The effect of database dirty data on h-index calculation

As all databases, the bibliometric ones (e.g. Scopus, Web of Knowledge and Google Scholar) are not exempt from errors, such as missing or wrong records, which may obviously affect publication/citation statistics and—more in general—the resulting bibliometric indicators. This paper tries to answer to the question “What is the effect of database uncertainty on the evaluation of the h-index?”, breaking the paradigm of deterministic database analysis and treating responses to database queries as random variables. Precisely an informetric model of the h-index is used to quantify the variability of this indicator with respect to the variability stemming from errors in database records. Some preliminary results are presented and discussed.