Is it true that articles in high-impact journals involve, in average, more coauthors?

A while ago, I was involved in writing a review article, which finally included around 270 references. Being not very experienced in using BibTeX, I had to manually enter many references in my TeX-file. There I noticed that Nature and Science entries took more time to handle because they usually come with more than 4 coauthors.

Today, I searched a bit on the internet (around 2 hours) and I did not find a solid answer to my question. Instead, I found very interesting articles from Eugen Tarnow about coauthorship in physics and junior physicists’ perception of authorship. In these articles, based on detailed surveys, Tarnow discusses that not all the authorships are awarded decently. As an example, he has found that: “The probability of any third and subsequent coathors being judged as inappropriate is 23% for the APS guideline, [and] 67% for the tighter guideline of the International Committee of Medical Journal Editors“.