Over the years, I've told colleagues and friends about things I have seen or experienced. Many times, people have said that I should write them down so that they won't be lost and forgotten, since some of them might be useful parts of our history. I've been writing them down, without being sure what I would do with them. I decided to gradually post them on this website, and see what reactions I get. I suggest reading from the bottom up (starting with the August 2017 post "The Meritocracy"). Thoughtful and kind feedback would be useful for me, and would help me to revise the exposition to make it as useful as possible. I hope that while you read my stories you will ask yourself "What can I learn from this?" I'm particularly interested in knowing what you see as the point of the story, or what you take away from it. Please send feedback to asilverb@gmail.com. Thanks for taking the time to read and hopefully reflect on them!

I often run the stories past the people I mention, even when they are anonymized, to get their feedback and give them a chance to correct the record or ask for changes. When they tell me they're happy to be named, I sometimes do so. When I give letters as pseudonyms, there is no correlation between those letters and the names of the real people.

Thursday, October 19, 2023

Conflicts of Interest --- Say what you mean, mean what you say, and don't change the rules of the game

My main advice on Conflict of Interest policies is: 
Say what you mean, and mean what you say.

My main advice on enforcing Conflict of Interest policies is: 
Don't change the rules in the middle of the game.

Someone I'll call Y was on a panel to evaluate proposals. He had followed the Conflict of Interest policy to the letter, reporting in advance, in writing, all the conflicts of interest he was supposed to report.

However, in the presence of the entire panel, the Program Officer who ran the panel claimed that Y needed to recuse himself from the discussion about a proposal written by someone I'll call Q. According to the Program Officer, Y had a conflict of interest with Q, because Q was a co-author of X, and X was Y's wife.

Due to her many experiences with discrimination based on marital status, X kept her personal and professional lives separate, and did not discuss her personal life in professional settings. Neither X nor Y had ever told the Program Officer about any personal relationship between them. Anything the Program Officer thought he knew came from gossip. The Program Officer's discussion, in front of the panel, of his assumptions about X's and Y's personal lives violated their privacy. 

The official Conflict of Interest policy did not state that being a relative of someone's co-author constituted a conflict of interest. Y pointed this out to the Program Officer, who still insisted on recusal.

Curiously, for a later panel, that same Program Officer decided not to recuse himself when he expressed support for a proposal submitted by the husband of his own collaborator.

The last people to realize they have a conflict of interest are the people who have a conflict of interest. That's why it's important to have a clear, unambiguous algorithm for recusal, rather than letting people decide for themselves.

Similarly, the last people to realize that it's not OK to deviate from the official policy or make ad hoc decisions on a case-by-case basis are the enforcers of the policy. Many cases I saw personally where someone was held to a more stringent Conflict of Interest policy than the official one involved a woman, and the attitude of the male enforcers was, "This doesn't smell right to me. It feels like she's trying to get away with something."

Don't change the policy in midstream, and especially, don't change it in midstream based on a particular case; it's easy for the enforcers' prejudices to override fairness. When something doesn't technically violate the policy but "just doesn't smell right" to you, that doesn't justify making new rules on the fly. If it's not right, your Conflict of Interest policy should have covered it. If an objective party determines that the policy needs changing, change it after the current round; don't enforce a different policy than the one that's in place at the time.

Conflict of Interest policies often include personal, financial, professional, and other types of conflicts. Some policies require giving out much more information than is necessary, and sometimes that information gets circulated to co-PIs or others who don't need to know it. It should suffice to recuse oneself and say that one has a conflict of interest that falls under the policy, without being forced to divulge details about the conflict (e.g., one's marital status, or who one is married to).

To summarize, conflict of Interest policies should:
    (1) be clear, unambiguous, and sensible;
    (2) not require people to divulge irrelevant or unnecessary information;
    (3) be sent to everyone who needs to abide by it, early enough to give them a chance to refuse to take part in the activity to which it applies.