I looked at this because although I specified a tolerance of 0.1, TAM kept iterating well beyond this point in terms of absolute deviance change. In tam.np, there is no reference to dev_crit (i.e. absolute or relative), but convD is always referenced against relative change.
I looked at this because although I specified a tolerance of 0.1, TAM kept iterating well beyond this point in terms of absolute deviance change. In tam.np, there is no reference to dev_crit (i.e. absolute or relative), but convD is always referenced against relative change.