Whatever the truth of the detail in the case of research data management you expose, there is enough undisputed evidence to suggest the practices may be more widespread. This is meat and drink to sociologists of scientific knowledge, and could add to the risk of undermining public confidence in science.
It also raises critical questions for the research assessment exercise. There has been considerable sensitivity about the question of multiple-author outputs and how sub-panels should assess the contributions of individuals. Now we have evidence that in some commercially sensitive industry-sponsored research authors may not have had access to the entire empirical data set; authors may not have participated in any of the data analysis or even in discussion about its nature; and authors may have had their manuscripts written by hired writers. How should we judge commercially sponsored research in this environment?