<p>Bibliometric reviews have increased in popularity as a way to summarize literature areas using quantitative, network-based methods. When conducting such reviews, it is assumed that bibliometric data are sufficiently accurate that errors associated with author names and references are unlikely to influence the results. We challenge this notion and discuss the types of errors (usually hidden) that compromise bibliometric data as well as the consequences of those errors. We provide a tutorial and R code for cleaning bibliometric data that focuses on identifying and fixing various types of errors. We demonstrate this with an illustrative example and also a re-analysis of a previously published bibliometric review. In both examples, we show that typical approaches to the handling of bibliometric data can produce inaccurate results and conclusions. The alternative that we develop and illustrate allows the bibliometrician to avoid these inaccuracies.</p>