How much of the conflict analysis can be explained by tit for tat? If you back down easily when things turn against you, the risk of initiating conflict with you is well known and low. I'd imagine we're biased towards finishing a project against all odds to avoid perceptions of unreliability or weakness in group dynamics.
See Strategy of Conflict. Backing down on a commitment to fight is very bad in repeated games.
These findings supported the
prediction that administrators may seek to justify an ineffective course of action by escalating their commitment if resources to it.
It sounds like maybe they were trying to make it work rather than justifying it.
In a section titled Norms for Consistency, it talks a bit about President Carter and the importance of your convictions.
I think most likely this is really about the importance of leaders being wise and knowledgeable to a degree that exceeds the comprehension of the average person. People who don't understand your decision judge you based on how unflinchingly confident you are in it's correctness even though other people don't understand it, don't think it's a good idea (presumably because they don't understand it) and are against it because they have some conflicting agenda (such as personal gain).
It's 11 pages. I can't manage to finish it in one sitting but it looks like a worthwhile read.
I find amusing that your reaction is exactly as described in their experiment: people have a favorable bias toward someone who stuck to his course of action even after multiple failures than someone trying multiple different approaches.
It's double amusing since the paper starts by outlining multiple cases where an administration of company stuck to decision to ultimate failure, yet the one example that you chose to single out is the one about president Carter being seen as waffling.
This effectively claims that administrators are biased into making Martingale style doubling bets. As an administrator of a large program, cancelation may mean the end of your career "you're the guy who wasted X billion dollars on Y". Meanwhile many ineffective courses of action will eventually achieve their goals sufficiently well given enough resources.
As a successful administrator you are likely to double down on failures rather than let them fail. With the goal to recoup your losses from the first round.
This tends to mirror my experience with large organizations. Projects tied to a manager are rarely canceled unless the manager leaves (willingly or not). I wouldn't be surprised if unicorns had a similar trajectory of justifying previous losses with even bigger bets.
See Strategy of Conflict. Backing down on a commitment to fight is very bad in repeated games.