Ah yes I fully agree on your final test! I always ask myself "If I presented this at a conference, what things would the cynical old professor point out, and can I make a plot to convince them that's not the case"
Also I have been diving into Gelman's Bayesian Data Analysis this week so good to see it get a mention!
Love the three tests. My own rather like the first is the counterfactual - what would other values need to be for this to be true /untrue. Great for ROI analysis to check expectations in advance.
Yes I like that a lot. The danger otherwise is that you get lost in the problem and forget how your conclusions relate to everything else. Then someone comes along and points out the obvious and you end up looking stupid!
Thanks @pstobbs!
Ah yes I fully agree on your final test! I always ask myself "If I presented this at a conference, what things would the cynical old professor point out, and can I make a plot to convince them that's not the case"
Also I have been diving into Gelman's Bayesian Data Analysis this week so good to see it get a mention!
Very wise. Always anticipate the next smug objection!
Love the three tests. My own rather like the first is the counterfactual - what would other values need to be for this to be true /untrue. Great for ROI analysis to check expectations in advance.
Yes I like that a lot. The danger otherwise is that you get lost in the problem and forget how your conclusions relate to everything else. Then someone comes along and points out the obvious and you end up looking stupid!
like the three tests Simon!