Distribution of Productivity is Non-normal
In many industries and fields, metrics of success often follow something like a power distribution. Examples include incomes of sports professionals, sales of writers and musicians, publication output and citation counts of academics (see some of the work by Aguinis and colleagues on star performers).
Success leads to more resources: In general, the world rewards the best and most capable in a field. This occurs though giving more exposure in high impact journals and so on. It also occurs where success leads to more resources (e.g., grant money; collaborators willing to put you on a paper; better collaborators; better PhD students; more time to do research; etc.).
Ability is non-linearly related to productivity: In addition, writing journal articles for high impact journals takes substantial expertise. There are many academics who never publish in high impact journals, presumably because their research skills, available data, or whatever are just not at that level. In contrast, other researchers are able to consistently produce output at that level. I'd also note that this is not a fixed characteristic. People do learn and improve. But ultimately, high level skills and persistence are required to consistently publish in top-tier journals.
In general, whenever you compare the publication of one academic to another, there are a number of things to consider:
- Quality of publications: The impact factor is a crude approximation in first instance but average impact factor becomes more informative when average over an academics portfolio of publications.
- Field-specific performance norms: Some fields publish more or less, and related to the next points, some fields have more co-authors per paper, which makes it easier to get many papers.
- Number of co-authors per paper: Some authors publish with few co-authors; some publish with many. A fairer comparison will often be obtained by dividing each publication by number of co-authors, and summing this fractionated total.
- Weighting by contribution to paper: In general, with more co-authors, any one person's contribution is less. In many fields, the first author makes the main contribution. Thus, one quick way of fractionating, is to get a count of number of first author papers only. Thus, if someone is the head of a big group, which get their name on many things, this often won't transfer to first author papers.
Aguinis, H., & O'Boyle, E. (2014). Star performers in twenty‐first century organizations. Personnel Psychology, 67(2), 313-350.