on a rather small number of published tombs while the often more representative majority of equally accessible graves remain unnoticed. In short, the paper is aiming to raise awareness for a pressing methodological problem. In doing so, it will address various rationales behind the practice of presenting selective evidence in excavation reports and suggest ways to cope with it.