Experimental data were excluded from analyses of effects of alendronate treatment on fatigue life of bone specimens from canine ribs.
We specify why the exclusion criteria were inappropriate.
Exclusion of these data changed statistically insignificant results to significant results for certain comparisons.
An additional one third of include tests stopped prior to complete failure, but these data were not analyzed ascensored data.
Modulus loss did not show any treatment effect whether all of the data or only selected data are considered.