Astronomers have it easy: EPO when you are asking for no sacrifices or changes in behavior

EPO for astronomers is relatively easy, because we do not have to say "Here is an image of material very close to the horizon of a massive black hole, and therefore you should vaccinate your children!" and many other examples. We have, therefore, I think, an enhanced responsibility to focus on accuracy, error bars, alternatives, and how these change with new data. Steady State was a good bet in 1948, not in 1968. One need is to preserve our ability to reassure: that Trinity was not going to ignite atmospheric nitrogen (Bethe) and LHC particles would not trigger collapse of the universe into a black hole, because cosmic rays come at much higher energies. We need to distinguish at least three audiences, (1) those who come to us (read Scientific American, attend April APS public lectures) for whom "how we know" is vital and what constitutes understanding (equations? words for your grandmother?), (2) students in breadth courses, for whom "why is this useful?" is important (you will be safer if you know how electricity and statistics work), and (3) everybody else, for whom we must start "where they are" as in Naomi Oreskes' interview in 30/11 New Scientist. Items across the board: both special and general relativity matter for GPS; quantum mechanics matters for everything. Latest announcement isn't always truest and best (NSX2 and r process). Two sides aren't always, or even usually, equal. Changing your mind when you learn something new is good not bad. Even Albert E in Sept 1947 letter to Lemaitre accepted that his cosmological constant might be a solution to time scale problem, not a blunder!