This study explores the extent to which authors with different impact and productivity levels cite journals, institutions, and other authors through an analysis of the scientific papers of 37,717 authors during 1990–2013. The results demonstrate that the core‐scatter distribution of cited authors, institutions, and journals varies for authors in each impact and productivity class. All authors in the science network receive the majority of their credit from high‐impact authors; however, this effect decreases as authors' impact levels decrease. Similarly, the proportion of citations that lower‐impact authors make to each other increases as authors' impact levels decrease. High‐impact authors, who have the highest degree of membership in the science network, publish fewer papers in comparison to highly productive authors. However, authors with the highest impact make both more references per paper and also more citations to papers in the science network. This suggests that high‐impact authors produce the most relevant work in the science network. Comparing practices by productivity level, authors receive the majority of their credit from highly productive authors and authors cite highly productive authors more frequently than less productive authors.