An ideal dossier gives the reader a sense of the contributions, strengths, weaknesses, and the vulnerabilities of the contributions. There is a detailed discussion of the candidate's work, critical and appreciative, indicating problems or disagreements on the part of the discussant. If there are issues such as too many publications and the candidate needs to focus more and slow down, we are told that (and so should the candidate). If there are issues that might come up, such as low productivity for a period of time, we are told why or that at least the department appreciate there is a problem there. Rather and hold back internal concerns, they are shared--in part because those concerns are very likely to come up in the university committee, and it is better to deal with it ahead of time. Comments from members of the university committee who know the candidate are unlikely to be helpful, and letters from those in the university who want to support the candidate (no matter how prestigious they are, especially if they are outside the department) rarely help.

In general, almost all rave letters, almost all letters from advisors, collaborators (except to attest to who did what in the work), almost all unbalanced assessments hurt the candidate. The strongest recommendations are those that consider the work and the contribution, perhaps even engage it quite seriously and contentiously, and then in the end say, she is terrific.

What always hurts are excuses, such as this second ranked journal is really important. (It's not that the excuse may not be true, but it smacks of special pleading.) If the work is good, that is the main issue, and if it is in a venue that is likely to be seen that matters too.

It's better to tell your true assessment of the candidate, than to wax eloquently but incredibly. If this is the best book on Icelandic dentisty, it won't help to claim that Icelandic dentistry is the major field, but it might help to know that the analysis in the book has deeply influenced how people think of dental health all over. If the Icelandic dentistry book wins an award from the Icelandic dentisty association (or another two modifier association, in which the number of potential awardees is typically 2-4 rather than 20) it's unlikely to impress people to say it is an award-winning book or article.

If the candidate does not get the grants normally required, saying that the standards have gone up is an excuse. Presumably, your university wants to have candidates who have received the grants, not the almosts. Of ocurse if the acceptance cutoff was 9% and this candidate stood at 10% it may be useful to mention that.

Finally, while the number of stars in the sky is very large, on Earth the number of academic stars is rather more limited. The claim should be reserved and used carefully. At least if you want people to believe you when it really matters.

I would never buy a used car from most dossiers, for I can smell a lemon. Now, I might discount the information, and offer to pay less for the car (or the candidate), but is that what you want me to do.