[The title of this article comes from a T-shirt with ten advices on what to do when everything else fails]

It has always surprised me to realize how confident we physicists are of the good faith of our colleagues. We may argue endlessly over one graph or result, getting to the point of publically casting doubts on the dexterity or intelligence of our peers (yes, I've seen that), but we never seem to doubt -privately or otherwise- their scientific integrity.

Physicists are not different from academics of other disciplines as far as internal wars for power in the Universities are concerned, or for what concerns arrivism, fights for funding, and the like. We might extend the parallelism to the larger zoo of all human activities in fact: competition for power, money, personal gain and prestige is ubiquitous, and there is no reason to expect that the fascination for Nature's laws is a vaccine against vileness.  In fact, it is not. But when it comes to producing Scientific results, there you get visceral reactions, and usually physicists exhibit high moral standards there.

I have seen similar levels of "devotion to the cause" in other areas of human activity, where the individual annihilates his or her ego for the common good. This happens when one realizes, or even acts by instinct as if realizing, that one has better put ones' own extemporaneous personal gain aside in order to follow the straightest path to an all-important goal. So physicists are not so strange in this respect. What might be worth emphasizing, however, is that there is a collective benefit from the trust that generates from this behaviour. We are like an ant colony!

Let us consider the following situation to clarify what I mean. You are a researcher trying to measure a fundamental physical quantity in a particle physics experiment, and you are close to publishing what you believe to be a correct, carefully produced result. A deadline for a very important international conference is approaching, and you need to get your result approved as soon as possible, in order to present it there; that would in turn make it very likely that you are offered a very nice position you have recently applied for.

In order for your result to be approved, it needs to be reviewed by your peers. In the course of a meeting dedicated to the internal scrutiny of the analysis, a reviewer  points out that you might have made a small mistake, and requests that you verify a part of  the analysis and produce some confirming results with a new cross-check. You are quite sure that your result is sound, and you consider the check a waste of time, but in principle you must oblige to your reviewer's request.

As you leave the meeting you reckon that doing that check would take you a week, and you would lose the chance to present your result at the conference. No talk, no job. You feel frustrated. On the other hand, it occurs to you that you are the only one who can produce that check since most of the analysis is based on your own private code, and it would be quite hard -and very, very unlikely- that any of your colleagues set up to fully verify your results independently. In principle, you could cook up a fake check in a minute.

So what do you do ? Do you produce fake data to support your result, or do you comply with the request ?

I have been in a situation similar to the one described above as an analyst, and as a reviewer. Of course as analyst I have always followed the right path, and complied with any requests of careful verification of my results, cross checks, additional studies, and the like (not without uttering profanities in the process, let that be clear). But as a reviewer, have I ever doubted of my counterparts and the results they fed me with ?

Well, it might surprise you, but despite being in general a die-hard sceptic (proof if this be, guess what, that I don't believe in Supersymmetry), as a reviewer I have never doubted the good faith of the analysts whose results I was screening. And reviews can be rough rides: I can recall several instances when fierce arguments were had. The authors of an analysis would insist that they had done things correctly and I would insist that their proofs were unconvincing. Usually we stopped short of calling each other names, but that was about the only limit left untrespassed.

Similar confrontations sometimes go on for weeks, and may become very unpleasant -you are the bad guy stopping authors from publishing and moving on, and your decision is based on nothing more than personal intuition and maybe some indicia that a result might be biased or that one error bar might be incorrect. If this were not a discussion among physicists you might be advised to watch behind your back as you walk home at night!

No. In no instance do I recall having suspected I was being fooled. And I realize this is remarkable: after all, we know that scientific fraud does exist. I just happen to believe that my field of research is practically free from it. Manipulate the data ? God forbid ! We spend years of our lives building these giant detectors, and invest our careers in making them as performant as possible and in understanding their output. And we accept that a reviewer may stop our work indefinitely as a necessary evil - the process of review and verification of analysis results is a fundamental part of the production of scientific knowledge.

So are we missionaries, or fanatics ? No, we just love science, and we want to know more about the inner workings of our universe. We value the accuracy of the scientific results we produce, and we are ready to make small sacrifices in order to continue to trust those results as well as those of other experiments. It is a sort of virtuous circle. We leave data manipulation for other human activities, where the ends may at times justify the means, as argued in Machiavelli's "The Prince". Or maybe we are Machiavellian ourselves, but our ends are knowledge and truth, and our means are at times masochistic...