Archive for November 2014
Roger Bacon- Consider the diverse benefits of randomization. Piercing through the vagaries of chance and deception, it allows us to discern truly the causes and consequences of action.
Trollius Maximus- Yet, there are costs to randomization. A trial requires an exceeding amount of time and treasure, while other methods can be done in short time by a single natural philosopher of modest means. Even if trials were as easy to conduct, they are harder to generalize to far-away lands and eras. To say nothing of the ethics of rewarding one man while spurning another, all according to the flip of a coin.
Bacon: I do not say that randomized studies are the only way. Where traditional methods fail, or when the question is of true import, we will find the costs of randomization to be of little matter.
Trollius: Even the ethical costs? Will you so lightly toss aside the question of the good?
Bacon: What could be the flaw of helping one while passing another by, so long as I do no harm to the other? Does not every good deed only help one or a few, while the multitudes remain ignorant of the deed?
Trollius: Have you indeed done no harm to those passed over in your trial, or those who bear witness to your study? Have you not convinced them that the world was a more random place than they had thought, that their own actions matter little compared to the all-powerful, uncaring hand of chance?
Bacon: Perhaps I have. Indeed, I stand convinced. I shall demonstrate in a paper using a randomized trial that exposure to randomization undermines people’s conviction that they are the master of their fate, the locus of their control, and I shall show that this new belief causes them great harm.
Trollius: Your paper would roil the world of randomistas.
Bacon: Yet I worry that natural philosophers will still turn too readily to randomization, since they gain most of the benefits from doing such studies, while experiencing only a fraction of these costs. Witness how psychologists continue to use deception, while economists and others spam the world with audit studies.
Trollius: Ah, but your work would be so convincing, your brilliance could shatter the randomista movement with a single blow. They would return to running cross-country regressions, and Sophists will carry the day once more.
Bacon: I see. Your trolling has convinced me to stay silent, for the good of the world I must maintain the Noble Lie that randomization is the ideal and the future of natural philosophy.
Trollius: By silence, you mean not writing a paper. Certainly a blog post could do no harm; people would find it funny, rather than a failed attempt at cleverness and a lame imitation of Brad DeLong (who you shouldn’t be trying to imitate anyway).
Bacon: Indeed. To the blogosphere!