The set of microbiologists is only one of a virtually infinite number of sets of similar size for which one could dream up some tenuous, hypothetical link to a worldwide terrorist plot. The set itself has been "cherry-picked" after the fact. Not only that, but the precise borders of the set have been tailored and defined to make the coincidence appear bigger than it really is. For example, was the most prominent scientist in this set, Don Wiley, really a microbiologist? Only in a very loose sense -- he would normally be considered a structural biologist. If the real, untailored set is "biologists" rather than "microbiologists," we might easily be talking about a pool of 200,000 people instead of 20,000 people, in which case the "coincidence" would disappear into the background. This is how we make mountains out of molehills, statistically speaking.
I think I see your objection. Because the cohort is not truly randomly selected, it cannot be correlated against the random distribution of unnatural death rate in the general population, is that correct?
Actually, there is a simpler flaw in the analysis that no one has detected yet: there are only six people mentioned who were U.S. citizens: Que*, Wiley*, Schwartz, Holzmayer, Huang, Mostow. Out of these, only the starred names were members of the American Society for Microbiology. Without some more digging that I don't have the time for, I cannot tell if the others were active participants in microbiology research.
For example, was the most prominent scientist in this set, Don Wiley, really a microbiologist?
He certainly possessed enough interest in the field to register as a member of the American Society for Microbiology. He was a structural biologist, but very active in in applying crystallography and his other structural biology work towards virological and immunological research as far as I can tell.
I must concede that without a more rigorous background in statistical theory, I cannot tell whether applying an outcome given a random distribution population to a relatively randomly selected group out of the population affects the validity, and so from a science standpoint I must say that I cannot commit to either view (yours included) without more information. I think that you are saying without the morbidity rates for microbiologists (or just biologists), we cannot tell if the recent events rise above the noise level of the morbidity rate for that specific group.
Based upon what I know about the insurance business, I don't think that is necessary; my insurance agent confirmed to me that Farmer's doesn't write different life insurance policies for microbiologists (or biologists) than the general office worker. Given that, my hunch is that any grouping whose morbidity rates do not rise above that of the general distribution is just as valid a set as the general population set. Confirmation from a commercial application of morbidity rates is somewhat persuasive however, absent any government regulation that prohibits underwriting based upon profession (which I am not aware exists).
In the most restrictive sense, if only Que and Wiley were counted, then I would say that this falls under the standard morbidity rate. The flaw with this view is that we don't even agree if they were microbiologists. Perhaps if we counted all people who would be active researchers that could assist a biowarfare effort (offensive or defensive), we can get away from the (probably media-mangled) "microbiologist" label. However, I suppose both of us probably don't have enough information to really tell for sure either way whether these deaths are significant. Maybe we should agree to disagree?