id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_37pxtlmwtzhtted3jppc4tyu5m	Katie Steele	The Diversity of Model Tuning Practices in Climate Science	2016	13	.pdf	application/pdf	6099	455	57	We examine one example and show that, in employing classical hypothesis testing, it involves calibrating a base model against data that are also used to confirm the of base models/theories: that use-novel data have a special role in confirmation and, more strongly, that data cannot be used twice, both for calibration (2007) then compare the performance of these 16 base models, assuming that inclusion of the term biTi is necessary just in case the estimated bi is significantly different from zero (at the 95% level). Contrary to the intuitive position, classical hypothesis testing does not respect use novelty and the no-double-counting rule: calibration is the assessment of particular model-instance hypotheses—these hypotheses are either model (when the confidence interval for some free parameter does not include zero).3 Thus, there is double counting, and data used for confirmation the score estimating the average predictive accuracy of the base-model procedure given n data points:	./cache/work_37pxtlmwtzhtted3jppc4tyu5m.pdf	./txt/work_37pxtlmwtzhtted3jppc4tyu5m.txt